training_log.txt

Epoch [1/3], Step [1/3236], Loss: 9.2222, Perplexity: 10119.8081
Epoch [1/3], Step [2/3236], Loss: 8.9599, Perplexity: 7784.4743
Epoch [1/3], Step [3/3236], Loss: 8.7379, Perplexity: 6234.9778
Epoch [1/3], Step [4/3236], Loss: 8.3448, Perplexity: 4208.3934
Epoch [1/3], Step [5/3236], Loss: 7.9231, Perplexity: 2760.3687
Epoch [1/3], Step [6/3236], Loss: 7.3448, Perplexity: 1548.1231
Epoch [1/3], Step [7/3236], Loss: 6.6534, Perplexity: 775.4096
Epoch [1/3], Step [8/3236], Loss: 5.8596, Perplexity: 350.5727
Epoch [1/3], Step [9/3236], Loss: 5.4005, Perplexity: 221.5211
Epoch [1/3], Step [10/3236], Loss: 5.1418, Perplexity: 171.0311
Epoch [1/3], Step [11/3236], Loss: 4.9463, Perplexity: 140.6581
Epoch [1/3], Step [12/3236], Loss: 4.7334, Perplexity: 113.6760
Epoch [1/3], Step [13/3236], Loss: 4.7012, Perplexity: 110.0836
Epoch [1/3], Step [14/3236], Loss: 4.6168, Perplexity: 101.1734
Epoch [1/3], Step [15/3236], Loss: 4.7380, Perplexity: 114.1999
Epoch [1/3], Step [16/3236], Loss: 4.6842, Perplexity: 108.2203
Epoch [1/3], Step [17/3236], Loss: 4.7004, Perplexity: 109.9936
Epoch [1/3], Step [18/3236], Loss: 4.6697, Perplexity: 106.6614
Epoch [1/3], Step [19/3236], Loss: 4.6944, Perplexity: 109.3329
Epoch [1/3], Step [20/3236], Loss: 4.6329, Perplexity: 102.8121
Epoch [1/3], Step [21/3236], Loss: 4.4227, Perplexity: 83.3241
Epoch [1/3], Step [22/3236], Loss: 4.3504, Perplexity: 77.5083
Epoch [1/3], Step [23/3236], Loss: 4.5945, Perplexity: 98.9370
Epoch [1/3], Step [24/3236], Loss: 4.5773, Perplexity: 97.2487
Epoch [1/3], Step [25/3236], Loss: 4.3762, Perplexity: 79.5368
Epoch [1/3], Step [26/3236], Loss: 4.2919, Perplexity: 73.1049
Epoch [1/3], Step [27/3236], Loss: 4.2987, Perplexity: 73.6025
Epoch [1/3], Step [28/3236], Loss: 4.3280, Perplexity: 75.7936
Epoch [1/3], Step [29/3236], Loss: 4.4594, Perplexity: 86.4350
Epoch [1/3], Step [30/3236], Loss: 4.2410, Perplexity: 69.4792
Epoch [1/3], Step [31/3236], Loss: 4.2691, Perplexity: 71.4567
Epoch [1/3], Step [32/3236], Loss: 4.1340, Perplexity: 62.4293
Epoch [1/3], Step [33/3236], Loss: 4.3162, Perplexity: 74.9069
Epoch [1/3], Step [34/3236], Loss: 4.3710, Perplexity: 79.1212
Epoch [1/3], Step [35/3236], Loss: 4.1243, Perplexity: 61.8220
Epoch [1/3], Step [36/3236], Loss: 4.3204, Perplexity: 75.2174
Epoch [1/3], Step [37/3236], Loss: 4.3087, Perplexity: 74.3420
Epoch [1/3], Step [38/3236], Loss: 4.0647, Perplexity: 58.2451
Epoch [1/3], Step [39/3236], Loss: 4.0999, Perplexity: 60.3326
Epoch [1/3], Step [40/3236], Loss: 4.0279, Perplexity: 56.1410
Epoch [1/3], Step [41/3236], Loss: 4.0969, Perplexity: 60.1524
Epoch [1/3], Step [42/3236], Loss: 4.0332, Perplexity: 56.4432
Epoch [1/3], Step [43/3236], Loss: 4.0432, Perplexity: 57.0067
Epoch [1/3], Step [44/3236], Loss: 4.0604, Perplexity: 57.9948
Epoch [1/3], Step [45/3236], Loss: 4.1130, Perplexity: 61.1286
Epoch [1/3], Step [46/3236], Loss: 4.3886, Perplexity: 80.5274
Epoch [1/3], Step [47/3236], Loss: 4.0479, Perplexity: 57.2796
Epoch [1/3], Step [48/3236], Loss: 4.1846, Perplexity: 65.6697
Epoch [1/3], Step [49/3236], Loss: 4.0579, Perplexity: 57.8551
Epoch [1/3], Step [50/3236], Loss: 4.1244, Perplexity: 61.8310
Epoch [1/3], Step [51/3236], Loss: 3.9409, Perplexity: 51.4637
Epoch [1/3], Step [52/3236], Loss: 4.0012, Perplexity: 54.6647
Epoch [1/3], Step [53/3236], Loss: 4.0632, Perplexity: 58.1615
Epoch [1/3], Step [54/3236], Loss: 4.0012, Perplexity: 54.6624
Epoch [1/3], Step [55/3236], Loss: 3.9227, Perplexity: 50.5388
Epoch [1/3], Step [56/3236], Loss: 4.8273, Perplexity: 124.8763
Epoch [1/3], Step [57/3236], Loss: 3.8718, Perplexity: 48.0287
Epoch [1/3], Step [58/3236], Loss: 3.8701, Perplexity: 47.9464
Epoch [1/3], Step [59/3236], Loss: 3.8084, Perplexity: 45.0766
Epoch [1/3], Step [60/3236], Loss: 4.0130, Perplexity: 55.3135
Epoch [1/3], Step [61/3236], Loss: 3.9273, Perplexity: 50.7707
Epoch [1/3], Step [62/3236], Loss: 3.8100, Perplexity: 45.1510
Epoch [1/3], Step [63/3236], Loss: 3.7824, Perplexity: 43.9196
Epoch [1/3], Step [64/3236], Loss: 3.8291, Perplexity: 46.0209
Epoch [1/3], Step [65/3236], Loss: 3.9290, Perplexity: 50.8566
Epoch [1/3], Step [66/3236], Loss: 4.1535, Perplexity: 63.6569
Epoch [1/3], Step [67/3236], Loss: 3.7226, Perplexity: 41.3729
Epoch [1/3], Step [68/3236], Loss: 3.9497, Perplexity: 51.9172
Epoch [1/3], Step [69/3236], Loss: 3.8006, Perplexity: 44.7295
Epoch [1/3], Step [70/3236], Loss: 3.7500, Perplexity: 42.5225
Epoch [1/3], Step [71/3236], Loss: 3.7877, Perplexity: 44.1562
Epoch [1/3], Step [72/3236], Loss: 3.8163, Perplexity: 45.4342
Epoch [1/3], Step [73/3236], Loss: 3.7098, Perplexity: 40.8470
Epoch [1/3], Step [74/3236], Loss: 3.6285, Perplexity: 37.6562
Epoch [1/3], Step [75/3236], Loss: 3.7639, Perplexity: 43.1173
Epoch [1/3], Step [76/3236], Loss: 3.8892, Perplexity: 48.8696
Epoch [1/3], Step [77/3236], Loss: 3.9653, Perplexity: 52.7346
Epoch [1/3], Step [78/3236], Loss: 3.7310, Perplexity: 41.7204
Epoch [1/3], Step [79/3236], Loss: 3.6447, Perplexity: 38.2715
Epoch [1/3], Step [80/3236], Loss: 5.2081, Perplexity: 182.7526
Epoch [1/3], Step [81/3236], Loss: 3.7047, Perplexity: 40.6392
Epoch [1/3], Step [82/3236], Loss: 3.6895, Perplexity: 40.0266
Epoch [1/3], Step [83/3236], Loss: 3.7342, Perplexity: 41.8556
Epoch [1/3], Step [84/3236], Loss: 3.7245, Perplexity: 41.4513
Epoch [1/3], Step [85/3236], Loss: 3.8082, Perplexity: 45.0689
Epoch [1/3], Step [86/3236], Loss: 3.7638, Perplexity: 43.1109
Epoch [1/3], Step [87/3236], Loss: 3.7813, Perplexity: 43.8742
Epoch [1/3], Step [88/3236], Loss: 3.7554, Perplexity: 42.7501
Epoch [1/3], Step [89/3236], Loss: 3.5927, Perplexity: 36.3318
Epoch [1/3], Step [90/3236], Loss: 3.6556, Perplexity: 38.6890
Epoch [1/3], Step [91/3236], Loss: 3.5765, Perplexity: 35.7466
Epoch [1/3], Step [92/3236], Loss: 3.8339, Perplexity: 46.2407
Epoch [1/3], Step [93/3236], Loss: 3.5202, Perplexity: 33.7929
Epoch [1/3], Step [94/3236], Loss: 3.6599, Perplexity: 38.8585
Epoch [1/3], Step [95/3236], Loss: 3.7432, Perplexity: 42.2328
Epoch [1/3], Step [96/3236], Loss: 3.7462, Perplexity: 42.3590
Epoch [1/3], Step [97/3236], Loss: 3.7054, Perplexity: 40.6666
Epoch [1/3], Step [98/3236], Loss: 3.4217, Perplexity: 30.6226
Epoch [1/3], Step [99/3236], Loss: 3.6837, Perplexity: 39.7921
Epoch [1/3], Step [100/3236], Loss: 4.0734, Perplexity: 58.7584
Epoch [1/3], Step [101/3236], Loss: 3.6107, Perplexity: 36.9928
Epoch [1/3], Step [102/3236], Loss: 3.6247, Perplexity: 37.5117
Epoch [1/3], Step [103/3236], Loss: 3.6528, Perplexity: 38.5807
Epoch [1/3], Step [104/3236], Loss: 3.6431, Perplexity: 38.2119
Epoch [1/3], Step [105/3236], Loss: 3.5513, Perplexity: 34.8584
Epoch [1/3], Step [106/3236], Loss: 3.6161, Perplexity: 37.1907
Epoch [1/3], Step [107/3236], Loss: 3.6272, Perplexity: 37.6061
Epoch [1/3], Step [108/3236], Loss: 3.5055, Perplexity: 33.2976
Epoch [1/3], Step [109/3236], Loss: 3.5031, Perplexity: 33.2179
Epoch [1/3], Step [110/3236], Loss: 3.5878, Perplexity: 36.1539
Epoch [1/3], Step [111/3236], Loss: 3.5627, Perplexity: 35.2590
Epoch [1/3], Step [112/3236], Loss: 3.5829, Perplexity: 35.9766
Epoch [1/3], Step [113/3236], Loss: 3.5030, Perplexity: 33.2147
Epoch [1/3], Step [114/3236], Loss: 3.6102, Perplexity: 36.9734
Epoch [1/3], Step [115/3236], Loss: 3.5452, Perplexity: 34.6481
Epoch [1/3], Step [116/3236], Loss: 3.4900, Perplexity: 32.7862
Epoch [1/3], Step [117/3236], Loss: 3.5004, Perplexity: 33.1288
Epoch [1/3], Step [118/3236], Loss: 3.3593, Perplexity: 28.7702
Epoch [1/3], Step [119/3236], Loss: 3.3908, Perplexity: 29.6902
Epoch [1/3], Step [120/3236], Loss: 3.6826, Perplexity: 39.7498
Epoch [1/3], Step [121/3236], Loss: 4.1208, Perplexity: 61.6099
Epoch [1/3], Step [122/3236], Loss: 3.7408, Perplexity: 42.1299
Epoch [1/3], Step [123/3236], Loss: 3.4818, Perplexity: 32.5168
Epoch [1/3], Step [124/3236], Loss: 3.6786, Perplexity: 39.5905
Epoch [1/3], Step [125/3236], Loss: 4.3509, Perplexity: 77.5481
Epoch [1/3], Step [126/3236], Loss: 3.4851, Perplexity: 32.6265
Epoch [1/3], Step [127/3236], Loss: 3.5109, Perplexity: 33.4780
Epoch [1/3], Step [128/3236], Loss: 3.5351, Perplexity: 34.2990
Epoch [1/3], Step [129/3236], Loss: 3.3898, Perplexity: 29.6595
Epoch [1/3], Step [130/3236], Loss: 3.4621, Perplexity: 31.8850
Epoch [1/3], Step [131/3236], Loss: 3.5279, Perplexity: 34.0512
Epoch [1/3], Step [132/3236], Loss: 3.4377, Perplexity: 31.1165
Epoch [1/3], Step [133/3236], Loss: 3.5236, Perplexity: 33.9047
Epoch [1/3], Step [134/3236], Loss: 3.5730, Perplexity: 35.6249
Epoch [1/3], Step [135/3236], Loss: 3.4620, Perplexity: 31.8822
Epoch [1/3], Step [136/3236], Loss: 3.5098, Perplexity: 33.4414
Epoch [1/3], Step [137/3236], Loss: 3.4746, Perplexity: 32.2841
Epoch [1/3], Step [138/3236], Loss: 3.3975, Perplexity: 29.8886
Epoch [1/3], Step [139/3236], Loss: 3.5668, Perplexity: 35.4029
Epoch [1/3], Step [140/3236], Loss: 4.3918, Perplexity: 80.7872
Epoch [1/3], Step [141/3236], Loss: 3.3861, Perplexity: 29.5491
Epoch [1/3], Step [142/3236], Loss: 3.7855, Perplexity: 44.0574
Epoch [1/3], Step [143/3236], Loss: 3.5224, Perplexity: 33.8660
Epoch [1/3], Step [144/3236], Loss: 3.4769, Perplexity: 32.3588
Epoch [1/3], Step [145/3236], Loss: 3.4525, Perplexity: 31.5798
Epoch [1/3], Step [146/3236], Loss: 3.4971, Perplexity: 33.0191
Epoch [1/3], Step [147/3236], Loss: 3.3873, Perplexity: 29.5874
Epoch [1/3], Step [148/3236], Loss: 3.2972, Perplexity: 27.0376
Epoch [1/3], Step [149/3236], Loss: 3.4045, Perplexity: 30.0999
Epoch [1/3], Step [150/3236], Loss: 3.3088, Perplexity: 27.3520
Epoch [1/3], Step [151/3236], Loss: 4.2604, Perplexity: 70.8404
Epoch [1/3], Step [152/3236], Loss: 3.3812, Perplexity: 29.4067
Epoch [1/3], Step [153/3236], Loss: 3.3482, Perplexity: 28.4528
Epoch [1/3], Step [154/3236], Loss: 3.3064, Perplexity: 27.2868
Epoch [1/3], Step [155/3236], Loss: 3.4883, Perplexity: 32.7319
Epoch [1/3], Step [156/3236], Loss: 3.4992, Perplexity: 33.0876
Epoch [1/3], Step [157/3236], Loss: 3.3791, Perplexity: 29.3431
Epoch [1/3], Step [158/3236], Loss: 3.9048, Perplexity: 49.6393
Epoch [1/3], Step [159/3236], Loss: 3.4104, Perplexity: 30.2782
Epoch [1/3], Step [160/3236], Loss: 3.4439, Perplexity: 31.3083
Epoch [1/3], Step [161/3236], Loss: 3.3780, Perplexity: 29.3128
Epoch [1/3], Step [162/3236], Loss: 3.6665, Perplexity: 39.1156
Epoch [1/3], Step [163/3236], Loss: 4.1140, Perplexity: 61.1912
Epoch [1/3], Step [164/3236], Loss: 3.5583, Perplexity: 35.1052
Epoch [1/3], Step [165/3236], Loss: 4.2100, Perplexity: 67.3588
Epoch [1/3], Step [166/3236], Loss: 3.3606, Perplexity: 28.8076
Epoch [1/3], Step [167/3236], Loss: 3.3136, Perplexity: 27.4847
Epoch [1/3], Step [168/3236], Loss: 3.2885, Perplexity: 26.8035
Epoch [1/3], Step [169/3236], Loss: 3.1875, Perplexity: 24.2268
Epoch [1/3], Step [170/3236], Loss: 3.4357, Perplexity: 31.0528
Epoch [1/3], Step [171/3236], Loss: 3.1591, Perplexity: 23.5486
Epoch [1/3], Step [172/3236], Loss: 3.3397, Perplexity: 28.2117
Epoch [1/3], Step [173/3236], Loss: 3.2903, Perplexity: 26.8518
Epoch [1/3], Step [174/3236], Loss: 3.3378, Perplexity: 28.1580
Epoch [1/3], Step [175/3236], Loss: 3.3338, Perplexity: 28.0436
Epoch [1/3], Step [176/3236], Loss: 3.2755, Perplexity: 26.4557
Epoch [1/3], Step [177/3236], Loss: 3.3880, Perplexity: 29.6058
Epoch [1/3], Step [178/3236], Loss: 3.3959, Perplexity: 29.8407
Epoch [1/3], Step [179/3236], Loss: 3.7205, Perplexity: 41.2835
Epoch [1/3], Step [180/3236], Loss: 3.2629, Perplexity: 26.1259
Epoch [1/3], Step [181/3236], Loss: 3.3516, Perplexity: 28.5494
Epoch [1/3], Step [182/3236], Loss: 3.4118, Perplexity: 30.3199
Epoch [1/3], Step [183/3236], Loss: 3.2706, Perplexity: 26.3281
Epoch [1/3], Step [184/3236], Loss: 3.4328, Perplexity: 30.9625
Epoch [1/3], Step [185/3236], Loss: 3.4364, Perplexity: 31.0742
Epoch [1/3], Step [186/3236], Loss: 3.3656, Perplexity: 28.9509
Epoch [1/3], Step [187/3236], Loss: 3.2897, Perplexity: 26.8348
Epoch [1/3], Step [188/3236], Loss: 3.3451, Perplexity: 28.3646
Epoch [1/3], Step [189/3236], Loss: 3.3045, Perplexity: 27.2337
Epoch [1/3], Step [190/3236], Loss: 3.5584, Perplexity: 35.1079
Epoch [1/3], Step [191/3236], Loss: 3.4016, Perplexity: 30.0135
Epoch [1/3], Step [192/3236], Loss: 3.3696, Perplexity: 29.0679
Epoch [1/3], Step [193/3236], Loss: 3.4726, Perplexity: 32.2188
Epoch [1/3], Step [194/3236], Loss: 3.4698, Perplexity: 32.1314
Epoch [1/3], Step [195/3236], Loss: 3.8899, Perplexity: 48.9084
Epoch [1/3], Step [196/3236], Loss: 3.2921, Perplexity: 26.8996
Epoch [1/3], Step [197/3236], Loss: 3.2313, Perplexity: 25.3137
Epoch [1/3], Step [198/3236], Loss: 3.1702, Perplexity: 23.8133
Epoch [1/3], Step [199/3236], Loss: 3.2872, Perplexity: 26.7686
Epoch [1/3], Step [200/3236], Loss: 3.4223, Perplexity: 30.6390
Epoch [1/3], Step [201/3236], Loss: 3.1055, Perplexity: 22.3208
Epoch [1/3], Step [202/3236], Loss: 3.3125, Perplexity: 27.4546
Epoch [1/3], Step [203/3236], Loss: 3.6019, Perplexity: 36.6675
Epoch [1/3], Step [204/3236], Loss: 3.7654, Perplexity: 43.1803
Epoch [1/3], Step [205/3236], Loss: 3.3089, Perplexity: 27.3551
Epoch [1/3], Step [206/3236], Loss: 3.3180, Perplexity: 27.6043
Epoch [1/3], Step [207/3236], Loss: 3.2032, Perplexity: 24.6118
Epoch [1/3], Step [208/3236], Loss: 3.3494, Perplexity: 28.4844
Epoch [1/3], Step [209/3236], Loss: 3.2449, Perplexity: 25.6586
Epoch [1/3], Step [210/3236], Loss: 3.2040, Perplexity: 24.6309
Epoch [1/3], Step [211/3236], Loss: 3.0143, Perplexity: 20.3751
Epoch [1/3], Step [212/3236], Loss: 3.2350, Perplexity: 25.4075
Epoch [1/3], Step [213/3236], Loss: 3.2449, Perplexity: 25.6597
Epoch [1/3], Step [214/3236], Loss: 3.2603, Perplexity: 26.0585
Epoch [1/3], Step [215/3236], Loss: 3.2999, Perplexity: 27.1095
Epoch [1/3], Step [216/3236], Loss: 3.1772, Perplexity: 23.9802
Epoch [1/3], Step [217/3236], Loss: 3.2919, Perplexity: 26.8935
Epoch [1/3], Step [218/3236], Loss: 3.1516, Perplexity: 23.3736
Epoch [1/3], Step [219/3236], Loss: 3.2176, Perplexity: 24.9672
Epoch [1/3], Step [220/3236], Loss: 3.3398, Perplexity: 28.2146
Epoch [1/3], Step [221/3236], Loss: 3.2541, Perplexity: 25.8976
Epoch [1/3], Step [222/3236], Loss: 3.1593, Perplexity: 23.5542
Epoch [1/3], Step [223/3236], Loss: 3.3378, Perplexity: 28.1583
Epoch [1/3], Step [224/3236], Loss: 3.3465, Perplexity: 28.4039
Epoch [1/3], Step [225/3236], Loss: 3.1215, Perplexity: 22.6806
Epoch [1/3], Step [226/3236], Loss: 3.2810, Perplexity: 26.6013
Epoch [1/3], Step [227/3236], Loss: 3.2150, Perplexity: 24.9041
Epoch [1/3], Step [228/3236], Loss: 3.1048, Perplexity: 22.3051
Epoch [1/3], Step [229/3236], Loss: 3.3622, Perplexity: 28.8520
Epoch [1/3], Step [230/3236], Loss: 3.7003, Perplexity: 40.4599
Epoch [1/3], Step [231/3236], Loss: 3.3874, Perplexity: 29.5891
Epoch [1/3], Step [232/3236], Loss: 3.1836, Perplexity: 24.1346
Epoch [1/3], Step [233/3236], Loss: 3.5456, Perplexity: 34.6588
Epoch [1/3], Step [234/3236], Loss: 3.3345, Perplexity: 28.0644
Epoch [1/3], Step [235/3236], Loss: 3.3415, Perplexity: 28.2604
Epoch [1/3], Step [236/3236], Loss: 3.9304, Perplexity: 50.9277
Epoch [1/3], Step [237/3236], Loss: 3.2307, Perplexity: 25.2986
Epoch [1/3], Step [238/3236], Loss: 3.1415, Perplexity: 23.1386
Epoch [1/3], Step [239/3236], Loss: 3.3464, Perplexity: 28.4013
Epoch [1/3], Step [240/3236], Loss: 3.3045, Perplexity: 27.2347
Epoch [1/3], Step [241/3236], Loss: 3.4432, Perplexity: 31.2884
Epoch [1/3], Step [242/3236], Loss: 3.3766, Perplexity: 29.2715
Epoch [1/3], Step [243/3236], Loss: 3.3236, Perplexity: 27.7611
Epoch [1/3], Step [244/3236], Loss: 3.2664, Perplexity: 26.2174
Epoch [1/3], Step [245/3236], Loss: 3.1576, Perplexity: 23.5135
Epoch [1/3], Step [246/3236], Loss: 3.3996, Perplexity: 29.9521
Epoch [1/3], Step [247/3236], Loss: 3.3228, Perplexity: 27.7365
Epoch [1/3], Step [248/3236], Loss: 3.3144, Perplexity: 27.5050
Epoch [1/3], Step [249/3236], Loss: 3.2473, Perplexity: 25.7206
Epoch [1/3], Step [250/3236], Loss: 3.4387, Perplexity: 31.1461
Epoch [1/3], Step [251/3236], Loss: 3.2201, Perplexity: 25.0296
Epoch [1/3], Step [252/3236], Loss: 3.3941, Perplexity: 29.7871
Epoch [1/3], Step [253/3236], Loss: 3.3466, Perplexity: 28.4048
Epoch [1/3], Step [254/3236], Loss: 3.2583, Perplexity: 26.0048
Epoch [1/3], Step [255/3236], Loss: 3.3008, Perplexity: 27.1332
Epoch [1/3], Step [256/3236], Loss: 3.2708, Perplexity: 26.3314
Epoch [1/3], Step [257/3236], Loss: 3.2883, Perplexity: 26.7966
Epoch [1/3], Step [258/3236], Loss: 3.2280, Perplexity: 25.2290
Epoch [1/3], Step [259/3236], Loss: 3.1932, Perplexity: 24.3660
Epoch [1/3], Step [260/3236], Loss: 3.9703, Perplexity: 53.0010
Epoch [1/3], Step [261/3236], Loss: 3.5272, Perplexity: 34.0302
Epoch [1/3], Step [262/3236], Loss: 3.2423, Perplexity: 25.5921
Epoch [1/3], Step [263/3236], Loss: 3.1567, Perplexity: 23.4938
Epoch [1/3], Step [264/3236], Loss: 3.3118, Perplexity: 27.4354
Epoch [1/3], Step [265/3236], Loss: 3.2465, Perplexity: 25.7000
Epoch [1/3], Step [266/3236], Loss: 3.2187, Perplexity: 24.9966
Epoch [1/3], Step [267/3236], Loss: 3.0987, Perplexity: 22.1696
Epoch [1/3], Step [268/3236], Loss: 3.1849, Perplexity: 24.1637
Epoch [1/3], Step [269/3236], Loss: 3.0904, Perplexity: 21.9855
Epoch [1/3], Step [270/3236], Loss: 3.1997, Perplexity: 24.5260
Epoch [1/3], Step [271/3236], Loss: 3.0549, Perplexity: 21.2183
Epoch [1/3], Step [272/3236], Loss: 3.2731, Perplexity: 26.3917
Epoch [1/3], Step [273/3236], Loss: 3.3766, Perplexity: 29.2713
Epoch [1/3], Step [274/3236], Loss: 3.1429, Perplexity: 23.1699
Epoch [1/3], Step [275/3236], Loss: 3.2328, Perplexity: 25.3500
Epoch [1/3], Step [276/3236], Loss: 3.3017, Perplexity: 27.1583
Epoch [1/3], Step [277/3236], Loss: 3.0633, Perplexity: 21.3985
Epoch [1/3], Step [278/3236], Loss: 3.1024, Perplexity: 22.2510
Epoch [1/3], Step [279/3236], Loss: 3.4813, Perplexity: 32.5011
Epoch [1/3], Step [280/3236], Loss: 3.1071, Perplexity: 22.3556
Epoch [1/3], Step [281/3236], Loss: 3.6083, Perplexity: 36.9024
Epoch [1/3], Step [282/3236], Loss: 3.2219, Perplexity: 25.0748
Epoch [1/3], Step [283/3236], Loss: 2.9989, Perplexity: 20.0632
Epoch [1/3], Step [284/3236], Loss: 3.1311, Perplexity: 22.8985
Epoch [1/3], Step [285/3236], Loss: 3.2457, Perplexity: 25.6795
Epoch [1/3], Step [286/3236], Loss: 3.1413, Perplexity: 23.1347
Epoch [1/3], Step [287/3236], Loss: 3.1130, Perplexity: 22.4878
Epoch [1/3], Step [288/3236], Loss: 3.0904, Perplexity: 21.9866
Epoch [1/3], Step [289/3236], Loss: 3.2094, Perplexity: 24.7649
Epoch [1/3], Step [290/3236], Loss: 3.8441, Perplexity: 46.7157
Epoch [1/3], Step [291/3236], Loss: 3.0945, Perplexity: 22.0770
Epoch [1/3], Step [292/3236], Loss: 3.5393, Perplexity: 34.4427
Epoch [1/3], Step [293/3236], Loss: 3.1406, Perplexity: 23.1170
Epoch [1/3], Step [294/3236], Loss: 3.1708, Perplexity: 23.8273
Epoch [1/3], Step [295/3236], Loss: 3.8164, Perplexity: 45.4388
Epoch [1/3], Step [296/3236], Loss: 3.5390, Perplexity: 34.4320
Epoch [1/3], Step [297/3236], Loss: 3.1443, Perplexity: 23.2039
Epoch [1/3], Step [298/3236], Loss: 3.1649, Perplexity: 23.6855
Epoch [1/3], Step [299/3236], Loss: 3.5703, Perplexity: 35.5279
Epoch [1/3], Step [300/3236], Loss: 3.2764, Perplexity: 26.4809
Epoch [1/3], Step [301/3236], Loss: 3.2912, Perplexity: 26.8753
Epoch [1/3], Step [302/3236], Loss: 3.1076, Perplexity: 22.3668
Epoch [1/3], Step [303/3236], Loss: 2.9906, Perplexity: 19.8976
Epoch [1/3], Step [304/3236], Loss: 3.1813, Perplexity: 24.0787
Epoch [1/3], Step [305/3236], Loss: 3.1813, Perplexity: 24.0791
Epoch [1/3], Step [306/3236], Loss: 3.2225, Perplexity: 25.0904
Epoch [1/3], Step [307/3236], Loss: 3.2136, Perplexity: 24.8680
Epoch [1/3], Step [308/3236], Loss: 3.0412, Perplexity: 20.9307
Epoch [1/3], Step [309/3236], Loss: 2.9787, Perplexity: 19.6632
Epoch [1/3], Step [310/3236], Loss: 3.1473, Perplexity: 23.2722
Epoch [1/3], Step [311/3236], Loss: 3.1377, Perplexity: 23.0497
Epoch [1/3], Step [312/3236], Loss: 3.2325, Perplexity: 25.3435
Epoch [1/3], Step [313/3236], Loss: 3.0006, Perplexity: 20.0984
Epoch [1/3], Step [314/3236], Loss: 3.0986, Perplexity: 22.1659
Epoch [1/3], Step [315/3236], Loss: 2.9621, Perplexity: 19.3379
Epoch [1/3], Step [316/3236], Loss: 3.0701, Perplexity: 21.5439
Epoch [1/3], Step [317/3236], Loss: 3.1670, Perplexity: 23.7372
Epoch [1/3], Step [318/3236], Loss: 2.9591, Perplexity: 19.2800
Epoch [1/3], Step [319/3236], Loss: 3.1697, Perplexity: 23.8004
Epoch [1/3], Step [320/3236], Loss: 3.1488, Perplexity: 23.3085
Epoch [1/3], Step [321/3236], Loss: 3.0396, Perplexity: 20.8964
Epoch [1/3], Step [322/3236], Loss: 2.9924, Perplexity: 19.9332
Epoch [1/3], Step [323/3236], Loss: 3.1533, Perplexity: 23.4135
Epoch [1/3], Step [324/3236], Loss: 3.0720, Perplexity: 21.5845
Epoch [1/3], Step [325/3236], Loss: 3.1162, Perplexity: 22.5615
Epoch [1/3], Step [326/3236], Loss: 3.1142, Perplexity: 22.5149
Epoch [1/3], Step [327/3236], Loss: 3.0932, Perplexity: 22.0467
Epoch [1/3], Step [328/3236], Loss: 2.9953, Perplexity: 19.9917
Epoch [1/3], Step [329/3236], Loss: 3.2726, Perplexity: 26.3806
Epoch [1/3], Step [330/3236], Loss: 2.9881, Perplexity: 19.8480
Epoch [1/3], Step [331/3236], Loss: 3.1748, Perplexity: 23.9218
Epoch [1/3], Step [332/3236], Loss: 3.1093, Perplexity: 22.4050
Epoch [1/3], Step [333/3236], Loss: 2.8848, Perplexity: 17.8993
Epoch [1/3], Step [334/3236], Loss: 3.3301, Perplexity: 27.9411
Epoch [1/3], Step [335/3236], Loss: 3.1873, Perplexity: 24.2238
Epoch [1/3], Step [336/3236], Loss: 3.1346, Perplexity: 22.9791
Epoch [1/3], Step [337/3236], Loss: 3.0102, Perplexity: 20.2917
Epoch [1/3], Step [338/3236], Loss: 3.0564, Perplexity: 21.2510
Epoch [1/3], Step [339/3236], Loss: 3.2870, Perplexity: 26.7623
Epoch [1/3], Step [340/3236], Loss: 3.1972, Perplexity: 24.4640
Epoch [1/3], Step [341/3236], Loss: 3.0542, Perplexity: 21.2051
Epoch [1/3], Step [342/3236], Loss: 3.1512, Perplexity: 23.3638
Epoch [1/3], Step [343/3236], Loss: 3.0635, Perplexity: 21.4023
Epoch [1/3], Step [344/3236], Loss: 3.0224, Perplexity: 20.5402
Epoch [1/3], Step [345/3236], Loss: 3.0837, Perplexity: 21.8399
Epoch [1/3], Step [346/3236], Loss: 3.2817, Perplexity: 26.6220
Epoch [1/3], Step [347/3236], Loss: 3.1689, Perplexity: 23.7813
Epoch [1/3], Step [348/3236], Loss: 3.5251, Perplexity: 33.9578
Epoch [1/3], Step [349/3236], Loss: 2.9760, Perplexity: 19.6102
Epoch [1/3], Step [350/3236], Loss: 3.4850, Perplexity: 32.6209
Epoch [1/3], Step [351/3236], Loss: 3.0942, Perplexity: 22.0704
Epoch [1/3], Step [352/3236], Loss: 2.9755, Perplexity: 19.5996
Epoch [1/3], Step [353/3236], Loss: 3.1597, Perplexity: 23.5646
Epoch [1/3], Step [354/3236], Loss: 2.9332, Perplexity: 18.7885
Epoch [1/3], Step [355/3236], Loss: 3.0455, Perplexity: 21.0198
Epoch [1/3], Step [356/3236], Loss: 3.0320, Perplexity: 20.7394
Epoch [1/3], Step [357/3236], Loss: 2.8812, Perplexity: 17.8355
Epoch [1/3], Step [358/3236], Loss: 3.1689, Perplexity: 23.7809
Epoch [1/3], Step [359/3236], Loss: 3.0649, Perplexity: 21.4313
Epoch [1/3], Step [360/3236], Loss: 3.0742, Perplexity: 21.6316
Epoch [1/3], Step [361/3236], Loss: 3.1322, Perplexity: 22.9239
Epoch [1/3], Step [362/3236], Loss: 2.8585, Perplexity: 17.4346
Epoch [1/3], Step [363/3236], Loss: 2.9946, Perplexity: 19.9774
Epoch [1/3], Step [364/3236], Loss: 2.9844, Perplexity: 19.7749
Epoch [1/3], Step [365/3236], Loss: 2.9412, Perplexity: 18.9392
Epoch [1/3], Step [366/3236], Loss: 3.0697, Perplexity: 21.5355
Epoch [1/3], Step [367/3236], Loss: 3.3203, Perplexity: 27.6686
Epoch [1/3], Step [368/3236], Loss: 2.9626, Perplexity: 19.3485
Epoch [1/3], Step [369/3236], Loss: 2.9307, Perplexity: 18.7411
Epoch [1/3], Step [370/3236], Loss: 2.8831, Perplexity: 17.8693
Epoch [1/3], Step [371/3236], Loss: 3.4218, Perplexity: 30.6244
Epoch [1/3], Step [372/3236], Loss: 2.9633, Perplexity: 19.3620
Epoch [1/3], Step [373/3236], Loss: 3.0547, Perplexity: 21.2138
Epoch [1/3], Step [374/3236], Loss: 3.0307, Perplexity: 20.7116
Epoch [1/3], Step [375/3236], Loss: 2.9464, Perplexity: 19.0366
Epoch [1/3], Step [376/3236], Loss: 3.5227, Perplexity: 33.8762
Epoch [1/3], Step [377/3236], Loss: 3.1139, Perplexity: 22.5092
Epoch [1/3], Step [378/3236], Loss: 3.1473, Perplexity: 23.2726
Epoch [1/3], Step [379/3236], Loss: 3.0862, Perplexity: 21.8945
Epoch [1/3], Step [380/3236], Loss: 3.0543, Perplexity: 21.2060
Epoch [1/3], Step [381/3236], Loss: 3.2411, Perplexity: 25.5617
Epoch [1/3], Step [382/3236], Loss: 3.1065, Perplexity: 22.3421
Epoch [1/3], Step [383/3236], Loss: 3.3203, Perplexity: 27.6692
Epoch [1/3], Step [384/3236], Loss: 2.9956, Perplexity: 19.9967
Epoch [1/3], Step [385/3236], Loss: 3.0101, Perplexity: 20.2895
Epoch [1/3], Step [386/3236], Loss: 3.2336, Perplexity: 25.3705
Epoch [1/3], Step [387/3236], Loss: 2.8710, Perplexity: 17.6547
Epoch [1/3], Step [388/3236], Loss: 3.8961, Perplexity: 49.2091
Epoch [1/3], Step [389/3236], Loss: 2.9318, Perplexity: 18.7609
Epoch [1/3], Step [390/3236], Loss: 2.8934, Perplexity: 18.0547
Epoch [1/3], Step [391/3236], Loss: 2.8675, Perplexity: 17.5928
Epoch [1/3], Step [392/3236], Loss: 3.0710, Perplexity: 21.5638
Epoch [1/3], Step [393/3236], Loss: 3.2544, Perplexity: 25.9029
Epoch [1/3], Step [394/3236], Loss: 2.9994, Perplexity: 20.0735
Epoch [1/3], Step [395/3236], Loss: 3.0988, Perplexity: 22.1713
Epoch [1/3], Step [396/3236], Loss: 2.9520, Perplexity: 19.1447
Epoch [1/3], Step [397/3236], Loss: 3.3476, Perplexity: 28.4348
Epoch [1/3], Step [398/3236], Loss: 3.4963, Perplexity: 32.9926
Epoch [1/3], Step [399/3236], Loss: 2.9308, Perplexity: 18.7417
Epoch [1/3], Step [400/3236], Loss: 3.0443, Perplexity: 20.9956
Epoch [1/3], Step [401/3236], Loss: 2.9871, Perplexity: 19.8276
Epoch [1/3], Step [402/3236], Loss: 3.3739, Perplexity: 29.1914
Epoch [1/3], Step [403/3236], Loss: 3.0431, Perplexity: 20.9703
Epoch [1/3], Step [404/3236], Loss: 3.0511, Perplexity: 21.1392
Epoch [1/3], Step [405/3236], Loss: 3.3794, Perplexity: 29.3524
Epoch [1/3], Step [406/3236], Loss: 3.1211, Perplexity: 22.6720
Epoch [1/3], Step [407/3236], Loss: 2.9331, Perplexity: 18.7864
Epoch [1/3], Step [408/3236], Loss: 2.9510, Perplexity: 19.1244
Epoch [1/3], Step [409/3236], Loss: 3.1346, Perplexity: 22.9798
Epoch [1/3], Step [410/3236], Loss: 2.8599, Perplexity: 17.4598
Epoch [1/3], Step [411/3236], Loss: 3.1860, Perplexity: 24.1925
Epoch [1/3], Step [412/3236], Loss: 3.0992, Perplexity: 22.1792
Epoch [1/3], Step [413/3236], Loss: 3.1057, Perplexity: 22.3240
Epoch [1/3], Step [414/3236], Loss: 2.9839, Perplexity: 19.7656
Epoch [1/3], Step [415/3236], Loss: 2.9370, Perplexity: 18.8593
Epoch [1/3], Step [416/3236], Loss: 2.8320, Perplexity: 16.9793
Epoch [1/3], Step [417/3236], Loss: 3.4646, Perplexity: 31.9621
Epoch [1/3], Step [418/3236], Loss: 3.3896, Perplexity: 29.6527
Epoch [1/3], Step [419/3236], Loss: 2.9985, Perplexity: 20.0561
Epoch [1/3], Step [420/3236], Loss: 3.1398, Perplexity: 23.0996
Epoch [1/3], Step [421/3236], Loss: 3.2082, Perplexity: 24.7343
Epoch [1/3], Step [422/3236], Loss: 3.8626, Perplexity: 47.5888
Epoch [1/3], Step [423/3236], Loss: 3.4046, Perplexity: 30.1036
Epoch [1/3], Step [424/3236], Loss: 3.0389, Perplexity: 20.8822
Epoch [1/3], Step [425/3236], Loss: 3.0893, Perplexity: 21.9616
Epoch [1/3], Step [426/3236], Loss: 2.9920, Perplexity: 19.9254
Epoch [1/3], Step [427/3236], Loss: 3.2103, Perplexity: 24.7858
Epoch [1/3], Step [428/3236], Loss: 2.8981, Perplexity: 18.1395
Epoch [1/3], Step [429/3236], Loss: 3.0434, Perplexity: 20.9772
Epoch [1/3], Step [430/3236], Loss: 3.0572, Perplexity: 21.2688
Epoch [1/3], Step [431/3236], Loss: 3.0579, Perplexity: 21.2825
Epoch [1/3], Step [432/3236], Loss: 3.0564, Perplexity: 21.2515
Epoch [1/3], Step [433/3236], Loss: 2.9007, Perplexity: 18.1866
Epoch [1/3], Step [434/3236], Loss: 3.2023, Perplexity: 24.5880
Epoch [1/3], Step [435/3236], Loss: 2.9037, Perplexity: 18.2414
Epoch [1/3], Step [436/3236], Loss: 2.9510, Perplexity: 19.1247
Epoch [1/3], Step [437/3236], Loss: 2.9488, Perplexity: 19.0833
Epoch [1/3], Step [438/3236], Loss: 3.0372, Perplexity: 20.8474
Epoch [1/3], Step [439/3236], Loss: 3.0674, Perplexity: 21.4861
Epoch [1/3], Step [440/3236], Loss: 2.9959, Perplexity: 20.0035
Epoch [1/3], Step [441/3236], Loss: 2.9045, Perplexity: 18.2568
Epoch [1/3], Step [442/3236], Loss: 3.0178, Perplexity: 20.4467
Epoch [1/3], Step [443/3236], Loss: 3.0337, Perplexity: 20.7735
Epoch [1/3], Step [444/3236], Loss: 3.0114, Perplexity: 20.3167
Epoch [1/3], Step [445/3236], Loss: 3.1747, Perplexity: 23.9200
Epoch [1/3], Step [446/3236], Loss: 3.1004, Perplexity: 22.2067
Epoch [1/3], Step [447/3236], Loss: 3.0531, Perplexity: 21.1804
Epoch [1/3], Step [448/3236], Loss: 2.9120, Perplexity: 18.3927
Epoch [1/3], Step [449/3236], Loss: 3.1498, Perplexity: 23.3305
Epoch [1/3], Step [450/3236], Loss: 3.0114, Perplexity: 20.3154
Epoch [1/3], Step [451/3236], Loss: 2.8467, Perplexity: 17.2300
Epoch [1/3], Step [452/3236], Loss: 3.1262, Perplexity: 22.7862
Epoch [1/3], Step [453/3236], Loss: 2.9590, Perplexity: 19.2784
Epoch [1/3], Step [454/3236], Loss: 3.1081, Perplexity: 22.3793
Epoch [1/3], Step [455/3236], Loss: 2.8719, Perplexity: 17.6699
Epoch [1/3], Step [456/3236], Loss: 3.1344, Perplexity: 22.9750
Epoch [1/3], Step [457/3236], Loss: 2.9571, Perplexity: 19.2415
Epoch [1/3], Step [458/3236], Loss: 2.9159, Perplexity: 18.4645
Epoch [1/3], Step [459/3236], Loss: 3.0293, Perplexity: 20.6828
Epoch [1/3], Step [460/3236], Loss: 3.0851, Perplexity: 21.8702
Epoch [1/3], Step [461/3236], Loss: 2.9642, Perplexity: 19.3802
Epoch [1/3], Step [462/3236], Loss: 2.9775, Perplexity: 19.6391
Epoch [1/3], Step [463/3236], Loss: 3.2369, Perplexity: 25.4553
Epoch [1/3], Step [464/3236], Loss: 3.1300, Perplexity: 22.8750
Epoch [1/3], Step [465/3236], Loss: 2.8322, Perplexity: 16.9829
Epoch [1/3], Step [466/3236], Loss: 3.0060, Perplexity: 20.2071
Epoch [1/3], Step [467/3236], Loss: 3.3468, Perplexity: 28.4115
Epoch [1/3], Step [468/3236], Loss: 2.9134, Perplexity: 18.4193
Epoch [1/3], Step [469/3236], Loss: 2.8210, Perplexity: 16.7938
Epoch [1/3], Step [470/3236], Loss: 2.9268, Perplexity: 18.6681
Epoch [1/3], Step [471/3236], Loss: 3.2561, Perplexity: 25.9480
Epoch [1/3], Step [472/3236], Loss: 3.0444, Perplexity: 20.9972
Epoch [1/3], Step [473/3236], Loss: 2.8187, Perplexity: 16.7557
Epoch [1/3], Step [474/3236], Loss: 2.9110, Perplexity: 18.3760
Epoch [1/3], Step [475/3236], Loss: 2.9818, Perplexity: 19.7228
Epoch [1/3], Step [476/3236], Loss: 3.5475, Perplexity: 34.7267
Epoch [1/3], Step [477/3236], Loss: 2.8662, Perplexity: 17.5700
Epoch [1/3], Step [478/3236], Loss: 2.9547, Perplexity: 19.1964
Epoch [1/3], Step [479/3236], Loss: 3.0381, Perplexity: 20.8648
Epoch [1/3], Step [480/3236], Loss: 3.1387, Perplexity: 23.0743
Epoch [1/3], Step [481/3236], Loss: 2.7677, Perplexity: 15.9220
Epoch [1/3], Step [482/3236], Loss: 2.8689, Perplexity: 17.6170
Epoch [1/3], Step [483/3236], Loss: 3.0468, Perplexity: 21.0482
Epoch [1/3], Step [484/3236], Loss: 3.4327, Perplexity: 30.9605
Epoch [1/3], Step [485/3236], Loss: 3.0131, Perplexity: 20.3506
Epoch [1/3], Step [486/3236], Loss: 2.8168, Perplexity: 16.7229
Epoch [1/3], Step [487/3236], Loss: 2.8695, Perplexity: 17.6277
Epoch [1/3], Step [488/3236], Loss: 2.9272, Perplexity: 18.6752
Epoch [1/3], Step [489/3236], Loss: 3.4536, Perplexity: 31.6146
Epoch [1/3], Step [490/3236], Loss: 2.9771, Perplexity: 19.6309
Epoch [1/3], Step [491/3236], Loss: 2.8003, Perplexity: 16.4499
Epoch [1/3], Step [492/3236], Loss: 2.8922, Perplexity: 18.0322
Epoch [1/3], Step [493/3236], Loss: 3.0223, Perplexity: 20.5389
Epoch [1/3], Step [494/3236], Loss: 2.8638, Perplexity: 17.5278
Epoch [1/3], Step [495/3236], Loss: 2.8123, Perplexity: 16.6479
Epoch [1/3], Step [496/3236], Loss: 2.9271, Perplexity: 18.6737
Epoch [1/3], Step [497/3236], Loss: 2.8519, Perplexity: 17.3211
Epoch [1/3], Step [498/3236], Loss: 2.8306, Perplexity: 16.9551
Epoch [1/3], Step [499/3236], Loss: 2.9780, Perplexity: 19.6486
Epoch [1/3], Step [500/3236], Loss: 2.8287, Perplexity: 16.9232
Epoch [1/3], Step [501/3236], Loss: 2.9304, Perplexity: 18.7360
Epoch [1/3], Step [502/3236], Loss: 2.8425, Perplexity: 17.1590
Epoch [1/3], Step [503/3236], Loss: 2.8846, Perplexity: 17.8964
Epoch [1/3], Step [504/3236], Loss: 3.3105, Perplexity: 27.4002
Epoch [1/3], Step [505/3236], Loss: 2.9525, Perplexity: 19.1541
Epoch [1/3], Step [506/3236], Loss: 2.8147, Perplexity: 16.6880
Epoch [1/3], Step [507/3236], Loss: 3.0603, Perplexity: 21.3344
Epoch [1/3], Step [508/3236], Loss: 2.7242, Perplexity: 15.2443
Epoch [1/3], Step [509/3236], Loss: 2.7960, Perplexity: 16.3784
Epoch [1/3], Step [510/3236], Loss: 2.9905, Perplexity: 19.8952
Epoch [1/3], Step [511/3236], Loss: 2.8507, Perplexity: 17.3002
Epoch [1/3], Step [512/3236], Loss: 3.3192, Perplexity: 27.6387
Epoch [1/3], Step [513/3236], Loss: 2.7612, Perplexity: 15.8192
Epoch [1/3], Step [514/3236], Loss: 3.2937, Perplexity: 26.9432
Epoch [1/3], Step [515/3236], Loss: 2.8555, Perplexity: 17.3827
Epoch [1/3], Step [516/3236], Loss: 2.8817, Perplexity: 17.8448
Epoch [1/3], Step [517/3236], Loss: 3.3113, Perplexity: 27.4216
Epoch [1/3], Step [518/3236], Loss: 4.3065, Perplexity: 74.1826
Epoch [1/3], Step [519/3236], Loss: 3.0722, Perplexity: 21.5899
Epoch [1/3], Step [520/3236], Loss: 3.1217, Perplexity: 22.6853
Epoch [1/3], Step [521/3236], Loss: 3.2416, Perplexity: 25.5754
Epoch [1/3], Step [522/3236], Loss: 2.9476, Perplexity: 19.0608
Epoch [1/3], Step [523/3236], Loss: 3.0324, Perplexity: 20.7469
Epoch [1/3], Step [524/3236], Loss: 2.9748, Perplexity: 19.5853
Epoch [1/3], Step [525/3236], Loss: 3.5193, Perplexity: 33.7621
Epoch [1/3], Step [526/3236], Loss: 3.0253, Perplexity: 20.5996
Epoch [1/3], Step [527/3236], Loss: 2.9525, Perplexity: 19.1539
Epoch [1/3], Step [528/3236], Loss: 2.9765, Perplexity: 19.6194
Epoch [1/3], Step [529/3236], Loss: 2.7893, Perplexity: 16.2700
Epoch [1/3], Step [530/3236], Loss: 2.9548, Perplexity: 19.1973
Epoch [1/3], Step [531/3236], Loss: 2.9512, Perplexity: 19.1292
Epoch [1/3], Step [532/3236], Loss: 3.0051, Perplexity: 20.1888
Epoch [1/3], Step [533/3236], Loss: 3.1057, Perplexity: 22.3249
Epoch [1/3], Step [534/3236], Loss: 2.9622, Perplexity: 19.3404
Epoch [1/3], Step [535/3236], Loss: 2.8231, Perplexity: 16.8295
Epoch [1/3], Step [536/3236], Loss: 2.7964, Perplexity: 16.3852
Epoch [1/3], Step [537/3236], Loss: 3.1147, Perplexity: 22.5262
Epoch [1/3], Step [538/3236], Loss: 2.9378, Perplexity: 18.8743
Epoch [1/3], Step [539/3236], Loss: 2.9112, Perplexity: 18.3792
Epoch [1/3], Step [540/3236], Loss: 3.2297, Perplexity: 25.2732
Epoch [1/3], Step [541/3236], Loss: 2.9153, Perplexity: 18.4551
Epoch [1/3], Step [542/3236], Loss: 3.0911, Perplexity: 22.0012
Epoch [1/3], Step [543/3236], Loss: 2.8925, Perplexity: 18.0379
Epoch [1/3], Step [544/3236], Loss: 2.9956, Perplexity: 19.9977
Epoch [1/3], Step [545/3236], Loss: 2.8728, Perplexity: 17.6863
Epoch [1/3], Step [546/3236], Loss: 2.8522, Perplexity: 17.3252
Epoch [1/3], Step [547/3236], Loss: 3.0382, Perplexity: 20.8682
Epoch [1/3], Step [548/3236], Loss: 2.9694, Perplexity: 19.4793
Epoch [1/3], Step [549/3236], Loss: 2.9270, Perplexity: 18.6724
Epoch [1/3], Step [550/3236], Loss: 2.9986, Perplexity: 20.0579
Epoch [1/3], Step [551/3236], Loss: 2.8811, Perplexity: 17.8347
Epoch [1/3], Step [552/3236], Loss: 2.8646, Perplexity: 17.5423
Epoch [1/3], Step [553/3236], Loss: 2.7844, Perplexity: 16.1905
Epoch [1/3], Step [554/3236], Loss: 2.8974, Perplexity: 18.1275
Epoch [1/3], Step [555/3236], Loss: 2.8419, Perplexity: 17.1475
Epoch [1/3], Step [556/3236], Loss: 3.2892, Perplexity: 26.8220
Epoch [1/3], Step [557/3236], Loss: 3.6706, Perplexity: 39.2740
Epoch [1/3], Step [558/3236], Loss: 2.8049, Perplexity: 16.5257
Epoch [1/3], Step [559/3236], Loss: 3.1494, Perplexity: 23.3221
Epoch [1/3], Step [560/3236], Loss: 3.0870, Perplexity: 21.9117
Epoch [1/3], Step [561/3236], Loss: 2.8705, Perplexity: 17.6462
Epoch [1/3], Step [562/3236], Loss: 2.7443, Perplexity: 15.5539
Epoch [1/3], Step [563/3236], Loss: 3.2814, Perplexity: 26.6131
Epoch [1/3], Step [564/3236], Loss: 3.1198, Perplexity: 22.6412
Epoch [1/3], Step [565/3236], Loss: 2.7790, Perplexity: 16.1031
Epoch [1/3], Step [566/3236], Loss: 2.8047, Perplexity: 16.5213
Epoch [1/3], Step [567/3236], Loss: 2.8375, Perplexity: 17.0722
Epoch [1/3], Step [568/3236], Loss: 3.0936, Perplexity: 22.0571
Epoch [1/3], Step [569/3236], Loss: 3.0188, Perplexity: 20.4676
Epoch [1/3], Step [570/3236], Loss: 2.8727, Perplexity: 17.6846
Epoch [1/3], Step [571/3236], Loss: 2.9537, Perplexity: 19.1768
Epoch [1/3], Step [572/3236], Loss: 3.0489, Perplexity: 21.0925
Epoch [1/3], Step [573/3236], Loss: 3.3651, Perplexity: 28.9361
Epoch [1/3], Step [574/3236], Loss: 3.0593, Perplexity: 21.3119
Epoch [1/3], Step [575/3236], Loss: 2.7313, Perplexity: 15.3529
Epoch [1/3], Step [576/3236], Loss: 2.7645, Perplexity: 15.8715
Epoch [1/3], Step [577/3236], Loss: 2.9123, Perplexity: 18.3982
Epoch [1/3], Step [578/3236], Loss: 2.8867, Perplexity: 17.9343
Epoch [1/3], Step [579/3236], Loss: 2.9860, Perplexity: 19.8066
Epoch [1/3], Step [580/3236], Loss: 2.8854, Perplexity: 17.9100
Epoch [1/3], Step [581/3236], Loss: 2.8482, Perplexity: 17.2565
Epoch [1/3], Step [582/3236], Loss: 2.7207, Perplexity: 15.1912
Epoch [1/3], Step [583/3236], Loss: 2.7995, Perplexity: 16.4367
Epoch [1/3], Step [584/3236], Loss: 2.7895, Perplexity: 16.2736
Epoch [1/3], Step [585/3236], Loss: 3.0042, Perplexity: 20.1710
Epoch [1/3], Step [586/3236], Loss: 2.7562, Perplexity: 15.7392
Epoch [1/3], Step [587/3236], Loss: 2.7950, Perplexity: 16.3624
Epoch [1/3], Step [588/3236], Loss: 2.8925, Perplexity: 18.0382
Epoch [1/3], Step [589/3236], Loss: 2.8324, Perplexity: 16.9859
Epoch [1/3], Step [590/3236], Loss: 2.9128, Perplexity: 18.4091
Epoch [1/3], Step [591/3236], Loss: 2.7101, Perplexity: 15.0302
Epoch [1/3], Step [592/3236], Loss: 2.7953, Perplexity: 16.3680
Epoch [1/3], Step [593/3236], Loss: 2.8429, Perplexity: 17.1653
Epoch [1/3], Step [594/3236], Loss: 3.3970, Perplexity: 29.8744
Epoch [1/3], Step [595/3236], Loss: 2.7281, Perplexity: 15.3043
Epoch [1/3], Step [596/3236], Loss: 3.3569, Perplexity: 28.7000
Epoch [1/3], Step [597/3236], Loss: 2.9811, Perplexity: 19.7101
Epoch [1/3], Step [598/3236], Loss: 2.9006, Perplexity: 18.1845
Epoch [1/3], Step [599/3236], Loss: 2.9085, Perplexity: 18.3301
Epoch [1/3], Step [600/3236], Loss: 2.7373, Perplexity: 15.4450
Epoch [1/3], Step [601/3236], Loss: 2.6708, Perplexity: 14.4521
Epoch [1/3], Step [602/3236], Loss: 2.8471, Perplexity: 17.2380
Epoch [1/3], Step [603/3236], Loss: 3.0770, Perplexity: 21.6929
Epoch [1/3], Step [604/3236], Loss: 3.2885, Perplexity: 26.8030
Epoch [1/3], Step [605/3236], Loss: 3.5162, Perplexity: 33.6578
Epoch [1/3], Step [606/3236], Loss: 2.8565, Perplexity: 17.4002
Epoch [1/3], Step [607/3236], Loss: 2.9456, Perplexity: 19.0211
Epoch [1/3], Step [608/3236], Loss: 2.7825, Perplexity: 16.1591
Epoch [1/3], Step [609/3236], Loss: 3.0984, Perplexity: 22.1629
Epoch [1/3], Step [610/3236], Loss: 2.9437, Perplexity: 18.9856
Epoch [1/3], Step [611/3236], Loss: 3.2593, Perplexity: 26.0322
Epoch [1/3], Step [612/3236], Loss: 3.8103, Perplexity: 45.1646
Epoch [1/3], Step [613/3236], Loss: 2.9777, Perplexity: 19.6422
Epoch [1/3], Step [614/3236], Loss: 2.8656, Perplexity: 17.5588
Epoch [1/3], Step [615/3236], Loss: 2.9005, Perplexity: 18.1839
Epoch [1/3], Step [616/3236], Loss: 2.8598, Perplexity: 17.4580
Epoch [1/3], Step [617/3236], Loss: 2.8849, Perplexity: 17.9024
Epoch [1/3], Step [618/3236], Loss: 2.9543, Perplexity: 19.1882
Epoch [1/3], Step [619/3236], Loss: 2.9179, Perplexity: 18.5032
Epoch [1/3], Step [620/3236], Loss: 2.9198, Perplexity: 18.5385
Epoch [1/3], Step [621/3236], Loss: 2.7170, Perplexity: 15.1352
Epoch [1/3], Step [622/3236], Loss: 2.7973, Perplexity: 16.4007
Epoch [1/3], Step [623/3236], Loss: 2.6519, Perplexity: 14.1816
Epoch [1/3], Step [624/3236], Loss: 2.8207, Perplexity: 16.7890
Epoch [1/3], Step [625/3236], Loss: 2.8475, Perplexity: 17.2443
Epoch [1/3], Step [626/3236], Loss: 2.8788, Perplexity: 17.7922
Epoch [1/3], Step [627/3236], Loss: 2.6598, Perplexity: 14.2939
Epoch [1/3], Step [628/3236], Loss: 2.8253, Perplexity: 16.8652
Epoch [1/3], Step [629/3236], Loss: 2.8561, Perplexity: 17.3928
Epoch [1/3], Step [630/3236], Loss: 2.7246, Perplexity: 15.2504
Epoch [1/3], Step [631/3236], Loss: 2.8289, Perplexity: 16.9270
Epoch [1/3], Step [632/3236], Loss: 2.8856, Perplexity: 17.9145
Epoch [1/3], Step [633/3236], Loss: 2.9114, Perplexity: 18.3831
Epoch [1/3], Step [634/3236], Loss: 2.7697, Perplexity: 15.9541
Epoch [1/3], Step [635/3236], Loss: 2.9387, Perplexity: 18.8912
Epoch [1/3], Step [636/3236], Loss: 2.9593, Perplexity: 19.2843
Epoch [1/3], Step [637/3236], Loss: 3.2667, Perplexity: 26.2254
Epoch [1/3], Step [638/3236], Loss: 2.7542, Perplexity: 15.7089
Epoch [1/3], Step [639/3236], Loss: 3.2623, Perplexity: 26.1104
Epoch [1/3], Step [640/3236], Loss: 3.2315, Perplexity: 25.3180
Epoch [1/3], Step [641/3236], Loss: 3.1470, Perplexity: 23.2653
Epoch [1/3], Step [642/3236], Loss: 2.9519, Perplexity: 19.1424
Epoch [1/3], Step [643/3236], Loss: 2.7665, Perplexity: 15.9033
Epoch [1/3], Step [644/3236], Loss: 3.0150, Perplexity: 20.3894
Epoch [1/3], Step [645/3236], Loss: 2.8946, Perplexity: 18.0758
Epoch [1/3], Step [646/3236], Loss: 2.7528, Perplexity: 15.6866
Epoch [1/3], Step [647/3236], Loss: 2.7446, Perplexity: 15.5585
Epoch [1/3], Step [648/3236], Loss: 3.3367, Perplexity: 28.1256
Epoch [1/3], Step [649/3236], Loss: 3.0133, Perplexity: 20.3539
Epoch [1/3], Step [650/3236], Loss: 3.1670, Perplexity: 23.7354
Epoch [1/3], Step [651/3236], Loss: 2.8965, Perplexity: 18.1115
Epoch [1/3], Step [652/3236], Loss: 2.8848, Perplexity: 17.8993
Epoch [1/3], Step [653/3236], Loss: 2.8651, Perplexity: 17.5513
Epoch [1/3], Step [654/3236], Loss: 2.9232, Perplexity: 18.6004
Epoch [1/3], Step [655/3236], Loss: 2.8548, Perplexity: 17.3703
Epoch [1/3], Step [656/3236], Loss: 2.6227, Perplexity: 13.7724
Epoch [1/3], Step [657/3236], Loss: 2.9552, Perplexity: 19.2056
Epoch [1/3], Step [658/3236], Loss: 2.8831, Perplexity: 17.8705
Epoch [1/3], Step [659/3236], Loss: 2.7873, Perplexity: 16.2364
Epoch [1/3], Step [660/3236], Loss: 3.1964, Perplexity: 24.4444
Epoch [1/3], Step [661/3236], Loss: 2.6950, Perplexity: 14.8061
Epoch [1/3], Step [662/3236], Loss: 2.8476, Perplexity: 17.2456
Epoch [1/3], Step [663/3236], Loss: 2.6893, Perplexity: 14.7212
Epoch [1/3], Step [664/3236], Loss: 3.3128, Perplexity: 27.4609
Epoch [1/3], Step [665/3236], Loss: 2.9309, Perplexity: 18.7447
Epoch [1/3], Step [666/3236], Loss: 2.6702, Perplexity: 14.4432
Epoch [1/3], Step [667/3236], Loss: 2.8947, Perplexity: 18.0776
Epoch [1/3], Step [668/3236], Loss: 2.7051, Perplexity: 14.9554
Epoch [1/3], Step [669/3236], Loss: 2.9522, Perplexity: 19.1478
Epoch [1/3], Step [670/3236], Loss: 2.8473, Perplexity: 17.2420
Epoch [1/3], Step [671/3236], Loss: 2.7712, Perplexity: 15.9783
Epoch [1/3], Step [672/3236], Loss: 2.7226, Perplexity: 15.2195
Epoch [1/3], Step [673/3236], Loss: 2.7384, Perplexity: 15.4618
Epoch [1/3], Step [674/3236], Loss: 2.7508, Perplexity: 15.6544
Epoch [1/3], Step [675/3236], Loss: 2.8451, Perplexity: 17.2026
Epoch [1/3], Step [676/3236], Loss: 2.9165, Perplexity: 18.4756
Epoch [1/3], Step [677/3236], Loss: 2.8348, Perplexity: 17.0261
Epoch [1/3], Step [678/3236], Loss: 2.7715, Perplexity: 15.9834
Epoch [1/3], Step [679/3236], Loss: 2.9330, Perplexity: 18.7843
Epoch [1/3], Step [680/3236], Loss: 2.8336, Perplexity: 17.0064
Epoch [1/3], Step [681/3236], Loss: 2.8755, Perplexity: 17.7336
Epoch [1/3], Step [682/3236], Loss: 3.2801, Perplexity: 26.5771
Epoch [1/3], Step [683/3236], Loss: 2.6854, Perplexity: 14.6642
Epoch [1/3], Step [684/3236], Loss: 2.6366, Perplexity: 13.9661
Epoch [1/3], Step [685/3236], Loss: 3.0211, Perplexity: 20.5139
Epoch [1/3], Step [686/3236], Loss: 2.8340, Perplexity: 17.0129
Epoch [1/3], Step [687/3236], Loss: 2.9814, Perplexity: 19.7148
Epoch [1/3], Step [688/3236], Loss: 2.7457, Perplexity: 15.5749
Epoch [1/3], Step [689/3236], Loss: 2.6572, Perplexity: 14.2569
Epoch [1/3], Step [690/3236], Loss: 2.7818, Perplexity: 16.1484
Epoch [1/3], Step [691/3236], Loss: 2.6301, Perplexity: 13.8756
Epoch [1/3], Step [692/3236], Loss: 2.9131, Perplexity: 18.4129
Epoch [1/3], Step [693/3236], Loss: 2.7329, Perplexity: 15.3770
Epoch [1/3], Step [694/3236], Loss: 2.8580, Perplexity: 17.4263
Epoch [1/3], Step [695/3236], Loss: 2.7958, Perplexity: 16.3757
Epoch [1/3], Step [696/3236], Loss: 2.8295, Perplexity: 16.9377
Epoch [1/3], Step [697/3236], Loss: 2.7907, Perplexity: 16.2922
Epoch [1/3], Step [698/3236], Loss: 2.7784, Perplexity: 16.0937
Epoch [1/3], Step [699/3236], Loss: 3.2564, Perplexity: 25.9557
Epoch [1/3], Step [700/3236], Loss: 3.0472, Perplexity: 21.0560
Epoch [1/3], Step [701/3236], Loss: 3.2372, Perplexity: 25.4624
Epoch [1/3], Step [702/3236], Loss: 2.7019, Perplexity: 14.9076
Epoch [1/3], Step [703/3236], Loss: 2.8445, Perplexity: 17.1924
Epoch [1/3], Step [704/3236], Loss: 2.8103, Perplexity: 16.6144
Epoch [1/3], Step [705/3236], Loss: 2.8193, Perplexity: 16.7647
Epoch [1/3], Step [706/3236], Loss: 2.8119, Perplexity: 16.6416
Epoch [1/3], Step [707/3236], Loss: 2.7906, Perplexity: 16.2905
Epoch [1/3], Step [708/3236], Loss: 2.8200, Perplexity: 16.7773
Epoch [1/3], Step [709/3236], Loss: 3.0006, Perplexity: 20.0970
Epoch [1/3], Step [710/3236], Loss: 2.7816, Perplexity: 16.1448
Epoch [1/3], Step [711/3236], Loss: 2.7731, Perplexity: 16.0078
Epoch [1/3], Step [712/3236], Loss: 3.1405, Perplexity: 23.1156
Epoch [1/3], Step [713/3236], Loss: 2.8112, Perplexity: 16.6304
Epoch [1/3], Step [714/3236], Loss: 2.8502, Perplexity: 17.2907
Epoch [1/3], Step [715/3236], Loss: 2.8961, Perplexity: 18.1041
Epoch [1/3], Step [716/3236], Loss: 2.7991, Perplexity: 16.4303
Epoch [1/3], Step [717/3236], Loss: 2.8435, Perplexity: 17.1758
Epoch [1/3], Step [718/3236], Loss: 2.6518, Perplexity: 14.1796
Epoch [1/3], Step [719/3236], Loss: 2.8168, Perplexity: 16.7229
Epoch [1/3], Step [720/3236], Loss: 2.7527, Perplexity: 15.6852
Epoch [1/3], Step [721/3236], Loss: 2.7107, Perplexity: 15.0395
Epoch [1/3], Step [722/3236], Loss: 3.2453, Perplexity: 25.6681
Epoch [1/3], Step [723/3236], Loss: 2.6895, Perplexity: 14.7247
Epoch [1/3], Step [724/3236], Loss: 2.6561, Perplexity: 14.2413
Epoch [1/3], Step [725/3236], Loss: 2.7076, Perplexity: 14.9935
Epoch [1/3], Step [726/3236], Loss: 2.9743, Perplexity: 19.5754
Epoch [1/3], Step [727/3236], Loss: 2.8660, Perplexity: 17.5668
Epoch [1/3], Step [728/3236], Loss: 2.7920, Perplexity: 16.3129
Epoch [1/3], Step [729/3236], Loss: 2.7990, Perplexity: 16.4276
Epoch [1/3], Step [730/3236], Loss: 3.2976, Perplexity: 27.0468
Epoch [1/3], Step [731/3236], Loss: 3.0521, Perplexity: 21.1602
Epoch [1/3], Step [732/3236], Loss: 2.7925, Perplexity: 16.3215
Epoch [1/3], Step [733/3236], Loss: 2.8165, Perplexity: 16.7180
Epoch [1/3], Step [734/3236], Loss: 2.7686, Perplexity: 15.9367
Epoch [1/3], Step [735/3236], Loss: 2.7170, Perplexity: 15.1346
Epoch [1/3], Step [736/3236], Loss: 2.8240, Perplexity: 16.8440
Epoch [1/3], Step [737/3236], Loss: 3.0591, Perplexity: 21.3088
Epoch [1/3], Step [738/3236], Loss: 2.7675, Perplexity: 15.9193
Epoch [1/3], Step [739/3236], Loss: 2.8525, Perplexity: 17.3304
Epoch [1/3], Step [740/3236], Loss: 2.8388, Perplexity: 17.0949
Epoch [1/3], Step [741/3236], Loss: 2.9019, Perplexity: 18.2084
Epoch [1/3], Step [742/3236], Loss: 3.1860, Perplexity: 24.1909
Epoch [1/3], Step [743/3236], Loss: 2.7641, Perplexity: 15.8642
Epoch [1/3], Step [744/3236], Loss: 2.7687, Perplexity: 15.9386
Epoch [1/3], Step [745/3236], Loss: 2.8737, Perplexity: 17.7019
Epoch [1/3], Step [746/3236], Loss: 2.5950, Perplexity: 13.3971
Epoch [1/3], Step [747/3236], Loss: 2.7890, Perplexity: 16.2648
Epoch [1/3], Step [748/3236], Loss: 3.5863, Perplexity: 36.0988
Epoch [1/3], Step [749/3236], Loss: 3.1243, Perplexity: 22.7430
Epoch [1/3], Step [750/3236], Loss: 2.6298, Perplexity: 13.8710
Epoch [1/3], Step [751/3236], Loss: 2.8707, Perplexity: 17.6487
Epoch [1/3], Step [752/3236], Loss: 2.7047, Perplexity: 14.9500
Epoch [1/3], Step [753/3236], Loss: 2.8178, Perplexity: 16.7399
Epoch [1/3], Step [754/3236], Loss: 2.8485, Perplexity: 17.2624
Epoch [1/3], Step [755/3236], Loss: 2.7199, Perplexity: 15.1790
Epoch [1/3], Step [756/3236], Loss: 2.8438, Perplexity: 17.1801
Epoch [1/3], Step [757/3236], Loss: 2.6810, Perplexity: 14.6003
Epoch [1/3], Step [758/3236], Loss: 2.9054, Perplexity: 18.2719
Epoch [1/3], Step [759/3236], Loss: 3.2444, Perplexity: 25.6468
Epoch [1/3], Step [760/3236], Loss: 2.6313, Perplexity: 13.8912
Epoch [1/3], Step [761/3236], Loss: 2.8658, Perplexity: 17.5634
Epoch [1/3], Step [762/3236], Loss: 2.8500, Perplexity: 17.2876
Epoch [1/3], Step [763/3236], Loss: 2.7449, Perplexity: 15.5630
Epoch [1/3], Step [764/3236], Loss: 2.8740, Perplexity: 17.7079
Epoch [1/3], Step [765/3236], Loss: 2.7010, Perplexity: 14.8944
Epoch [1/3], Step [766/3236], Loss: 2.6744, Perplexity: 14.5034
Epoch [1/3], Step [767/3236], Loss: 2.8341, Perplexity: 17.0148
Epoch [1/3], Step [768/3236], Loss: 2.6655, Perplexity: 14.3752
Epoch [1/3], Step [769/3236], Loss: 2.6508, Perplexity: 14.1647
Epoch [1/3], Step [770/3236], Loss: 3.0082, Perplexity: 20.2505
Epoch [1/3], Step [771/3236], Loss: 3.5295, Perplexity: 34.1078
Epoch [1/3], Step [772/3236], Loss: 2.8723, Perplexity: 17.6769
Epoch [1/3], Step [773/3236], Loss: 2.6946, Perplexity: 14.7999
Epoch [1/3], Step [774/3236], Loss: 3.1494, Perplexity: 23.3214
Epoch [1/3], Step [775/3236], Loss: 2.8201, Perplexity: 16.7786
Epoch [1/3], Step [776/3236], Loss: 2.8353, Perplexity: 17.0361
Epoch [1/3], Step [777/3236], Loss: 2.9191, Perplexity: 18.5237
Epoch [1/3], Step [778/3236], Loss: 2.8870, Perplexity: 17.9393
Epoch [1/3], Step [779/3236], Loss: 2.8720, Perplexity: 17.6724
Epoch [1/3], Step [780/3236], Loss: 2.7658, Perplexity: 15.8921
Epoch [1/3], Step [781/3236], Loss: 2.7943, Perplexity: 16.3516
Epoch [1/3], Step [782/3236], Loss: 2.7719, Perplexity: 15.9886
Epoch [1/3], Step [783/3236], Loss: 2.7539, Perplexity: 15.7039
Epoch [1/3], Step [784/3236], Loss: 2.9028, Perplexity: 18.2259
Epoch [1/3], Step [785/3236], Loss: 3.5514, Perplexity: 34.8626
Epoch [1/3], Step [786/3236], Loss: 2.9733, Perplexity: 19.5564
Epoch [1/3], Step [787/3236], Loss: 2.6323, Perplexity: 13.9057
Epoch [1/3], Step [788/3236], Loss: 2.9717, Perplexity: 19.5248
Epoch [1/3], Step [789/3236], Loss: 2.7468, Perplexity: 15.5919
Epoch [1/3], Step [790/3236], Loss: 2.7574, Perplexity: 15.7585
Epoch [1/3], Step [791/3236], Loss: 2.7772, Perplexity: 16.0746
Epoch [1/3], Step [792/3236], Loss: 2.7072, Perplexity: 14.9877
Epoch [1/3], Step [793/3236], Loss: 2.6684, Perplexity: 14.4167
Epoch [1/3], Step [794/3236], Loss: 2.9092, Perplexity: 18.3430
Epoch [1/3], Step [795/3236], Loss: 2.8578, Perplexity: 17.4228
Epoch [1/3], Step [796/3236], Loss: 2.7641, Perplexity: 15.8646
Epoch [1/3], Step [797/3236], Loss: 2.8027, Perplexity: 16.4887
Epoch [1/3], Step [798/3236], Loss: 2.6128, Perplexity: 13.6374
Epoch [1/3], Step [799/3236], Loss: 2.7144, Perplexity: 15.0960
Epoch [1/3], Step [800/3236], Loss: 2.7848, Perplexity: 16.1971
Epoch [1/3], Step [801/3236], Loss: 2.8080, Perplexity: 16.5760
Epoch [1/3], Step [802/3236], Loss: 2.6846, Perplexity: 14.6526
Epoch [1/3], Step [803/3236], Loss: 2.5929, Perplexity: 13.3683
Epoch [1/3], Step [804/3236], Loss: 2.8734, Perplexity: 17.6963
Epoch [1/3], Step [805/3236], Loss: 2.6358, Perplexity: 13.9549
Epoch [1/3], Step [806/3236], Loss: 2.8295, Perplexity: 16.9373
Epoch [1/3], Step [807/3236], Loss: 2.7213, Perplexity: 15.1994
Epoch [1/3], Step [808/3236], Loss: 2.7956, Perplexity: 16.3731
Epoch [1/3], Step [809/3236], Loss: 2.6775, Perplexity: 14.5485
Epoch [1/3], Step [810/3236], Loss: 2.6239, Perplexity: 13.7888
Epoch [1/3], Step [811/3236], Loss: 2.7868, Perplexity: 16.2293
Epoch [1/3], Step [812/3236], Loss: 2.8248, Perplexity: 16.8568
Epoch [1/3], Step [813/3236], Loss: 2.6362, Perplexity: 13.9596
Epoch [1/3], Step [814/3236], Loss: 2.7427, Perplexity: 15.5287
Epoch [1/3], Step [815/3236], Loss: 2.7543, Perplexity: 15.7093
Epoch [1/3], Step [816/3236], Loss: 2.8046, Perplexity: 16.5199
Epoch [1/3], Step [817/3236], Loss: 3.1796, Perplexity: 24.0376
Epoch [1/3], Step [818/3236], Loss: 2.7474, Perplexity: 15.6018
Epoch [1/3], Step [819/3236], Loss: 2.5400, Perplexity: 12.6793
Epoch [1/3], Step [820/3236], Loss: 2.6588, Perplexity: 14.2796
Epoch [1/3], Step [821/3236], Loss: 2.6775, Perplexity: 14.5484
Epoch [1/3], Step [822/3236], Loss: 2.8216, Perplexity: 16.8031
Epoch [1/3], Step [823/3236], Loss: 2.6780, Perplexity: 14.5555
Epoch [1/3], Step [824/3236], Loss: 2.8742, Perplexity: 17.7107
Epoch [1/3], Step [825/3236], Loss: 2.9636, Perplexity: 19.3674
Epoch [1/3], Step [826/3236], Loss: 2.6497, Perplexity: 14.1495
Epoch [1/3], Step [827/3236], Loss: 2.6651, Perplexity: 14.3699
Epoch [1/3], Step [828/3236], Loss: 2.7857, Perplexity: 16.2112
Epoch [1/3], Step [829/3236], Loss: 2.9107, Perplexity: 18.3703
Epoch [1/3], Step [830/3236], Loss: 3.2417, Perplexity: 25.5772
Epoch [1/3], Step [831/3236], Loss: 2.7434, Perplexity: 15.5390
Epoch [1/3], Step [832/3236], Loss: 2.6965, Perplexity: 14.8275
Epoch [1/3], Step [833/3236], Loss: 2.7409, Perplexity: 15.5007
Epoch [1/3], Step [834/3236], Loss: 2.8091, Perplexity: 16.5944
Epoch [1/3], Step [835/3236], Loss: 2.8106, Perplexity: 16.6199
Epoch [1/3], Step [836/3236], Loss: 2.6817, Perplexity: 14.6101
Epoch [1/3], Step [837/3236], Loss: 2.8272, Perplexity: 16.8983
Epoch [1/3], Step [838/3236], Loss: 2.6091, Perplexity: 13.5870
Epoch [1/3], Step [839/3236], Loss: 2.6838, Perplexity: 14.6411
Epoch [1/3], Step [840/3236], Loss: 2.7599, Perplexity: 15.7985
Epoch [1/3], Step [841/3236], Loss: 2.8217, Perplexity: 16.8047
Epoch [1/3], Step [842/3236], Loss: 2.6920, Perplexity: 14.7607
Epoch [1/3], Step [843/3236], Loss: 2.8027, Perplexity: 16.4897
Epoch [1/3], Step [844/3236], Loss: 2.6152, Perplexity: 13.6701
Epoch [1/3], Step [845/3236], Loss: 2.6462, Perplexity: 14.1007
Epoch [1/3], Step [846/3236], Loss: 3.0360, Perplexity: 20.8211
Epoch [1/3], Step [847/3236], Loss: 2.6367, Perplexity: 13.9668
Epoch [1/3], Step [848/3236], Loss: 2.7054, Perplexity: 14.9604
Epoch [1/3], Step [849/3236], Loss: 2.6513, Perplexity: 14.1720
Epoch [1/3], Step [850/3236], Loss: 2.7781, Perplexity: 16.0884
Epoch [1/3], Step [851/3236], Loss: 2.9532, Perplexity: 19.1663
Epoch [1/3], Step [852/3236], Loss: 2.7691, Perplexity: 15.9435
Epoch [1/3], Step [853/3236], Loss: 2.6908, Perplexity: 14.7440
Epoch [1/3], Step [854/3236], Loss: 2.8114, Perplexity: 16.6332
Epoch [1/3], Step [855/3236], Loss: 2.7177, Perplexity: 15.1453
Epoch [1/3], Step [856/3236], Loss: 2.8280, Perplexity: 16.9123
Epoch [1/3], Step [857/3236], Loss: 3.0494, Perplexity: 21.1022
Epoch [1/3], Step [858/3236], Loss: 2.8211, Perplexity: 16.7952
Epoch [1/3], Step [859/3236], Loss: 2.8127, Perplexity: 16.6543
Epoch [1/3], Step [860/3236], Loss: 2.6367, Perplexity: 13.9673
Epoch [1/3], Step [861/3236], Loss: 3.0673, Perplexity: 21.4835
Epoch [1/3], Step [862/3236], Loss: 2.8228, Perplexity: 16.8231
Epoch [1/3], Step [863/3236], Loss: 2.7569, Perplexity: 15.7514
Epoch [1/3], Step [864/3236], Loss: 2.9365, Perplexity: 18.8499
Epoch [1/3], Step [865/3236], Loss: 2.9631, Perplexity: 19.3570
Epoch [1/3], Step [866/3236], Loss: 2.9028, Perplexity: 18.2254
Epoch [1/3], Step [867/3236], Loss: 2.6274, Perplexity: 13.8377
Epoch [1/3], Step [868/3236], Loss: 2.7285, Perplexity: 15.3103
Epoch [1/3], Step [869/3236], Loss: 2.8755, Perplexity: 17.7337
Epoch [1/3], Step [870/3236], Loss: 2.8918, Perplexity: 18.0259
Epoch [1/3], Step [871/3236], Loss: 2.5132, Perplexity: 12.3441
Epoch [1/3], Step [872/3236], Loss: 2.7562, Perplexity: 15.7397
Epoch [1/3], Step [873/3236], Loss: 2.8126, Perplexity: 16.6528
Epoch [1/3], Step [874/3236], Loss: 2.6790, Perplexity: 14.5702
Epoch [1/3], Step [875/3236], Loss: 2.8309, Perplexity: 16.9605
Epoch [1/3], Step [876/3236], Loss: 2.6316, Perplexity: 13.8956
Epoch [1/3], Step [877/3236], Loss: 2.6825, Perplexity: 14.6218
Epoch [1/3], Step [878/3236], Loss: 2.4677, Perplexity: 11.7948
Epoch [1/3], Step [879/3236], Loss: 2.6250, Perplexity: 13.8040
Epoch [1/3], Step [880/3236], Loss: 2.9529, Perplexity: 19.1620
Epoch [1/3], Step [881/3236], Loss: 2.7098, Perplexity: 15.0261
Epoch [1/3], Step [882/3236], Loss: 2.4546, Perplexity: 11.6415
Epoch [1/3], Step [883/3236], Loss: 2.7329, Perplexity: 15.3773
Epoch [1/3], Step [884/3236], Loss: 2.6199, Perplexity: 13.7341
Epoch [1/3], Step [885/3236], Loss: 2.7491, Perplexity: 15.6279
Epoch [1/3], Step [886/3236], Loss: 2.7561, Perplexity: 15.7388
Epoch [1/3], Step [887/3236], Loss: 2.6854, Perplexity: 14.6642
Epoch [1/3], Step [888/3236], Loss: 2.5828, Perplexity: 13.2338
Epoch [1/3], Step [889/3236], Loss: 2.5811, Perplexity: 13.2114
Epoch [1/3], Step [890/3236], Loss: 2.6750, Perplexity: 14.5124
Epoch [1/3], Step [891/3236], Loss: 2.5580, Perplexity: 12.9098
Epoch [1/3], Step [892/3236], Loss: 2.6736, Perplexity: 14.4920
Epoch [1/3], Step [893/3236], Loss: 2.7183, Perplexity: 15.1548
Epoch [1/3], Step [894/3236], Loss: 2.7076, Perplexity: 14.9927
Epoch [1/3], Step [895/3236], Loss: 2.7482, Perplexity: 15.6140
Epoch [1/3], Step [896/3236], Loss: 2.7582, Perplexity: 15.7713
Epoch [1/3], Step [897/3236], Loss: 2.6209, Perplexity: 13.7488
Epoch [1/3], Step [898/3236], Loss: 2.7233, Perplexity: 15.2303
Epoch [1/3], Step [899/3236], Loss: 3.4266, Perplexity: 30.7709
Epoch [1/3], Step [900/3236], Loss: 2.5802, Perplexity: 13.1993
Epoch [1/3], Step [901/3236], Loss: 2.6074, Perplexity: 13.5638
Epoch [1/3], Step [902/3236], Loss: 2.7019, Perplexity: 14.9086
Epoch [1/3], Step [903/3236], Loss: 2.4846, Perplexity: 11.9958
Epoch [1/3], Step [904/3236], Loss: 2.7506, Perplexity: 15.6518
Epoch [1/3], Step [905/3236], Loss: 2.7348, Perplexity: 15.4061
Epoch [1/3], Step [906/3236], Loss: 2.5871, Perplexity: 13.2915
Epoch [1/3], Step [907/3236], Loss: 2.7051, Perplexity: 14.9563
Epoch [1/3], Step [908/3236], Loss: 2.5664, Perplexity: 13.0185
Epoch [1/3], Step [909/3236], Loss: 2.4767, Perplexity: 11.9023
Epoch [1/3], Step [910/3236], Loss: 2.5687, Perplexity: 13.0483
Epoch [1/3], Step [911/3236], Loss: 2.9166, Perplexity: 18.4788
Epoch [1/3], Step [912/3236], Loss: 2.5677, Perplexity: 13.0352
Epoch [1/3], Step [913/3236], Loss: 2.7448, Perplexity: 15.5623
Epoch [1/3], Step [914/3236], Loss: 2.4676, Perplexity: 11.7939
Epoch [1/3], Step [915/3236], Loss: 2.6880, Perplexity: 14.7022
Epoch [1/3], Step [916/3236], Loss: 2.5826, Perplexity: 13.2313
Epoch [1/3], Step [917/3236], Loss: 2.4516, Perplexity: 11.6066
Epoch [1/3], Step [918/3236], Loss: 2.8849, Perplexity: 17.9023
Epoch [1/3], Step [919/3236], Loss: 2.5529, Perplexity: 12.8445
Epoch [1/3], Step [920/3236], Loss: 2.4828, Perplexity: 11.9752
Epoch [1/3], Step [921/3236], Loss: 2.5312, Perplexity: 12.5686
Epoch [1/3], Step [922/3236], Loss: 2.5082, Perplexity: 12.2829
Epoch [1/3], Step [923/3236], Loss: 2.7928, Perplexity: 16.3263
Epoch [1/3], Step [924/3236], Loss: 2.5931, Perplexity: 13.3708
Epoch [1/3], Step [925/3236], Loss: 2.7730, Perplexity: 16.0071
Epoch [1/3], Step [926/3236], Loss: 2.7075, Perplexity: 14.9912
Epoch [1/3], Step [927/3236], Loss: 2.5270, Perplexity: 12.5163
Epoch [1/3], Step [928/3236], Loss: 2.5631, Perplexity: 12.9762
Epoch [1/3], Step [929/3236], Loss: 2.5216, Perplexity: 12.4484
Epoch [1/3], Step [930/3236], Loss: 2.7250, Perplexity: 15.2567
Epoch [1/3], Step [931/3236], Loss: 2.6206, Perplexity: 13.7439
Epoch [1/3], Step [932/3236], Loss: 2.6876, Perplexity: 14.6969
Epoch [1/3], Step [933/3236], Loss: 2.4815, Perplexity: 11.9587
Epoch [1/3], Step [934/3236], Loss: 2.7208, Perplexity: 15.1923
Epoch [1/3], Step [935/3236], Loss: 2.7152, Perplexity: 15.1083
Epoch [1/3], Step [936/3236], Loss: 2.4670, Perplexity: 11.7867
Epoch [1/3], Step [937/3236], Loss: 2.6468, Perplexity: 14.1087
Epoch [1/3], Step [938/3236], Loss: 2.6194, Perplexity: 13.7277
Epoch [1/3], Step [939/3236], Loss: 2.5799, Perplexity: 13.1963
Epoch [1/3], Step [940/3236], Loss: 2.7741, Perplexity: 16.0235
Epoch [1/3], Step [941/3236], Loss: 2.6992, Perplexity: 14.8685
Epoch [1/3], Step [942/3236], Loss: 2.8303, Perplexity: 16.9502
Epoch [1/3], Step [943/3236], Loss: 2.9212, Perplexity: 18.5634
Epoch [1/3], Step [944/3236], Loss: 2.5847, Perplexity: 13.2593
Epoch [1/3], Step [945/3236], Loss: 2.5490, Perplexity: 12.7937
Epoch [1/3], Step [946/3236], Loss: 2.5658, Perplexity: 13.0110
Epoch [1/3], Step [947/3236], Loss: 2.7892, Perplexity: 16.2680
Epoch [1/3], Step [948/3236], Loss: 2.4870, Perplexity: 12.0249
Epoch [1/3], Step [949/3236], Loss: 3.3898, Perplexity: 29.6603
Epoch [1/3], Step [950/3236], Loss: 2.7460, Perplexity: 15.5803
Epoch [1/3], Step [951/3236], Loss: 2.5636, Perplexity: 12.9830
Epoch [1/3], Step [952/3236], Loss: 2.6566, Perplexity: 14.2481
Epoch [1/3], Step [953/3236], Loss: 2.5480, Perplexity: 12.7820
Epoch [1/3], Step [954/3236], Loss: 2.6362, Perplexity: 13.9599
Epoch [1/3], Step [955/3236], Loss: 2.8213, Perplexity: 16.7991
Epoch [1/3], Step [956/3236], Loss: 2.7996, Perplexity: 16.4384
Epoch [1/3], Step [957/3236], Loss: 3.1402, Perplexity: 23.1080
Epoch [1/3], Step [958/3236], Loss: 2.6223, Perplexity: 13.7676
Epoch [1/3], Step [959/3236], Loss: 2.6078, Perplexity: 13.5695
Epoch [1/3], Step [960/3236], Loss: 2.6463, Perplexity: 14.1023
Epoch [1/3], Step [961/3236], Loss: 2.6404, Perplexity: 14.0186
Epoch [1/3], Step [962/3236], Loss: 2.5654, Perplexity: 13.0065
Epoch [1/3], Step [963/3236], Loss: 2.6431, Perplexity: 14.0568
Epoch [1/3], Step [964/3236], Loss: 2.5734, Perplexity: 13.1098
Epoch [1/3], Step [965/3236], Loss: 3.0539, Perplexity: 21.1972
Epoch [1/3], Step [966/3236], Loss: 2.6512, Perplexity: 14.1704
Epoch [1/3], Step [967/3236], Loss: 2.5092, Perplexity: 12.2956
Epoch [1/3], Step [968/3236], Loss: 2.8409, Perplexity: 17.1310
Epoch [1/3], Step [969/3236], Loss: 2.5879, Perplexity: 13.3021
Epoch [1/3], Step [970/3236], Loss: 2.6150, Perplexity: 13.6668
Epoch [1/3], Step [971/3236], Loss: 2.6639, Perplexity: 14.3523
Epoch [1/3], Step [972/3236], Loss: 2.5483, Perplexity: 12.7857
Epoch [1/3], Step [973/3236], Loss: 2.6920, Perplexity: 14.7614
Epoch [1/3], Step [974/3236], Loss: 2.4659, Perplexity: 11.7737
Epoch [1/3], Step [975/3236], Loss: 2.6655, Perplexity: 14.3754
Epoch [1/3], Step [976/3236], Loss: 3.0957, Perplexity: 22.1038
Epoch [1/3], Step [977/3236], Loss: 2.5602, Perplexity: 12.9385
Epoch [1/3], Step [978/3236], Loss: 2.5152, Perplexity: 12.3694
Epoch [1/3], Step [979/3236], Loss: 2.8819, Perplexity: 17.8485
Epoch [1/3], Step [980/3236], Loss: 2.8910, Perplexity: 18.0115
Epoch [1/3], Step [981/3236], Loss: 3.0192, Perplexity: 20.4753
Epoch [1/3], Step [982/3236], Loss: 2.7150, Perplexity: 15.1053
Epoch [1/3], Step [983/3236], Loss: 2.5393, Perplexity: 12.6709
Epoch [1/3], Step [984/3236], Loss: 2.5796, Perplexity: 13.1913
Epoch [1/3], Step [985/3236], Loss: 3.1238, Perplexity: 22.7334
Epoch [1/3], Step [986/3236], Loss: 2.6623, Perplexity: 14.3296
Epoch [1/3], Step [987/3236], Loss: 2.4302, Perplexity: 11.3617
Epoch [1/3], Step [988/3236], Loss: 2.6154, Perplexity: 13.6732
Epoch [1/3], Step [989/3236], Loss: 2.7242, Perplexity: 15.2439
Epoch [1/3], Step [990/3236], Loss: 2.5444, Perplexity: 12.7358
Epoch [1/3], Step [991/3236], Loss: 2.6126, Perplexity: 13.6341
Epoch [1/3], Step [992/3236], Loss: 3.0374, Perplexity: 20.8506
Epoch [1/3], Step [993/3236], Loss: 2.6233, Perplexity: 13.7816
Epoch [1/3], Step [994/3236], Loss: 2.6185, Perplexity: 13.7152
Epoch [1/3], Step [995/3236], Loss: 2.4881, Perplexity: 12.0384
Epoch [1/3], Step [996/3236], Loss: 2.7259, Perplexity: 15.2708
Epoch [1/3], Step [997/3236], Loss: 2.5719, Perplexity: 13.0913
Epoch [1/3], Step [998/3236], Loss: 2.6087, Perplexity: 13.5813
Epoch [1/3], Step [999/3236], Loss: 2.8173, Perplexity: 16.7314
Epoch [1/3], Step [1000/3236], Loss: 2.5742, Perplexity: 13.1212
Epoch [1/3], Step [1001/3236], Loss: 2.6738, Perplexity: 14.4945
Epoch [1/3], Step [1002/3236], Loss: 2.7807, Perplexity: 16.1297
Epoch [1/3], Step [1003/3236], Loss: 2.4362, Perplexity: 11.4301
Epoch [1/3], Step [1004/3236], Loss: 2.5102, Perplexity: 12.3072
Epoch [1/3], Step [1005/3236], Loss: 2.7083, Perplexity: 15.0032
Epoch [1/3], Step [1006/3236], Loss: 2.6426, Perplexity: 14.0502
Epoch [1/3], Step [1007/3236], Loss: 2.5341, Perplexity: 12.6047
Epoch [1/3], Step [1008/3236], Loss: 2.4862, Perplexity: 12.0150
Epoch [1/3], Step [1009/3236], Loss: 2.6902, Perplexity: 14.7348
Epoch [1/3], Step [1010/3236], Loss: 2.7012, Perplexity: 14.8981
Epoch [1/3], Step [1011/3236], Loss: 2.4777, Perplexity: 11.9136
Epoch [1/3], Step [1012/3236], Loss: 2.6536, Perplexity: 14.2055
Epoch [1/3], Step [1013/3236], Loss: 2.4568, Perplexity: 11.6678
Epoch [1/3], Step [1014/3236], Loss: 3.0185, Perplexity: 20.4607
Epoch [1/3], Step [1015/3236], Loss: 2.8479, Perplexity: 17.2514
Epoch [1/3], Step [1016/3236], Loss: 2.4394, Perplexity: 11.4657
Epoch [1/3], Step [1017/3236], Loss: 2.8761, Perplexity: 17.7457
Epoch [1/3], Step [1018/3236], Loss: 2.7275, Perplexity: 15.2944
Epoch [1/3], Step [1019/3236], Loss: 2.5412, Perplexity: 12.6955
Epoch [1/3], Step [1020/3236], Loss: 2.7508, Perplexity: 15.6550
Epoch [1/3], Step [1021/3236], Loss: 2.4836, Perplexity: 11.9849
Epoch [1/3], Step [1022/3236], Loss: 2.8335, Perplexity: 17.0053
Epoch [1/3], Step [1023/3236], Loss: 2.6725, Perplexity: 14.4757
Epoch [1/3], Step [1024/3236], Loss: 2.5898, Perplexity: 13.3277
Epoch [1/3], Step [1025/3236], Loss: 2.6697, Perplexity: 14.4357
Epoch [1/3], Step [1026/3236], Loss: 2.5089, Perplexity: 12.2909
Epoch [1/3], Step [1027/3236], Loss: 2.5140, Perplexity: 12.3538
Epoch [1/3], Step [1028/3236], Loss: 2.7636, Perplexity: 15.8568
Epoch [1/3], Step [1029/3236], Loss: 2.5784, Perplexity: 13.1756
Epoch [1/3], Step [1030/3236], Loss: 2.6235, Perplexity: 13.7832
Epoch [1/3], Step [1031/3236], Loss: 2.6901, Perplexity: 14.7325
Epoch [1/3], Step [1032/3236], Loss: 2.8820, Perplexity: 17.8507
Epoch [1/3], Step [1033/3236], Loss: 2.7371, Perplexity: 15.4424
Epoch [1/3], Step [1034/3236], Loss: 2.5154, Perplexity: 12.3719
Epoch [1/3], Step [1035/3236], Loss: 2.7010, Perplexity: 14.8952
Epoch [1/3], Step [1036/3236], Loss: 2.6186, Perplexity: 13.7159
Epoch [1/3], Step [1037/3236], Loss: 2.7144, Perplexity: 15.0953
Epoch [1/3], Step [1038/3236], Loss: 2.7315, Perplexity: 15.3565
Epoch [1/3], Step [1039/3236], Loss: 2.4804, Perplexity: 11.9461
Epoch [1/3], Step [1040/3236], Loss: 2.8588, Perplexity: 17.4399
Epoch [1/3], Step [1041/3236], Loss: 2.5689, Perplexity: 13.0509
Epoch [1/3], Step [1042/3236], Loss: 2.4863, Perplexity: 12.0168
Epoch [1/3], Step [1043/3236], Loss: 2.6046, Perplexity: 13.5254
Epoch [1/3], Step [1044/3236], Loss: 2.5080, Perplexity: 12.2801
Epoch [1/3], Step [1045/3236], Loss: 2.9023, Perplexity: 18.2154
Epoch [1/3], Step [1046/3236], Loss: 3.0197, Perplexity: 20.4851
Epoch [1/3], Step [1047/3236], Loss: 2.5141, Perplexity: 12.3555
Epoch [1/3], Step [1048/3236], Loss: 2.5778, Perplexity: 13.1682
Epoch [1/3], Step [1049/3236], Loss: 2.7843, Perplexity: 16.1883
Epoch [1/3], Step [1050/3236], Loss: 2.5337, Perplexity: 12.5996
Epoch [1/3], Step [1051/3236], Loss: 2.6342, Perplexity: 13.9328
Epoch [1/3], Step [1052/3236], Loss: 2.4681, Perplexity: 11.7995
Epoch [1/3], Step [1053/3236], Loss: 2.6497, Perplexity: 14.1494
Epoch [1/3], Step [1054/3236], Loss: 2.8798, Perplexity: 17.8113
Epoch [1/3], Step [1055/3236], Loss: 2.9442, Perplexity: 18.9957
Epoch [1/3], Step [1056/3236], Loss: 2.5652, Perplexity: 13.0031
Epoch [1/3], Step [1057/3236], Loss: 2.4429, Perplexity: 11.5066
Epoch [1/3], Step [1058/3236], Loss: 2.5480, Perplexity: 12.7817
Epoch [1/3], Step [1059/3236], Loss: 2.5331, Perplexity: 12.5926
Epoch [1/3], Step [1060/3236], Loss: 2.6574, Perplexity: 14.2595
Epoch [1/3], Step [1061/3236], Loss: 2.4111, Perplexity: 11.1466
Epoch [1/3], Step [1062/3236], Loss: 2.6616, Perplexity: 14.3194
Epoch [1/3], Step [1063/3236], Loss: 2.9429, Perplexity: 18.9699
Epoch [1/3], Step [1064/3236], Loss: 2.3502, Perplexity: 10.4881
Epoch [1/3], Step [1065/3236], Loss: 2.5366, Perplexity: 12.6372
Epoch [1/3], Step [1066/3236], Loss: 2.8900, Perplexity: 17.9928
Epoch [1/3], Step [1067/3236], Loss: 2.3820, Perplexity: 10.8263
Epoch [1/3], Step [1068/3236], Loss: 2.5734, Perplexity: 13.1101
Epoch [1/3], Step [1069/3236], Loss: 2.6453, Perplexity: 14.0871
Epoch [1/3], Step [1070/3236], Loss: 2.5264, Perplexity: 12.5081
Epoch [1/3], Step [1071/3236], Loss: 2.3162, Perplexity: 10.1373
Epoch [1/3], Step [1072/3236], Loss: 2.6597, Perplexity: 14.2922
Epoch [1/3], Step [1073/3236], Loss: 2.9848, Perplexity: 19.7831
Epoch [1/3], Step [1074/3236], Loss: 2.4388, Perplexity: 11.4590
Epoch [1/3], Step [1075/3236], Loss: 2.4456, Perplexity: 11.5380
Epoch [1/3], Step [1076/3236], Loss: 2.7544, Perplexity: 15.7116
Epoch [1/3], Step [1077/3236], Loss: 2.3380, Perplexity: 10.3608
Epoch [1/3], Step [1078/3236], Loss: 3.4600, Perplexity: 31.8181
Epoch [1/3], Step [1079/3236], Loss: 2.7770, Perplexity: 16.0710
Epoch [1/3], Step [1080/3236], Loss: 2.5157, Perplexity: 12.3758
Epoch [1/3], Step [1081/3236], Loss: 2.5890, Perplexity: 13.3162
Epoch [1/3], Step [1082/3236], Loss: 2.5522, Perplexity: 12.8355
Epoch [1/3], Step [1083/3236], Loss: 2.4493, Perplexity: 11.5805
Epoch [1/3], Step [1084/3236], Loss: 2.6545, Perplexity: 14.2174
Epoch [1/3], Step [1085/3236], Loss: 2.5676, Perplexity: 13.0349
Epoch [1/3], Step [1086/3236], Loss: 2.4582, Perplexity: 11.6841
Epoch [1/3], Step [1087/3236], Loss: 2.6891, Perplexity: 14.7183
Epoch [1/3], Step [1088/3236], Loss: 2.4170, Perplexity: 11.2124
Epoch [1/3], Step [1089/3236], Loss: 2.4307, Perplexity: 11.3666
Epoch [1/3], Step [1090/3236], Loss: 2.5422, Perplexity: 12.7075
Epoch [1/3], Step [1091/3236], Loss: 2.5047, Perplexity: 12.2404
Epoch [1/3], Step [1092/3236], Loss: 2.6530, Perplexity: 14.1962
Epoch [1/3], Step [1093/3236], Loss: 2.6971, Perplexity: 14.8364
Epoch [1/3], Step [1094/3236], Loss: 2.4853, Perplexity: 12.0053
Epoch [1/3], Step [1095/3236], Loss: 2.5903, Perplexity: 13.3335
Epoch [1/3], Step [1096/3236], Loss: 2.4686, Perplexity: 11.8054
Epoch [1/3], Step [1097/3236], Loss: 2.8691, Perplexity: 17.6209
Epoch [1/3], Step [1098/3236], Loss: 2.5989, Perplexity: 13.4486
Epoch [1/3], Step [1099/3236], Loss: 2.4716, Perplexity: 11.8412
Epoch [1/3], Step [1100/3236], Loss: 3.2200, Perplexity: 25.0288
Epoch [1/3], Step [1101/3236], Loss: 2.4395, Perplexity: 11.4670
Epoch [1/3], Step [1102/3236], Loss: 2.3874, Perplexity: 10.8847
Epoch [1/3], Step [1103/3236], Loss: 2.4508, Perplexity: 11.5976
Epoch [1/3], Step [1104/3236], Loss: 2.4669, Perplexity: 11.7853
Epoch [1/3], Step [1105/3236], Loss: 2.5567, Perplexity: 12.8935
Epoch [1/3], Step [1106/3236], Loss: 2.6472, Perplexity: 14.1147
Epoch [1/3], Step [1107/3236], Loss: 2.5191, Perplexity: 12.4177
Epoch [1/3], Step [1108/3236], Loss: 2.5224, Perplexity: 12.4585
Epoch [1/3], Step [1109/3236], Loss: 2.4910, Perplexity: 12.0730
Epoch [1/3], Step [1110/3236], Loss: 2.4224, Perplexity: 11.2733
Epoch [1/3], Step [1111/3236], Loss: 2.4874, Perplexity: 12.0297
Epoch [1/3], Step [1112/3236], Loss: 2.6629, Perplexity: 14.3376
Epoch [1/3], Step [1113/3236], Loss: 2.6117, Perplexity: 13.6216
Epoch [1/3], Step [1114/3236], Loss: 2.5824, Perplexity: 13.2288
Epoch [1/3], Step [1115/3236], Loss: 2.4666, Perplexity: 11.7823
Epoch [1/3], Step [1116/3236], Loss: 2.4716, Perplexity: 11.8412
Epoch [1/3], Step [1117/3236], Loss: 2.4836, Perplexity: 11.9843
Epoch [1/3], Step [1118/3236], Loss: 2.4300, Perplexity: 11.3592
Epoch [1/3], Step [1119/3236], Loss: 2.4335, Perplexity: 11.3982
Epoch [1/3], Step [1120/3236], Loss: 2.5941, Perplexity: 13.3848
Epoch [1/3], Step [1121/3236], Loss: 2.4052, Perplexity: 11.0810
Epoch [1/3], Step [1122/3236], Loss: 2.5130, Perplexity: 12.3416
Epoch [1/3], Step [1123/3236], Loss: 2.7248, Perplexity: 15.2540
Epoch [1/3], Step [1124/3236], Loss: 2.8976, Perplexity: 18.1306
Epoch [1/3], Step [1125/3236], Loss: 3.1512, Perplexity: 23.3646
Epoch [1/3], Step [1126/3236], Loss: 2.8136, Perplexity: 16.6693
Epoch [1/3], Step [1127/3236], Loss: 2.8925, Perplexity: 18.0384
Epoch [1/3], Step [1128/3236], Loss: 2.9166, Perplexity: 18.4776
Epoch [1/3], Step [1129/3236], Loss: 2.9964, Perplexity: 20.0143
Epoch [1/3], Step [1130/3236], Loss: 2.5279, Perplexity: 12.5266
Epoch [1/3], Step [1131/3236], Loss: 2.5352, Perplexity: 12.6190
Epoch [1/3], Step [1132/3236], Loss: 2.6096, Perplexity: 13.5932
Epoch [1/3], Step [1133/3236], Loss: 2.6977, Perplexity: 14.8450
Epoch [1/3], Step [1134/3236], Loss: 2.8959, Perplexity: 18.0997
Epoch [1/3], Step [1135/3236], Loss: 2.6010, Perplexity: 13.4773
Epoch [1/3], Step [1136/3236], Loss: 2.6230, Perplexity: 13.7775
Epoch [1/3], Step [1137/3236], Loss: 2.6683, Perplexity: 14.4160
Epoch [1/3], Step [1138/3236], Loss: 2.6294, Perplexity: 13.8648
Epoch [1/3], Step [1139/3236], Loss: 2.6575, Perplexity: 14.2599
Epoch [1/3], Step [1140/3236], Loss: 2.5280, Perplexity: 12.5289
Epoch [1/3], Step [1141/3236], Loss: 2.5476, Perplexity: 12.7769
Epoch [1/3], Step [1142/3236], Loss: 2.5524, Perplexity: 12.8378
Epoch [1/3], Step [1143/3236], Loss: 2.5292, Perplexity: 12.5433
Epoch [1/3], Step [1144/3236], Loss: 3.2343, Perplexity: 25.3898
Epoch [1/3], Step [1145/3236], Loss: 2.4848, Perplexity: 11.9988
Epoch [1/3], Step [1146/3236], Loss: 2.6170, Perplexity: 13.6948
Epoch [1/3], Step [1147/3236], Loss: 2.5936, Perplexity: 13.3775
Epoch [1/3], Step [1148/3236], Loss: 2.4144, Perplexity: 11.1825
Epoch [1/3], Step [1149/3236], Loss: 3.0535, Perplexity: 21.1899
Epoch [1/3], Step [1150/3236], Loss: 2.5166, Perplexity: 12.3861
Epoch [1/3], Step [1151/3236], Loss: 2.5062, Perplexity: 12.2580
Epoch [1/3], Step [1152/3236], Loss: 3.0589, Perplexity: 21.3050
Epoch [1/3], Step [1153/3236], Loss: 2.5684, Perplexity: 13.0450
Epoch [1/3], Step [1154/3236], Loss: 2.4976, Perplexity: 12.1536
Epoch [1/3], Step [1155/3236], Loss: 2.6106, Perplexity: 13.6075
Epoch [1/3], Step [1156/3236], Loss: 2.4734, Perplexity: 11.8622
Epoch [1/3], Step [1157/3236], Loss: 2.6096, Perplexity: 13.5931
Epoch [1/3], Step [1158/3236], Loss: 3.1110, Perplexity: 22.4438
Epoch [1/3], Step [1159/3236], Loss: 2.5913, Perplexity: 13.3470
Epoch [1/3], Step [1160/3236], Loss: 2.8790, Perplexity: 17.7964
Epoch [1/3], Step [1161/3236], Loss: 2.5499, Perplexity: 12.8060
Epoch [1/3], Step [1162/3236], Loss: 2.5955, Perplexity: 13.4032
Epoch [1/3], Step [1163/3236], Loss: 3.3804, Perplexity: 29.3822
Epoch [1/3], Step [1164/3236], Loss: 2.9762, Perplexity: 19.6133
Epoch [1/3], Step [1165/3236], Loss: 2.7403, Perplexity: 15.4919
Epoch [1/3], Step [1166/3236], Loss: 2.4303, Perplexity: 11.3620
Epoch [1/3], Step [1167/3236], Loss: 2.6529, Perplexity: 14.1945
Epoch [1/3], Step [1168/3236], Loss: 2.6133, Perplexity: 13.6445
Epoch [1/3], Step [1169/3236], Loss: 2.5069, Perplexity: 12.2665
Epoch [1/3], Step [1170/3236], Loss: 2.5879, Perplexity: 13.3020
Epoch [1/3], Step [1171/3236], Loss: 3.0116, Perplexity: 20.3191
Epoch [1/3], Step [1172/3236], Loss: 2.4873, Perplexity: 12.0282
Epoch [1/3], Step [1173/3236], Loss: 2.5968, Perplexity: 13.4202
Epoch [1/3], Step [1174/3236], Loss: 2.3599, Perplexity: 10.5903
Epoch [1/3], Step [1175/3236], Loss: 2.4145, Perplexity: 11.1847
Epoch [1/3], Step [1176/3236], Loss: 2.7356, Perplexity: 15.4198
Epoch [1/3], Step [1177/3236], Loss: 2.5465, Perplexity: 12.7622
Epoch [1/3], Step [1178/3236], Loss: 2.4080, Perplexity: 11.1113
Epoch [1/3], Step [1179/3236], Loss: 2.5998, Perplexity: 13.4611
Epoch [1/3], Step [1180/3236], Loss: 2.5118, Perplexity: 12.3268
Epoch [1/3], Step [1181/3236], Loss: 2.4654, Perplexity: 11.7687
Epoch [1/3], Step [1182/3236], Loss: 2.5034, Perplexity: 12.2239
Epoch [1/3], Step [1183/3236], Loss: 2.6150, Perplexity: 13.6671
Epoch [1/3], Step [1184/3236], Loss: 2.4059, Perplexity: 11.0887
Epoch [1/3], Step [1185/3236], Loss: 2.3912, Perplexity: 10.9270
Epoch [1/3], Step [1186/3236], Loss: 2.4777, Perplexity: 11.9136
Epoch [1/3], Step [1187/3236], Loss: 2.4570, Perplexity: 11.6697
Epoch [1/3], Step [1188/3236], Loss: 2.6257, Perplexity: 13.8142
Epoch [1/3], Step [1189/3236], Loss: 2.5762, Perplexity: 13.1474
Epoch [1/3], Step [1190/3236], Loss: 2.5096, Perplexity: 12.2997
Epoch [1/3], Step [1191/3236], Loss: 3.3428, Perplexity: 28.2969
Epoch [1/3], Step [1192/3236], Loss: 2.6822, Perplexity: 14.6172
Epoch [1/3], Step [1193/3236], Loss: 2.3956, Perplexity: 10.9745
Epoch [1/3], Step [1194/3236], Loss: 2.4407, Perplexity: 11.4813
Epoch [1/3], Step [1195/3236], Loss: 2.3850, Perplexity: 10.8588
Epoch [1/3], Step [1196/3236], Loss: 2.9079, Perplexity: 18.3189
Epoch [1/3], Step [1197/3236], Loss: 2.5309, Perplexity: 12.5645
Epoch [1/3], Step [1198/3236], Loss: 2.3765, Perplexity: 10.7668
Epoch [1/3], Step [1199/3236], Loss: 2.4584, Perplexity: 11.6861
Epoch [1/3], Step [1200/3236], Loss: 2.5187, Perplexity: 12.4119
Epoch [1/3], Step [1201/3236], Loss: 2.6974, Perplexity: 14.8408
Epoch [1/3], Step [1202/3236], Loss: 2.7125, Perplexity: 15.0668
Epoch [1/3], Step [1203/3236], Loss: 3.2806, Perplexity: 26.5924
Epoch [1/3], Step [1204/3236], Loss: 2.7173, Perplexity: 15.1397
Epoch [1/3], Step [1205/3236], Loss: 2.7601, Perplexity: 15.8010
Epoch [1/3], Step [1206/3236], Loss: 2.3760, Perplexity: 10.7618
Epoch [1/3], Step [1207/3236], Loss: 2.3886, Perplexity: 10.8977
Epoch [1/3], Step [1208/3236], Loss: 2.4728, Perplexity: 11.8550
Epoch [1/3], Step [1209/3236], Loss: 2.3915, Perplexity: 10.9295
Epoch [1/3], Step [1210/3236], Loss: 2.3429, Perplexity: 10.4110
Epoch [1/3], Step [1211/3236], Loss: 2.5622, Perplexity: 12.9643
Epoch [1/3], Step [1212/3236], Loss: 3.2339, Perplexity: 25.3791
Epoch [1/3], Step [1213/3236], Loss: 2.4540, Perplexity: 11.6345
Epoch [1/3], Step [1214/3236], Loss: 2.4318, Perplexity: 11.3795
Epoch [1/3], Step [1215/3236], Loss: 2.3480, Perplexity: 10.4645
Epoch [1/3], Step [1216/3236], Loss: 2.5032, Perplexity: 12.2211
Epoch [1/3], Step [1217/3236], Loss: 2.6472, Perplexity: 14.1145
Epoch [1/3], Step [1218/3236], Loss: 2.4557, Perplexity: 11.6547
Epoch [1/3], Step [1219/3236], Loss: 2.2963, Perplexity: 9.9371
Epoch [1/3], Step [1220/3236], Loss: 2.4446, Perplexity: 11.5256
Epoch [1/3], Step [1221/3236], Loss: 2.3904, Perplexity: 10.9181
Epoch [1/3], Step [1222/3236], Loss: 2.4277, Perplexity: 11.3327
Epoch [1/3], Step [1223/3236], Loss: 2.3960, Perplexity: 10.9793
Epoch [1/3], Step [1224/3236], Loss: 2.3912, Perplexity: 10.9267
Epoch [1/3], Step [1225/3236], Loss: 2.2804, Perplexity: 9.7801
Epoch [1/3], Step [1226/3236], Loss: 2.3988, Perplexity: 11.0099
Epoch [1/3], Step [1227/3236], Loss: 2.5237, Perplexity: 12.4751
Epoch [1/3], Step [1228/3236], Loss: 2.4048, Perplexity: 11.0760
Epoch [1/3], Step [1229/3236], Loss: 2.2548, Perplexity: 9.5337
Epoch [1/3], Step [1230/3236], Loss: 2.6206, Perplexity: 13.7439
Epoch [1/3], Step [1231/3236], Loss: 2.4685, Perplexity: 11.8045
Epoch [1/3], Step [1232/3236], Loss: 2.4947, Perplexity: 12.1183
Epoch [1/3], Step [1233/3236], Loss: 2.4431, Perplexity: 11.5081
Epoch [1/3], Step [1234/3236], Loss: 2.3973, Perplexity: 10.9933
Epoch [1/3], Step [1235/3236], Loss: 2.9359, Perplexity: 18.8383
Epoch [1/3], Step [1236/3236], Loss: 2.4559, Perplexity: 11.6566
Epoch [1/3], Step [1237/3236], Loss: 2.5083, Perplexity: 12.2843
Epoch [1/3], Step [1238/3236], Loss: 2.3345, Perplexity: 10.3244
Epoch [1/3], Step [1239/3236], Loss: 2.3802, Perplexity: 10.8074
Epoch [1/3], Step [1240/3236], Loss: 2.5366, Perplexity: 12.6361
Epoch [1/3], Step [1241/3236], Loss: 2.4964, Perplexity: 12.1387
Epoch [1/3], Step [1242/3236], Loss: 2.4655, Perplexity: 11.7699
Epoch [1/3], Step [1243/3236], Loss: 2.4689, Perplexity: 11.8095
Epoch [1/3], Step [1244/3236], Loss: 2.4367, Perplexity: 11.4355
Epoch [1/3], Step [1245/3236], Loss: 2.4659, Perplexity: 11.7742
Epoch [1/3], Step [1246/3236], Loss: 2.5726, Perplexity: 13.0998
Epoch [1/3], Step [1247/3236], Loss: 2.4072, Perplexity: 11.1025
Epoch [1/3], Step [1248/3236], Loss: 2.8274, Perplexity: 16.9017
Epoch [1/3], Step [1249/3236], Loss: 2.3042, Perplexity: 10.0158
Epoch [1/3], Step [1250/3236], Loss: 2.3644, Perplexity: 10.6375
Epoch [1/3], Step [1251/3236], Loss: 2.4823, Perplexity: 11.9691
Epoch [1/3], Step [1252/3236], Loss: 2.5699, Perplexity: 13.0646
Epoch [1/3], Step [1253/3236], Loss: 2.5047, Perplexity: 12.2401
Epoch [1/3], Step [1254/3236], Loss: 2.5076, Perplexity: 12.2749
Epoch [1/3], Step [1255/3236], Loss: 2.3749, Perplexity: 10.7504
Epoch [1/3], Step [1256/3236], Loss: 2.5072, Perplexity: 12.2708
Epoch [1/3], Step [1257/3236], Loss: 2.3621, Perplexity: 10.6131
Epoch [1/3], Step [1258/3236], Loss: 2.5091, Perplexity: 12.2939
Epoch [1/3], Step [1259/3236], Loss: 2.4397, Perplexity: 11.4693
Epoch [1/3], Step [1260/3236], Loss: 2.7261, Perplexity: 15.2725
Epoch [1/3], Step [1261/3236], Loss: 3.4165, Perplexity: 30.4615
Epoch [1/3], Step [1262/3236], Loss: 2.7535, Perplexity: 15.6980
Epoch [1/3], Step [1263/3236], Loss: 2.4182, Perplexity: 11.2256
Epoch [1/3], Step [1264/3236], Loss: 2.3320, Perplexity: 10.2982
Epoch [1/3], Step [1265/3236], Loss: 2.8927, Perplexity: 18.0426
Epoch [1/3], Step [1266/3236], Loss: 2.3741, Perplexity: 10.7415
Epoch [1/3], Step [1267/3236], Loss: 2.3735, Perplexity: 10.7353
Epoch [1/3], Step [1268/3236], Loss: 2.3325, Perplexity: 10.3033
Epoch [1/3], Step [1269/3236], Loss: 2.4101, Perplexity: 11.1356
Epoch [1/3], Step [1270/3236], Loss: 2.5100, Perplexity: 12.3055
Epoch [1/3], Step [1271/3236], Loss: 2.6222, Perplexity: 13.7665
Epoch [1/3], Step [1272/3236], Loss: 2.5080, Perplexity: 12.2799
Epoch [1/3], Step [1273/3236], Loss: 2.4562, Perplexity: 11.6601
Epoch [1/3], Step [1274/3236], Loss: 2.4542, Perplexity: 11.6370
Epoch [1/3], Step [1275/3236], Loss: 2.3869, Perplexity: 10.8798
Epoch [1/3], Step [1276/3236], Loss: 2.4683, Perplexity: 11.8024
Epoch [1/3], Step [1277/3236], Loss: 2.3936, Perplexity: 10.9526
Epoch [1/3], Step [1278/3236], Loss: 2.4448, Perplexity: 11.5288
Epoch [1/3], Step [1279/3236], Loss: 2.5991, Perplexity: 13.4514
Epoch [1/3], Step [1280/3236], Loss: 2.3785, Perplexity: 10.7892
Epoch [1/3], Step [1281/3236], Loss: 2.3928, Perplexity: 10.9443
Epoch [1/3], Step [1282/3236], Loss: 2.3458, Perplexity: 10.4413
Epoch [1/3], Step [1283/3236], Loss: 2.5176, Perplexity: 12.3992
Epoch [1/3], Step [1284/3236], Loss: 2.6121, Perplexity: 13.6281
Epoch [1/3], Step [1285/3236], Loss: 2.4059, Perplexity: 11.0880
Epoch [1/3], Step [1286/3236], Loss: 2.3990, Perplexity: 11.0123
Epoch [1/3], Step [1287/3236], Loss: 2.6697, Perplexity: 14.4352
Epoch [1/3], Step [1288/3236], Loss: 2.3022, Perplexity: 9.9961
Epoch [1/3], Step [1289/3236], Loss: 2.3805, Perplexity: 10.8106
Epoch [1/3], Step [1290/3236], Loss: 2.5194, Perplexity: 12.4206
Epoch [1/3], Step [1291/3236], Loss: 2.3777, Perplexity: 10.7797
Epoch [1/3], Step [1292/3236], Loss: 2.6271, Perplexity: 13.8331
Epoch [1/3], Step [1293/3236], Loss: 2.5596, Perplexity: 12.9305
Epoch [1/3], Step [1294/3236], Loss: 2.3045, Perplexity: 10.0189
Epoch [1/3], Step [1295/3236], Loss: 2.4291, Perplexity: 11.3487
Epoch [1/3], Step [1296/3236], Loss: 2.4750, Perplexity: 11.8820
Epoch [1/3], Step [1297/3236], Loss: 2.3228, Perplexity: 10.2043
Epoch [1/3], Step [1298/3236], Loss: 3.1602, Perplexity: 23.5749
Epoch [1/3], Step [1299/3236], Loss: 2.4024, Perplexity: 11.0499
Epoch [1/3], Step [1300/3236], Loss: 2.5697, Perplexity: 13.0615
Epoch [1/3], Step [1301/3236], Loss: 2.4803, Perplexity: 11.9452
Epoch [1/3], Step [1302/3236], Loss: 3.3690, Perplexity: 29.0492
Epoch [1/3], Step [1303/3236], Loss: 2.6882, Perplexity: 14.7048
Epoch [1/3], Step [1304/3236], Loss: 2.5382, Perplexity: 12.6569
Epoch [1/3], Step [1305/3236], Loss: 2.5829, Perplexity: 13.2358
Epoch [1/3], Step [1306/3236], Loss: 2.8734, Perplexity: 17.6963
Epoch [1/3], Step [1307/3236], Loss: 2.4293, Perplexity: 11.3507
Epoch [1/3], Step [1308/3236], Loss: 2.5602, Perplexity: 12.9378
Epoch [1/3], Step [1309/3236], Loss: 2.3807, Perplexity: 10.8121
Epoch [1/3], Step [1310/3236], Loss: 2.4478, Perplexity: 11.5628
Epoch [1/3], Step [1311/3236], Loss: 2.4120, Perplexity: 11.1563
Epoch [1/3], Step [1312/3236], Loss: 2.4441, Perplexity: 11.5197
Epoch [1/3], Step [1313/3236], Loss: 2.4881, Perplexity: 12.0383
Epoch [1/3], Step [1314/3236], Loss: 2.4449, Perplexity: 11.5297
Epoch [1/3], Step [1315/3236], Loss: 2.7035, Perplexity: 14.9314
Epoch [1/3], Step [1316/3236], Loss: 2.5284, Perplexity: 12.5337
Epoch [1/3], Step [1317/3236], Loss: 2.4065, Perplexity: 11.0947
Epoch [1/3], Step [1318/3236], Loss: 2.4573, Perplexity: 11.6736
Epoch [1/3], Step [1319/3236], Loss: 2.4930, Perplexity: 12.0971
Epoch [1/3], Step [1320/3236], Loss: 2.6211, Perplexity: 13.7510
Epoch [1/3], Step [1321/3236], Loss: 2.6213, Perplexity: 13.7532
Epoch [1/3], Step [1322/3236], Loss: 2.3794, Perplexity: 10.7979
Epoch [1/3], Step [1323/3236], Loss: 2.6090, Perplexity: 13.5857
Epoch [1/3], Step [1324/3236], Loss: 2.4444, Perplexity: 11.5237
Epoch [1/3], Step [1325/3236], Loss: 2.5469, Perplexity: 12.7677
Epoch [1/3], Step [1326/3236], Loss: 2.3085, Perplexity: 10.0598
Epoch [1/3], Step [1327/3236], Loss: 2.4292, Perplexity: 11.3503
Epoch [1/3], Step [1328/3236], Loss: 2.9271, Perplexity: 18.6731
Epoch [1/3], Step [1329/3236], Loss: 2.3769, Perplexity: 10.7712
Epoch [1/3], Step [1330/3236], Loss: 2.5102, Perplexity: 12.3078
Epoch [1/3], Step [1331/3236], Loss: 2.7338, Perplexity: 15.3919
Epoch [1/3], Step [1332/3236], Loss: 2.4612, Perplexity: 11.7189
Epoch [1/3], Step [1333/3236], Loss: 2.6573, Perplexity: 14.2579
Epoch [1/3], Step [1334/3236], Loss: 2.4948, Perplexity: 12.1192
Epoch [1/3], Step [1335/3236], Loss: 2.3064, Perplexity: 10.0382
Epoch [1/3], Step [1336/3236], Loss: 2.8542, Perplexity: 17.3610
Epoch [1/3], Step [1337/3236], Loss: 2.4154, Perplexity: 11.1944
Epoch [1/3], Step [1338/3236], Loss: 2.6902, Perplexity: 14.7340
Epoch [1/3], Step [1339/3236], Loss: 2.3891, Perplexity: 10.9038
Epoch [1/3], Step [1340/3236], Loss: 2.3238, Perplexity: 10.2149
Epoch [1/3], Step [1341/3236], Loss: 2.7613, Perplexity: 15.8204
Epoch [1/3], Step [1342/3236], Loss: 2.6696, Perplexity: 14.4345
Epoch [1/3], Step [1343/3236], Loss: 2.4680, Perplexity: 11.7990
Epoch [1/3], Step [1344/3236], Loss: 2.3778, Perplexity: 10.7813
Epoch [1/3], Step [1345/3236], Loss: 2.5774, Perplexity: 13.1627
Epoch [1/3], Step [1346/3236], Loss: 2.5138, Perplexity: 12.3522
Epoch [1/3], Step [1347/3236], Loss: 3.2156, Perplexity: 24.9183
Epoch [1/3], Step [1348/3236], Loss: 2.5752, Perplexity: 13.1339
Epoch [1/3], Step [1349/3236], Loss: 2.3729, Perplexity: 10.7279
Epoch [1/3], Step [1350/3236], Loss: 2.4702, Perplexity: 11.8243
Epoch [1/3], Step [1351/3236], Loss: 2.8505, Perplexity: 17.2960
Epoch [1/3], Step [1352/3236], Loss: 2.2798, Perplexity: 9.7745
Epoch [1/3], Step [1353/3236], Loss: 2.5200, Perplexity: 12.4290
Epoch [1/3], Step [1354/3236], Loss: 2.4965, Perplexity: 12.1395
Epoch [1/3], Step [1355/3236], Loss: 2.3898, Perplexity: 10.9111
Epoch [1/3], Step [1356/3236], Loss: 2.4252, Perplexity: 11.3047
Epoch [1/3], Step [1357/3236], Loss: 2.5632, Perplexity: 12.9775
Epoch [1/3], Step [1358/3236], Loss: 2.3804, Perplexity: 10.8092
Epoch [1/3], Step [1359/3236], Loss: 2.2227, Perplexity: 9.2325
Epoch [1/3], Step [1360/3236], Loss: 2.3870, Perplexity: 10.8809
Epoch [1/3], Step [1361/3236], Loss: 2.4141, Perplexity: 11.1800
Epoch [1/3], Step [1362/3236], Loss: 2.4375, Perplexity: 11.4448
Epoch [1/3], Step [1363/3236], Loss: 2.6977, Perplexity: 14.8457
Epoch [1/3], Step [1364/3236], Loss: 2.5384, Perplexity: 12.6588
Epoch [1/3], Step [1365/3236], Loss: 2.3903, Perplexity: 10.9167
Epoch [1/3], Step [1366/3236], Loss: 2.6465, Perplexity: 14.1048
Epoch [1/3], Step [1367/3236], Loss: 2.4721, Perplexity: 11.8474
Epoch [1/3], Step [1368/3236], Loss: 2.4367, Perplexity: 11.4348
Epoch [1/3], Step [1369/3236], Loss: 2.9989, Perplexity: 20.0632
Epoch [1/3], Step [1370/3236], Loss: 2.4918, Perplexity: 12.0832
Epoch [1/3], Step [1371/3236], Loss: 2.5683, Perplexity: 13.0434
Epoch [1/3], Step [1372/3236], Loss: 2.4959, Perplexity: 12.1324
Epoch [1/3], Step [1373/3236], Loss: 2.3907, Perplexity: 10.9212
Epoch [1/3], Step [1374/3236], Loss: 2.5160, Perplexity: 12.3787
Epoch [1/3], Step [1375/3236], Loss: 2.5516, Perplexity: 12.8272
Epoch [1/3], Step [1376/3236], Loss: 2.5833, Perplexity: 13.2403
Epoch [1/3], Step [1377/3236], Loss: 2.6115, Perplexity: 13.6200
Epoch [1/3], Step [1378/3236], Loss: 2.6236, Perplexity: 13.7852
Epoch [1/3], Step [1379/3236], Loss: 2.4833, Perplexity: 11.9810
Epoch [1/3], Step [1380/3236], Loss: 2.7701, Perplexity: 15.9599
Epoch [1/3], Step [1381/3236], Loss: 2.2612, Perplexity: 9.5951
Epoch [1/3], Step [1382/3236], Loss: 2.6029, Perplexity: 13.5032
Epoch [1/3], Step [1383/3236], Loss: 2.3711, Perplexity: 10.7091
Epoch [1/3], Step [1384/3236], Loss: 2.2590, Perplexity: 9.5731
Epoch [1/3], Step [1385/3236], Loss: 2.4607, Perplexity: 11.7129
Epoch [1/3], Step [1386/3236], Loss: 2.3912, Perplexity: 10.9264
Epoch [1/3], Step [1387/3236], Loss: 2.2871, Perplexity: 9.8464
Epoch [1/3], Step [1388/3236], Loss: 2.3903, Perplexity: 10.9169
Epoch [1/3], Step [1389/3236], Loss: 2.4534, Perplexity: 11.6274
Epoch [1/3], Step [1390/3236], Loss: 2.3888, Perplexity: 10.9009
Epoch [1/3], Step [1391/3236], Loss: 3.1130, Perplexity: 22.4877
Epoch [1/3], Step [1392/3236], Loss: 2.4469, Perplexity: 11.5526
Epoch [1/3], Step [1393/3236], Loss: 2.6533, Perplexity: 14.2013
Epoch [1/3], Step [1394/3236], Loss: 2.3957, Perplexity: 10.9762
Epoch [1/3], Step [1395/3236], Loss: 2.4796, Perplexity: 11.9363
Epoch [1/3], Step [1396/3236], Loss: 2.3489, Perplexity: 10.4739
Epoch [1/3], Step [1397/3236], Loss: 3.1226, Perplexity: 22.7060
Epoch [1/3], Step [1398/3236], Loss: 2.5297, Perplexity: 12.5493
Epoch [1/3], Step [1399/3236], Loss: 2.7524, Perplexity: 15.6800
Epoch [1/3], Step [1400/3236], Loss: 2.2405, Perplexity: 9.3984
Epoch [1/3], Step [1401/3236], Loss: 2.5094, Perplexity: 12.2981
Epoch [1/3], Step [1402/3236], Loss: 2.3410, Perplexity: 10.3913
Epoch [1/3], Step [1403/3236], Loss: 2.3615, Perplexity: 10.6071
Epoch [1/3], Step [1404/3236], Loss: 2.3763, Perplexity: 10.7649
Epoch [1/3], Step [1405/3236], Loss: 2.4865, Perplexity: 12.0186
Epoch [1/3], Step [1406/3236], Loss: 2.3411, Perplexity: 10.3923
Epoch [1/3], Step [1407/3236], Loss: 2.3886, Perplexity: 10.8982
Epoch [1/3], Step [1408/3236], Loss: 2.3769, Perplexity: 10.7719
Epoch [1/3], Step [1409/3236], Loss: 2.3602, Perplexity: 10.5932
Epoch [1/3], Step [1410/3236], Loss: 2.5073, Perplexity: 12.2712
Epoch [1/3], Step [1411/3236], Loss: 2.4204, Perplexity: 11.2501
Epoch [1/3], Step [1412/3236], Loss: 2.5120, Perplexity: 12.3292
Epoch [1/3], Step [1413/3236], Loss: 2.4572, Perplexity: 11.6719
Epoch [1/3], Step [1414/3236], Loss: 2.5134, Perplexity: 12.3469
Epoch [1/3], Step [1415/3236], Loss: 2.3535, Perplexity: 10.5224
Epoch [1/3], Step [1416/3236], Loss: 2.3459, Perplexity: 10.4425
Epoch [1/3], Step [1417/3236], Loss: 2.5389, Perplexity: 12.6658
Epoch [1/3], Step [1418/3236], Loss: 2.5533, Perplexity: 12.8491
Epoch [1/3], Step [1419/3236], Loss: 2.4291, Perplexity: 11.3484
Epoch [1/3], Step [1420/3236], Loss: 2.3469, Perplexity: 10.4535
Epoch [1/3], Step [1421/3236], Loss: 3.0480, Perplexity: 21.0726
Epoch [1/3], Step [1422/3236], Loss: 2.3453, Perplexity: 10.4368
Epoch [1/3], Step [1423/3236], Loss: 2.2094, Perplexity: 9.1100
Epoch [1/3], Step [1424/3236], Loss: 2.3068, Perplexity: 10.0425
Epoch [1/3], Step [1425/3236], Loss: 2.4162, Perplexity: 11.2029
Epoch [1/3], Step [1426/3236], Loss: 2.3655, Perplexity: 10.6494
Epoch [1/3], Step [1427/3236], Loss: 2.5099, Perplexity: 12.3038
Epoch [1/3], Step [1428/3236], Loss: 2.4576, Perplexity: 11.6763
Epoch [1/3], Step [1429/3236], Loss: 2.2624, Perplexity: 9.6057
Epoch [1/3], Step [1430/3236], Loss: 2.3334, Perplexity: 10.3128
Epoch [1/3], Step [1431/3236], Loss: 2.8551, Perplexity: 17.3759
Epoch [1/3], Step [1432/3236], Loss: 2.3491, Perplexity: 10.4764
Epoch [1/3], Step [1433/3236], Loss: 2.7984, Perplexity: 16.4191
Epoch [1/3], Step [1434/3236], Loss: 2.3939, Perplexity: 10.9559
Epoch [1/3], Step [1435/3236], Loss: 2.2984, Perplexity: 9.9579
Epoch [1/3], Step [1436/3236], Loss: 2.3592, Perplexity: 10.5822
Epoch [1/3], Step [1437/3236], Loss: 2.4731, Perplexity: 11.8594
Epoch [1/3], Step [1438/3236], Loss: 2.4432, Perplexity: 11.5103
Epoch [1/3], Step [1439/3236], Loss: 2.1903, Perplexity: 8.9382
Epoch [1/3], Step [1440/3236], Loss: 2.6774, Perplexity: 14.5466
Epoch [1/3], Step [1441/3236], Loss: 2.8710, Perplexity: 17.6553
Epoch [1/3], Step [1442/3236], Loss: 2.3037, Perplexity: 10.0110
Epoch [1/3], Step [1443/3236], Loss: 2.2605, Perplexity: 9.5877
Epoch [1/3], Step [1444/3236], Loss: 3.1194, Perplexity: 22.6338
Epoch [1/3], Step [1445/3236], Loss: 2.2260, Perplexity: 9.2629
Epoch [1/3], Step [1446/3236], Loss: 2.2439, Perplexity: 9.4303
Epoch [1/3], Step [1447/3236], Loss: 2.5863, Perplexity: 13.2807
Epoch [1/3], Step [1448/3236], Loss: 2.3554, Perplexity: 10.5429
Epoch [1/3], Step [1449/3236], Loss: 2.2482, Perplexity: 9.4708
Epoch [1/3], Step [1450/3236], Loss: 2.5738, Perplexity: 13.1155
Epoch [1/3], Step [1451/3236], Loss: 2.4746, Perplexity: 11.8765
Epoch [1/3], Step [1452/3236], Loss: 2.3241, Perplexity: 10.2176
Epoch [1/3], Step [1453/3236], Loss: 2.5087, Perplexity: 12.2892
Epoch [1/3], Step [1454/3236], Loss: 2.9140, Perplexity: 18.4301
Epoch [1/3], Step [1455/3236], Loss: 2.5097, Perplexity: 12.3007
Epoch [1/3], Step [1456/3236], Loss: 2.4818, Perplexity: 11.9623
Epoch [1/3], Step [1457/3236], Loss: 2.4229, Perplexity: 11.2788
Epoch [1/3], Step [1458/3236], Loss: 2.4448, Perplexity: 11.5288
Epoch [1/3], Step [1459/3236], Loss: 2.3574, Perplexity: 10.5632
Epoch [1/3], Step [1460/3236], Loss: 2.5468, Perplexity: 12.7657
Epoch [1/3], Step [1461/3236], Loss: 2.5319, Perplexity: 12.5774
Epoch [1/3], Step [1462/3236], Loss: 2.3885, Perplexity: 10.8976
Epoch [1/3], Step [1463/3236], Loss: 2.6852, Perplexity: 14.6610
Epoch [1/3], Step [1464/3236], Loss: 2.6472, Perplexity: 14.1146
Epoch [1/3], Step [1465/3236], Loss: 3.2282, Perplexity: 25.2354
Epoch [1/3], Step [1466/3236], Loss: 2.4491, Perplexity: 11.5778
Epoch [1/3], Step [1467/3236], Loss: 2.5376, Perplexity: 12.6488
Epoch [1/3], Step [1468/3236], Loss: 2.4517, Perplexity: 11.6083
Epoch [1/3], Step [1469/3236], Loss: 2.4920, Perplexity: 12.0849
Epoch [1/3], Step [1470/3236], Loss: 2.4206, Perplexity: 11.2529
Epoch [1/3], Step [1471/3236], Loss: 2.4855, Perplexity: 12.0073
Epoch [1/3], Step [1472/3236], Loss: 2.4657, Perplexity: 11.7716
Epoch [1/3], Step [1473/3236], Loss: 2.7405, Perplexity: 15.4945
Epoch [1/3], Step [1474/3236], Loss: 2.3398, Perplexity: 10.3787
Epoch [1/3], Step [1475/3236], Loss: 2.5259, Perplexity: 12.5024
Epoch [1/3], Step [1476/3236], Loss: 2.4982, Perplexity: 12.1605
Epoch [1/3], Step [1477/3236], Loss: 2.4617, Perplexity: 11.7245
Epoch [1/3], Step [1478/3236], Loss: 3.0054, Perplexity: 20.1951
Epoch [1/3], Step [1479/3236], Loss: 2.3222, Perplexity: 10.1977
Epoch [1/3], Step [1480/3236], Loss: 2.3100, Perplexity: 10.0740
Epoch [1/3], Step [1481/3236], Loss: 3.2565, Perplexity: 25.9573
Epoch [1/3], Step [1482/3236], Loss: 2.4688, Perplexity: 11.8081
Epoch [1/3], Step [1483/3236], Loss: 2.5122, Perplexity: 12.3324
Epoch [1/3], Step [1484/3236], Loss: 2.3960, Perplexity: 10.9787
Epoch [1/3], Step [1485/3236], Loss: 2.4489, Perplexity: 11.5758
Epoch [1/3], Step [1486/3236], Loss: 2.4101, Perplexity: 11.1348
Epoch [1/3], Step [1487/3236], Loss: 2.3374, Perplexity: 10.3544
Epoch [1/3], Step [1488/3236], Loss: 2.7845, Perplexity: 16.1917
Epoch [1/3], Step [1489/3236], Loss: 2.5006, Perplexity: 12.1901
Epoch [1/3], Step [1490/3236], Loss: 2.4502, Perplexity: 11.5906
Epoch [1/3], Step [1491/3236], Loss: 2.8496, Perplexity: 17.2806
Epoch [1/3], Step [1492/3236], Loss: 2.3470, Perplexity: 10.4544
Epoch [1/3], Step [1493/3236], Loss: 2.3941, Perplexity: 10.9581
Epoch [1/3], Step [1494/3236], Loss: 3.1812, Perplexity: 24.0760
Epoch [1/3], Step [1495/3236], Loss: 2.3736, Perplexity: 10.7364
Epoch [1/3], Step [1496/3236], Loss: 2.4936, Perplexity: 12.1043
Epoch [1/3], Step [1497/3236], Loss: 2.5112, Perplexity: 12.3201
Epoch [1/3], Step [1498/3236], Loss: 2.3335, Perplexity: 10.3136
Epoch [1/3], Step [1499/3236], Loss: 2.3240, Perplexity: 10.2165
Epoch [1/3], Step [1500/3236], Loss: 2.4900, Perplexity: 12.0609
Epoch [1/3], Step [1501/3236], Loss: 3.1177, Perplexity: 22.5953
Epoch [1/3], Step [1502/3236], Loss: 2.4241, Perplexity: 11.2916
Epoch [1/3], Step [1503/3236], Loss: 2.4020, Perplexity: 11.0448
Epoch [1/3], Step [1504/3236], Loss: 2.5109, Perplexity: 12.3157
Epoch [1/3], Step [1505/3236], Loss: 2.5662, Perplexity: 13.0161
Epoch [1/3], Step [1506/3236], Loss: 2.4429, Perplexity: 11.5060
Epoch [1/3], Step [1507/3236], Loss: 2.3789, Perplexity: 10.7934
Epoch [1/3], Step [1508/3236], Loss: 2.5071, Perplexity: 12.2694
Epoch [1/3], Step [1509/3236], Loss: 2.4750, Perplexity: 11.8821
Epoch [1/3], Step [1510/3236], Loss: 2.4268, Perplexity: 11.3229
Epoch [1/3], Step [1511/3236], Loss: 2.4107, Perplexity: 11.1418
Epoch [1/3], Step [1512/3236], Loss: 2.4997, Perplexity: 12.1791
Epoch [1/3], Step [1513/3236], Loss: 2.3432, Perplexity: 10.4149
Epoch [1/3], Step [1514/3236], Loss: 2.4091, Perplexity: 11.1245
Epoch [1/3], Step [1515/3236], Loss: 2.2595, Perplexity: 9.5778
Epoch [1/3], Step [1516/3236], Loss: 2.2436, Perplexity: 9.4271
Epoch [1/3], Step [1517/3236], Loss: 2.6433, Perplexity: 14.0601
Epoch [1/3], Step [1518/3236], Loss: 2.3819, Perplexity: 10.8249
Epoch [1/3], Step [1519/3236], Loss: 2.5173, Perplexity: 12.3957
Epoch [1/3], Step [1520/3236], Loss: 2.3417, Perplexity: 10.3993
Epoch [1/3], Step [1521/3236], Loss: 2.5438, Perplexity: 12.7277
Epoch [1/3], Step [1522/3236], Loss: 3.0357, Perplexity: 20.8157
Epoch [1/3], Step [1523/3236], Loss: 2.5601, Perplexity: 12.9369
Epoch [1/3], Step [1524/3236], Loss: 3.2015, Perplexity: 24.5692
Epoch [1/3], Step [1525/3236], Loss: 2.4791, Perplexity: 11.9300
Epoch [1/3], Step [1526/3236], Loss: 2.3610, Perplexity: 10.6012
Epoch [1/3], Step [1527/3236], Loss: 2.4758, Perplexity: 11.8910
Epoch [1/3], Step [1528/3236], Loss: 2.4501, Perplexity: 11.5892
Epoch [1/3], Step [1529/3236], Loss: 2.4867, Perplexity: 12.0219
Epoch [1/3], Step [1530/3236], Loss: 2.7422, Perplexity: 15.5204
Epoch [1/3], Step [1531/3236], Loss: 2.3640, Perplexity: 10.6338
Epoch [1/3], Step [1532/3236], Loss: 2.4419, Perplexity: 11.4951
Epoch [1/3], Step [1533/3236], Loss: 2.7536, Perplexity: 15.6993
Epoch [1/3], Step [1534/3236], Loss: 2.1829, Perplexity: 8.8721
Epoch [1/3], Step [1535/3236], Loss: 2.5302, Perplexity: 12.5559
Epoch [1/3], Step [1536/3236], Loss: 2.5227, Perplexity: 12.4625
Epoch [1/3], Step [1537/3236], Loss: 2.2304, Perplexity: 9.3040
Epoch [1/3], Step [1538/3236], Loss: 2.3536, Perplexity: 10.5233
Epoch [1/3], Step [1539/3236], Loss: 2.2939, Perplexity: 9.9140
Epoch [1/3], Step [1540/3236], Loss: 2.3795, Perplexity: 10.7993
Epoch [1/3], Step [1541/3236], Loss: 2.3654, Perplexity: 10.6488
Epoch [1/3], Step [1542/3236], Loss: 3.5865, Perplexity: 36.1080
Epoch [1/3], Step [1543/3236], Loss: 2.1855, Perplexity: 8.8954
Epoch [1/3], Step [1544/3236], Loss: 2.2646, Perplexity: 9.6277
Epoch [1/3], Step [1545/3236], Loss: 2.3716, Perplexity: 10.7144
Epoch [1/3], Step [1546/3236], Loss: 3.0133, Perplexity: 20.3550
Epoch [1/3], Step [1547/3236], Loss: 2.2086, Perplexity: 9.1027
Epoch [1/3], Step [1548/3236], Loss: 2.7001, Perplexity: 14.8816
Epoch [1/3], Step [1549/3236], Loss: 3.0829, Perplexity: 21.8214
Epoch [1/3], Step [1550/3236], Loss: 2.3177, Perplexity: 10.1525
Epoch [1/3], Step [1551/3236], Loss: 2.7683, Perplexity: 15.9322
Epoch [1/3], Step [1552/3236], Loss: 2.4128, Perplexity: 11.1648
Epoch [1/3], Step [1553/3236], Loss: 2.5709, Perplexity: 13.0779
Epoch [1/3], Step [1554/3236], Loss: 2.2698, Perplexity: 9.6776
Epoch [1/3], Step [1555/3236], Loss: 2.6433, Perplexity: 14.0598
Epoch [1/3], Step [1556/3236], Loss: 2.5331, Perplexity: 12.5923
Epoch [1/3], Step [1557/3236], Loss: 2.3888, Perplexity: 10.9004
Epoch [1/3], Step [1558/3236], Loss: 2.3140, Perplexity: 10.1150
Epoch [1/3], Step [1559/3236], Loss: 2.5072, Perplexity: 12.2701
Epoch [1/3], Step [1560/3236], Loss: 2.4344, Perplexity: 11.4095
Epoch [1/3], Step [1561/3236], Loss: 2.3780, Perplexity: 10.7834
Epoch [1/3], Step [1562/3236], Loss: 2.5030, Perplexity: 12.2188
Epoch [1/3], Step [1563/3236], Loss: 2.3525, Perplexity: 10.5123
Epoch [1/3], Step [1564/3236], Loss: 2.5800, Perplexity: 13.1970
Epoch [1/3], Step [1565/3236], Loss: 2.5296, Perplexity: 12.5484
Epoch [1/3], Step [1566/3236], Loss: 2.2717, Perplexity: 9.6958
Epoch [1/3], Step [1567/3236], Loss: 2.3497, Perplexity: 10.4825
Epoch [1/3], Step [1568/3236], Loss: 2.3494, Perplexity: 10.4788
Epoch [1/3], Step [1569/3236], Loss: 2.3362, Perplexity: 10.3423
Epoch [1/3], Step [1570/3236], Loss: 2.4773, Perplexity: 11.9093
Epoch [1/3], Step [1571/3236], Loss: 2.5520, Perplexity: 12.8328
Epoch [1/3], Step [1572/3236], Loss: 2.5210, Perplexity: 12.4414
Epoch [1/3], Step [1573/3236], Loss: 2.4385, Perplexity: 11.4554
Epoch [1/3], Step [1574/3236], Loss: 2.4296, Perplexity: 11.3542
Epoch [1/3], Step [1575/3236], Loss: 2.4089, Perplexity: 11.1217
Epoch [1/3], Step [1576/3236], Loss: 2.6881, Perplexity: 14.7038
Epoch [1/3], Step [1577/3236], Loss: 2.4572, Perplexity: 11.6716
Epoch [1/3], Step [1578/3236], Loss: 2.3518, Perplexity: 10.5041
Epoch [1/3], Step [1579/3236], Loss: 2.3987, Perplexity: 11.0089
Epoch [1/3], Step [1580/3236], Loss: 2.4743, Perplexity: 11.8730
Epoch [1/3], Step [1581/3236], Loss: 2.2848, Perplexity: 9.8242
Epoch [1/3], Step [1582/3236], Loss: 2.3542, Perplexity: 10.5301
Epoch [1/3], Step [1583/3236], Loss: 2.3856, Perplexity: 10.8659
Epoch [1/3], Step [1584/3236], Loss: 2.6439, Perplexity: 14.0677
Epoch [1/3], Step [1585/3236], Loss: 2.5582, Perplexity: 12.9122
Epoch [1/3], Step [1586/3236], Loss: 2.3554, Perplexity: 10.5421
Epoch [1/3], Step [1587/3236], Loss: 2.5997, Perplexity: 13.4598
Epoch [1/3], Step [1588/3236], Loss: 2.2775, Perplexity: 9.7528
Epoch [1/3], Step [1589/3236], Loss: 2.4157, Perplexity: 11.1975
Epoch [1/3], Step [1590/3236], Loss: 2.3726, Perplexity: 10.7251
Epoch [1/3], Step [1591/3236], Loss: 2.5513, Perplexity: 12.8242
Epoch [1/3], Step [1592/3236], Loss: 2.4680, Perplexity: 11.7991
Epoch [1/3], Step [1593/3236], Loss: 2.4311, Perplexity: 11.3711
Epoch [1/3], Step [1594/3236], Loss: 2.8483, Perplexity: 17.2576
Epoch [1/3], Step [1595/3236], Loss: 2.4239, Perplexity: 11.2902
Epoch [1/3], Step [1596/3236], Loss: 2.9501, Perplexity: 19.1075
Epoch [1/3], Step [1597/3236], Loss: 2.7216, Perplexity: 15.2045
Epoch [1/3], Step [1598/3236], Loss: 2.4550, Perplexity: 11.6462
Epoch [1/3], Step [1599/3236], Loss: 2.3642, Perplexity: 10.6352
Epoch [1/3], Step [1600/3236], Loss: 2.5428, Perplexity: 12.7146
Epoch [1/3], Step [1601/3236], Loss: 2.3426, Perplexity: 10.4083
Epoch [1/3], Step [1602/3236], Loss: 2.3422, Perplexity: 10.4044
Epoch [1/3], Step [1603/3236], Loss: 2.5096, Perplexity: 12.2999
Epoch [1/3], Step [1604/3236], Loss: 2.4007, Perplexity: 11.0314
Epoch [1/3], Step [1605/3236], Loss: 2.3460, Perplexity: 10.4438
Epoch [1/3], Step [1606/3236], Loss: 2.7921, Perplexity: 16.3153
Epoch [1/3], Step [1607/3236], Loss: 2.4449, Perplexity: 11.5295
Epoch [1/3], Step [1608/3236], Loss: 2.3749, Perplexity: 10.7504
Epoch [1/3], Step [1609/3236], Loss: 2.2038, Perplexity: 9.0594
Epoch [1/3], Step [1610/3236], Loss: 2.2672, Perplexity: 9.6525
Epoch [1/3], Step [1611/3236], Loss: 2.3763, Perplexity: 10.7649
Epoch [1/3], Step [1612/3236], Loss: 2.5407, Perplexity: 12.6888
Epoch [1/3], Step [1613/3236], Loss: 2.3461, Perplexity: 10.4448
Epoch [1/3], Step [1614/3236], Loss: 2.3739, Perplexity: 10.7392
Epoch [1/3], Step [1615/3236], Loss: 2.2690, Perplexity: 9.6696
Epoch [1/3], Step [1616/3236], Loss: 2.3227, Perplexity: 10.2034
Epoch [1/3], Step [1617/3236], Loss: 2.3614, Perplexity: 10.6061
Epoch [1/3], Step [1618/3236], Loss: 2.2698, Perplexity: 9.6773
Epoch [1/3], Step [1619/3236], Loss: 2.3800, Perplexity: 10.8049
Epoch [1/3], Step [1620/3236], Loss: 2.3361, Perplexity: 10.3407
Epoch [1/3], Step [1621/3236], Loss: 2.3724, Perplexity: 10.7228
Epoch [1/3], Step [1622/3236], Loss: 2.1777, Perplexity: 8.8264
Epoch [1/3], Step [1623/3236], Loss: 2.3347, Perplexity: 10.3259
Epoch [1/3], Step [1624/3236], Loss: 2.5250, Perplexity: 12.4907
Epoch [1/3], Step [1625/3236], Loss: 2.3234, Perplexity: 10.2102
Epoch [1/3], Step [1626/3236], Loss: 2.3699, Perplexity: 10.6968
Epoch [1/3], Step [1627/3236], Loss: 2.4291, Perplexity: 11.3486
Epoch [1/3], Step [1628/3236], Loss: 2.5289, Perplexity: 12.5395
Epoch [1/3], Step [1629/3236], Loss: 2.4105, Perplexity: 11.1398
Epoch [1/3], Step [1630/3236], Loss: 2.5096, Perplexity: 12.3001
Epoch [1/3], Step [1631/3236], Loss: 2.3166, Perplexity: 10.1416
Epoch [1/3], Step [1632/3236], Loss: 2.3330, Perplexity: 10.3085
Epoch [1/3], Step [1633/3236], Loss: 2.2070, Perplexity: 9.0884
Epoch [1/3], Step [1634/3236], Loss: 2.1652, Perplexity: 8.7160
Epoch [1/3], Step [1635/3236], Loss: 2.2930, Perplexity: 9.9046
Epoch [1/3], Step [1636/3236], Loss: 2.2221, Perplexity: 9.2265
Epoch [1/3], Step [1637/3236], Loss: 2.3773, Perplexity: 10.7756
Epoch [1/3], Step [1638/3236], Loss: 2.4610, Perplexity: 11.7166
Epoch [1/3], Step [1639/3236], Loss: 2.4003, Perplexity: 11.0268
Epoch [1/3], Step [1640/3236], Loss: 2.3660, Perplexity: 10.6545
Epoch [1/3], Step [1641/3236], Loss: 2.4071, Perplexity: 11.1022
Epoch [1/3], Step [1642/3236], Loss: 2.4078, Perplexity: 11.1096
Epoch [1/3], Step [1643/3236], Loss: 2.3681, Perplexity: 10.6768
Epoch [1/3], Step [1644/3236], Loss: 2.2209, Perplexity: 9.2158
Epoch [1/3], Step [1645/3236], Loss: 2.2065, Perplexity: 9.0842
Epoch [1/3], Step [1646/3236], Loss: 2.2828, Perplexity: 9.8044
Epoch [1/3], Step [1647/3236], Loss: 2.3229, Perplexity: 10.2052
Epoch [1/3], Step [1648/3236], Loss: 2.4510, Perplexity: 11.5998
Epoch [1/3], Step [1649/3236], Loss: 2.4927, Perplexity: 12.0939
Epoch [1/3], Step [1650/3236], Loss: 2.8255, Perplexity: 16.8694
Epoch [1/3], Step [1651/3236], Loss: 2.4313, Perplexity: 11.3735
Epoch [1/3], Step [1652/3236], Loss: 2.4988, Perplexity: 12.1680
Epoch [1/3], Step [1653/3236], Loss: 2.5525, Perplexity: 12.8397
Epoch [1/3], Step [1654/3236], Loss: 2.4483, Perplexity: 11.5682
Epoch [1/3], Step [1655/3236], Loss: 2.3195, Perplexity: 10.1707
Epoch [1/3], Step [1656/3236], Loss: 2.1684, Perplexity: 8.7447
Epoch [1/3], Step [1657/3236], Loss: 2.5549, Perplexity: 12.8700
Epoch [1/3], Step [1658/3236], Loss: 2.2110, Perplexity: 9.1252
Epoch [1/3], Step [1659/3236], Loss: 2.5632, Perplexity: 12.9776
Epoch [1/3], Step [1660/3236], Loss: 2.3961, Perplexity: 10.9806
Epoch [1/3], Step [1661/3236], Loss: 2.4988, Perplexity: 12.1684
Epoch [1/3], Step [1662/3236], Loss: 2.1921, Perplexity: 8.9539
Epoch [1/3], Step [1663/3236], Loss: 2.4199, Perplexity: 11.2453
Epoch [1/3], Step [1664/3236], Loss: 2.4406, Perplexity: 11.4794
Epoch [1/3], Step [1665/3236], Loss: 2.5362, Perplexity: 12.6317
Epoch [1/3], Step [1666/3236], Loss: 2.4437, Perplexity: 11.5154
Epoch [1/3], Step [1667/3236], Loss: 2.6949, Perplexity: 14.8047
Epoch [1/3], Step [1668/3236], Loss: 2.5856, Perplexity: 13.2716
Epoch [1/3], Step [1669/3236], Loss: 2.4128, Perplexity: 11.1653
Epoch [1/3], Step [1670/3236], Loss: 2.4815, Perplexity: 11.9593
Epoch [1/3], Step [1671/3236], Loss: 2.4190, Perplexity: 11.2345
Epoch [1/3], Step [1672/3236], Loss: 2.2755, Perplexity: 9.7329
Epoch [1/3], Step [1673/3236], Loss: 2.4854, Perplexity: 12.0063
Epoch [1/3], Step [1674/3236], Loss: 2.1647, Perplexity: 8.7119
Epoch [1/3], Step [1675/3236], Loss: 2.3896, Perplexity: 10.9088
Epoch [1/3], Step [1676/3236], Loss: 2.4092, Perplexity: 11.1252
Epoch [1/3], Step [1677/3236], Loss: 2.3448, Perplexity: 10.4309
Epoch [1/3], Step [1678/3236], Loss: 2.2938, Perplexity: 9.9123
Epoch [1/3], Step [1679/3236], Loss: 2.3349, Perplexity: 10.3287
Epoch [1/3], Step [1680/3236], Loss: 2.3587, Perplexity: 10.5767
Epoch [1/3], Step [1681/3236], Loss: 3.2004, Perplexity: 24.5417
Epoch [1/3], Step [1682/3236], Loss: 2.4275, Perplexity: 11.3300
Epoch [1/3], Step [1683/3236], Loss: 2.2073, Perplexity: 9.0914
Epoch [1/3], Step [1684/3236], Loss: 2.4348, Perplexity: 11.4134
Epoch [1/3], Step [1685/3236], Loss: 2.2965, Perplexity: 9.9389
Epoch [1/3], Step [1686/3236], Loss: 2.3072, Perplexity: 10.0458
Epoch [1/3], Step [1687/3236], Loss: 2.3570, Perplexity: 10.5593
Epoch [1/3], Step [1688/3236], Loss: 2.4253, Perplexity: 11.3056
Epoch [1/3], Step [1689/3236], Loss: 2.4351, Perplexity: 11.4171
Epoch [1/3], Step [1690/3236], Loss: 2.4267, Perplexity: 11.3217
Epoch [1/3], Step [1691/3236], Loss: 3.4215, Perplexity: 30.6164
Epoch [1/3], Step [1692/3236], Loss: 2.8985, Perplexity: 18.1474
Epoch [1/3], Step [1693/3236], Loss: 2.2576, Perplexity: 9.5602
Epoch [1/3], Step [1694/3236], Loss: 2.7220, Perplexity: 15.2105
Epoch [1/3], Step [1695/3236], Loss: 2.6968, Perplexity: 14.8323
Epoch [1/3], Step [1696/3236], Loss: 3.0947, Perplexity: 22.0796
Epoch [1/3], Step [1697/3236], Loss: 2.4874, Perplexity: 12.0300
Epoch [1/3], Step [1698/3236], Loss: 2.3363, Perplexity: 10.3434
Epoch [1/3], Step [1699/3236], Loss: 2.3659, Perplexity: 10.6541
Epoch [1/3], Step [1700/3236], Loss: 2.4627, Perplexity: 11.7363
Epoch [1/3], Step [1701/3236], Loss: 2.3052, Perplexity: 10.0262
Epoch [1/3], Step [1702/3236], Loss: 2.2492, Perplexity: 9.4801
Epoch [1/3], Step [1703/3236], Loss: 2.6033, Perplexity: 13.5082
Epoch [1/3], Step [1704/3236], Loss: 2.3236, Perplexity: 10.2120
Epoch [1/3], Step [1705/3236], Loss: 2.3976, Perplexity: 10.9962
Epoch [1/3], Step [1706/3236], Loss: 2.3089, Perplexity: 10.0633
Epoch [1/3], Step [1707/3236], Loss: 2.3893, Perplexity: 10.9058
Epoch [1/3], Step [1708/3236], Loss: 2.2136, Perplexity: 9.1487
Epoch [1/3], Step [1709/3236], Loss: 2.6023, Perplexity: 13.4946
Epoch [1/3], Step [1710/3236], Loss: 3.0428, Perplexity: 20.9630
Epoch [1/3], Step [1711/3236], Loss: 2.3602, Perplexity: 10.5926
Epoch [1/3], Step [1712/3236], Loss: 2.3206, Perplexity: 10.1813
Epoch [1/3], Step [1713/3236], Loss: 2.2543, Perplexity: 9.5284
Epoch [1/3], Step [1714/3236], Loss: 2.4870, Perplexity: 12.0251
Epoch [1/3], Step [1715/3236], Loss: 2.3712, Perplexity: 10.7107
Epoch [1/3], Step [1716/3236], Loss: 2.3814, Perplexity: 10.8205
Epoch [1/3], Step [1717/3236], Loss: 2.3364, Perplexity: 10.3441
Epoch [1/3], Step [1718/3236], Loss: 2.2637, Perplexity: 9.6187
Epoch [1/3], Step [1719/3236], Loss: 2.2892, Perplexity: 9.8666
Epoch [1/3], Step [1720/3236], Loss: 2.3522, Perplexity: 10.5091
Epoch [1/3], Step [1721/3236], Loss: 2.2702, Perplexity: 9.6812
Epoch [1/3], Step [1722/3236], Loss: 2.4038, Perplexity: 11.0655
Epoch [1/3], Step [1723/3236], Loss: 2.4130, Perplexity: 11.1670
Epoch [1/3], Step [1724/3236], Loss: 2.3354, Perplexity: 10.3336
Epoch [1/3], Step [1725/3236], Loss: 2.4415, Perplexity: 11.4898
Epoch [1/3], Step [1726/3236], Loss: 2.2497, Perplexity: 9.4849
Epoch [1/3], Step [1727/3236], Loss: 2.2398, Perplexity: 9.3913
Epoch [1/3], Step [1728/3236], Loss: 2.3106, Perplexity: 10.0804
Epoch [1/3], Step [1729/3236], Loss: 2.2818, Perplexity: 9.7945
Epoch [1/3], Step [1730/3236], Loss: 2.3131, Perplexity: 10.1057
Epoch [1/3], Step [1731/3236], Loss: 2.7082, Perplexity: 15.0018
Epoch [1/3], Step [1732/3236], Loss: 2.3363, Perplexity: 10.3427
Epoch [1/3], Step [1733/3236], Loss: 2.4323, Perplexity: 11.3852
Epoch [1/3], Step [1734/3236], Loss: 2.2731, Perplexity: 9.7094
Epoch [1/3], Step [1735/3236], Loss: 2.4317, Perplexity: 11.3786
Epoch [1/3], Step [1736/3236], Loss: 2.4541, Perplexity: 11.6360
Epoch [1/3], Step [1737/3236], Loss: 2.2542, Perplexity: 9.5273
Epoch [1/3], Step [1738/3236], Loss: 2.2523, Perplexity: 9.5095
Epoch [1/3], Step [1739/3236], Loss: 2.3203, Perplexity: 10.1791
Epoch [1/3], Step [1740/3236], Loss: 2.2161, Perplexity: 9.1715
Epoch [1/3], Step [1741/3236], Loss: 2.6825, Perplexity: 14.6215
Epoch [1/3], Step [1742/3236], Loss: 2.6197, Perplexity: 13.7319
Epoch [1/3], Step [1743/3236], Loss: 2.3956, Perplexity: 10.9751
Epoch [1/3], Step [1744/3236], Loss: 2.6995, Perplexity: 14.8728
Epoch [1/3], Step [1745/3236], Loss: 2.2762, Perplexity: 9.7400
Epoch [1/3], Step [1746/3236], Loss: 2.1824, Perplexity: 8.8678
Epoch [1/3], Step [1747/3236], Loss: 2.4351, Perplexity: 11.4168
Epoch [1/3], Step [1748/3236], Loss: 2.7401, Perplexity: 15.4890
Epoch [1/3], Step [1749/3236], Loss: 2.3160, Perplexity: 10.1353
Epoch [1/3], Step [1750/3236], Loss: 2.3879, Perplexity: 10.8904
Epoch [1/3], Step [1751/3236], Loss: 2.5365, Perplexity: 12.6352
Epoch [1/3], Step [1752/3236], Loss: 2.4042, Perplexity: 11.0690
Epoch [1/3], Step [1753/3236], Loss: 2.2784, Perplexity: 9.7614
Epoch [1/3], Step [1754/3236], Loss: 2.1988, Perplexity: 9.0142
Epoch [1/3], Step [1755/3236], Loss: 3.1517, Perplexity: 23.3749
Epoch [1/3], Step [1756/3236], Loss: 2.4288, Perplexity: 11.3457
Epoch [1/3], Step [1757/3236], Loss: 2.3623, Perplexity: 10.6150
Epoch [1/3], Step [1758/3236], Loss: 2.3664, Perplexity: 10.6590
Epoch [1/3], Step [1759/3236], Loss: 2.7238, Perplexity: 15.2376
Epoch [1/3], Step [1760/3236], Loss: 2.2408, Perplexity: 9.4005
Epoch [1/3], Step [1761/3236], Loss: 2.2053, Perplexity: 9.0727
Epoch [1/3], Step [1762/3236], Loss: 2.1934, Perplexity: 8.9658
Epoch [1/3], Step [1763/3236], Loss: 2.3145, Perplexity: 10.1200
Epoch [1/3], Step [1764/3236], Loss: 2.7013, Perplexity: 14.8993
Epoch [1/3], Step [1765/3236], Loss: 2.9688, Perplexity: 19.4684
Epoch [1/3], Step [1766/3236], Loss: 2.6529, Perplexity: 14.1956
Epoch [1/3], Step [1767/3236], Loss: 2.4152, Perplexity: 11.1916
Epoch [1/3], Step [1768/3236], Loss: 2.5463, Perplexity: 12.7601
Epoch [1/3], Step [1769/3236], Loss: 2.3086, Perplexity: 10.0603
Epoch [1/3], Step [1770/3236], Loss: 2.3303, Perplexity: 10.2809
Epoch [1/3], Step [1771/3236], Loss: 2.3281, Perplexity: 10.2586
Epoch [1/3], Step [1772/3236], Loss: 2.3919, Perplexity: 10.9342
Epoch [1/3], Step [1773/3236], Loss: 2.2624, Perplexity: 9.6061
Epoch [1/3], Step [1774/3236], Loss: 2.3516, Perplexity: 10.5027
Epoch [1/3], Step [1775/3236], Loss: 2.2265, Perplexity: 9.2678
Epoch [1/3], Step [1776/3236], Loss: 2.5795, Perplexity: 13.1910
Epoch [1/3], Step [1777/3236], Loss: 2.2467, Perplexity: 9.4565
Epoch [1/3], Step [1778/3236], Loss: 2.5194, Perplexity: 12.4213
Epoch [1/3], Step [1779/3236], Loss: 2.3933, Perplexity: 10.9496
Epoch [1/3], Step [1780/3236], Loss: 2.8358, Perplexity: 17.0444
Epoch [1/3], Step [1781/3236], Loss: 2.6345, Perplexity: 13.9368
Epoch [1/3], Step [1782/3236], Loss: 2.2344, Perplexity: 9.3411
Epoch [1/3], Step [1783/3236], Loss: 2.2195, Perplexity: 9.2028
Epoch [1/3], Step [1784/3236], Loss: 2.4491, Perplexity: 11.5781
Epoch [1/3], Step [1785/3236], Loss: 2.3920, Perplexity: 10.9349
Epoch [1/3], Step [1786/3236], Loss: 2.2550, Perplexity: 9.5349
Epoch [1/3], Step [1787/3236], Loss: 2.4127, Perplexity: 11.1644
Epoch [1/3], Step [1788/3236], Loss: 2.3788, Perplexity: 10.7921
Epoch [1/3], Step [1789/3236], Loss: 2.5242, Perplexity: 12.4810
Epoch [1/3], Step [1790/3236], Loss: 2.4860, Perplexity: 12.0128
Epoch [1/3], Step [1791/3236], Loss: 3.0504, Perplexity: 21.1243
Epoch [1/3], Step [1792/3236], Loss: 2.5581, Perplexity: 12.9119
Epoch [1/3], Step [1793/3236], Loss: 2.3498, Perplexity: 10.4836
Epoch [1/3], Step [1794/3236], Loss: 2.5590, Perplexity: 12.9232
Epoch [1/3], Step [1795/3236], Loss: 2.2944, Perplexity: 9.9180
Epoch [1/3], Step [1796/3236], Loss: 2.3421, Perplexity: 10.4027
Epoch [1/3], Step [1797/3236], Loss: 2.6556, Perplexity: 14.2335
Epoch [1/3], Step [1798/3236], Loss: 2.5254, Perplexity: 12.4957
Epoch [1/3], Step [1799/3236], Loss: 2.3620, Perplexity: 10.6117
Epoch [1/3], Step [1800/3236], Loss: 2.3923, Perplexity: 10.9387
Epoch [1/3], Step [1801/3236], Loss: 2.5739, Perplexity: 13.1172
Epoch [1/3], Step [1802/3236], Loss: 2.3585, Perplexity: 10.5746
Epoch [1/3], Step [1803/3236], Loss: 2.3540, Perplexity: 10.5271
Epoch [1/3], Step [1804/3236], Loss: 2.3560, Perplexity: 10.5482
Epoch [1/3], Step [1805/3236], Loss: 2.4086, Perplexity: 11.1178
Epoch [1/3], Step [1806/3236], Loss: 2.5316, Perplexity: 12.5733
Epoch [1/3], Step [1807/3236], Loss: 2.3566, Perplexity: 10.5546
Epoch [1/3], Step [1808/3236], Loss: 2.4580, Perplexity: 11.6811
Epoch [1/3], Step [1809/3236], Loss: 2.5194, Perplexity: 12.4215
Epoch [1/3], Step [1810/3236], Loss: 2.2914, Perplexity: 9.8892
Epoch [1/3], Step [1811/3236], Loss: 2.5938, Perplexity: 13.3802
Epoch [1/3], Step [1812/3236], Loss: 2.3414, Perplexity: 10.3956
Epoch [1/3], Step [1813/3236], Loss: 2.2445, Perplexity: 9.4358
Epoch [1/3], Step [1814/3236], Loss: 2.3500, Perplexity: 10.4861
Epoch [1/3], Step [1815/3236], Loss: 2.3318, Perplexity: 10.2966
Epoch [1/3], Step [1816/3236], Loss: 2.3576, Perplexity: 10.5660
Epoch [1/3], Step [1817/3236], Loss: 2.3929, Perplexity: 10.9449
Epoch [1/3], Step [1818/3236], Loss: 2.4207, Perplexity: 11.2536
Epoch [1/3], Step [1819/3236], Loss: 2.3558, Perplexity: 10.5466
Epoch [1/3], Step [1820/3236], Loss: 2.4484, Perplexity: 11.5697
Epoch [1/3], Step [1821/3236], Loss: 2.3763, Perplexity: 10.7652
Epoch [1/3], Step [1822/3236], Loss: 2.3585, Perplexity: 10.5748
Epoch [1/3], Step [1823/3236], Loss: 2.3078, Perplexity: 10.0518
Epoch [1/3], Step [1824/3236], Loss: 2.3801, Perplexity: 10.8056
Epoch [1/3], Step [1825/3236], Loss: 2.1662, Perplexity: 8.7251
Epoch [1/3], Step [1826/3236], Loss: 2.4278, Perplexity: 11.3334
Epoch [1/3], Step [1827/3236], Loss: 2.5237, Perplexity: 12.4750
Epoch [1/3], Step [1828/3236], Loss: 2.3900, Perplexity: 10.9137
Epoch [1/3], Step [1829/3236], Loss: 2.3376, Perplexity: 10.3565
Epoch [1/3], Step [1830/3236], Loss: 2.3651, Perplexity: 10.6456
Epoch [1/3], Step [1831/3236], Loss: 2.8623, Perplexity: 17.5009
Epoch [1/3], Step [1832/3236], Loss: 3.7028, Perplexity: 40.5622
Epoch [1/3], Step [1833/3236], Loss: 2.2454, Perplexity: 9.4445
Epoch [1/3], Step [1834/3236], Loss: 2.2058, Perplexity: 9.0775
Epoch [1/3], Step [1835/3236], Loss: 2.3701, Perplexity: 10.6983
Epoch [1/3], Step [1836/3236], Loss: 2.3195, Perplexity: 10.1701
Epoch [1/3], Step [1837/3236], Loss: 2.3615, Perplexity: 10.6069
Epoch [1/3], Step [1838/3236], Loss: 2.3025, Perplexity: 9.9988
Epoch [1/3], Step [1839/3236], Loss: 2.2757, Perplexity: 9.7344
Epoch [1/3], Step [1840/3236], Loss: 2.3670, Perplexity: 10.6653
Epoch [1/3], Step [1841/3236], Loss: 2.1970, Perplexity: 8.9984
Epoch [1/3], Step [1842/3236], Loss: 2.3728, Perplexity: 10.7275
Epoch [1/3], Step [1843/3236], Loss: 2.1617, Perplexity: 8.6855
Epoch [1/3], Step [1844/3236], Loss: 2.3102, Perplexity: 10.0767
Epoch [1/3], Step [1845/3236], Loss: 2.2699, Perplexity: 9.6788
Epoch [1/3], Step [1846/3236], Loss: 2.2694, Perplexity: 9.6735
Epoch [1/3], Step [1847/3236], Loss: 2.2273, Perplexity: 9.2746
Epoch [1/3], Step [1848/3236], Loss: 2.2909, Perplexity: 9.8834
Epoch [1/3], Step [1849/3236], Loss: 2.2225, Perplexity: 9.2304
Epoch [1/3], Step [1850/3236], Loss: 2.2361, Perplexity: 9.3565
Epoch [1/3], Step [1851/3236], Loss: 2.2821, Perplexity: 9.7968
Epoch [1/3], Step [1852/3236], Loss: 2.5627, Perplexity: 12.9714
Epoch [1/3], Step [1853/3236], Loss: 2.3390, Perplexity: 10.3709
Epoch [1/3], Step [1854/3236], Loss: 2.2714, Perplexity: 9.6932
Epoch [1/3], Step [1855/3236], Loss: 2.2211, Perplexity: 9.2175
Epoch [1/3], Step [1856/3236], Loss: 2.2108, Perplexity: 9.1228
Epoch [1/3], Step [1857/3236], Loss: 2.3892, Perplexity: 10.9048
Epoch [1/3], Step [1858/3236], Loss: 2.5656, Perplexity: 13.0080
Epoch [1/3], Step [1859/3236], Loss: 3.4614, Perplexity: 31.8617
Epoch [1/3], Step [1860/3236], Loss: 2.2479, Perplexity: 9.4680
Epoch [1/3], Step [1861/3236], Loss: 2.3144, Perplexity: 10.1192
Epoch [1/3], Step [1862/3236], Loss: 2.2666, Perplexity: 9.6465
Epoch [1/3], Step [1863/3236], Loss: 2.3698, Perplexity: 10.6948
Epoch [1/3], Step [1864/3236], Loss: 2.4356, Perplexity: 11.4224
Epoch [1/3], Step [1865/3236], Loss: 2.2893, Perplexity: 9.8684
Epoch [1/3], Step [1866/3236], Loss: 2.3978, Perplexity: 10.9990
Epoch [1/3], Step [1867/3236], Loss: 2.2165, Perplexity: 9.1752
Epoch [1/3], Step [1868/3236], Loss: 2.4138, Perplexity: 11.1758
Epoch [1/3], Step [1869/3236], Loss: 2.6367, Perplexity: 13.9677
Epoch [1/3], Step [1870/3236], Loss: 2.6621, Perplexity: 14.3271
Epoch [1/3], Step [1871/3236], Loss: 2.5122, Perplexity: 12.3315
Epoch [1/3], Step [1872/3236], Loss: 2.6404, Perplexity: 14.0193
Epoch [1/3], Step [1873/3236], Loss: 2.2627, Perplexity: 9.6090
Epoch [1/3], Step [1874/3236], Loss: 2.5045, Perplexity: 12.2373
Epoch [1/3], Step [1875/3236], Loss: 2.2372, Perplexity: 9.3675
Epoch [1/3], Step [1876/3236], Loss: 2.3101, Perplexity: 10.0757
Epoch [1/3], Step [1877/3236], Loss: 2.6374, Perplexity: 13.9765
Epoch [1/3], Step [1878/3236], Loss: 2.4930, Perplexity: 12.0981
Epoch [1/3], Step [1879/3236], Loss: 2.2169, Perplexity: 9.1790
Epoch [1/3], Step [1880/3236], Loss: 2.1879, Perplexity: 8.9162
Epoch [1/3], Step [1881/3236], Loss: 2.8915, Perplexity: 18.0199
Epoch [1/3], Step [1882/3236], Loss: 2.5914, Perplexity: 13.3489
Epoch [1/3], Step [1883/3236], Loss: 2.6974, Perplexity: 14.8404
Epoch [1/3], Step [1884/3236], Loss: 2.3891, Perplexity: 10.9040
Epoch [1/3], Step [1885/3236], Loss: 2.2362, Perplexity: 9.3578
Epoch [1/3], Step [1886/3236], Loss: 2.3446, Perplexity: 10.4290
Epoch [1/3], Step [1887/3236], Loss: 2.2927, Perplexity: 9.9012
Epoch [1/3], Step [1888/3236], Loss: 2.3706, Perplexity: 10.7035
Epoch [1/3], Step [1889/3236], Loss: 2.2022, Perplexity: 9.0447
Epoch [1/3], Step [1890/3236], Loss: 2.2436, Perplexity: 9.4268
Epoch [1/3], Step [1891/3236], Loss: 2.3782, Perplexity: 10.7852
Epoch [1/3], Step [1892/3236], Loss: 2.1567, Perplexity: 8.6427
Epoch [1/3], Step [1893/3236], Loss: 2.2679, Perplexity: 9.6590
Epoch [1/3], Step [1894/3236], Loss: 2.3965, Perplexity: 10.9846
Epoch [1/3], Step [1895/3236], Loss: 2.7706, Perplexity: 15.9685
Epoch [1/3], Step [1896/3236], Loss: 2.2502, Perplexity: 9.4893
Epoch [1/3], Step [1897/3236], Loss: 2.3629, Perplexity: 10.6218
Epoch [1/3], Step [1898/3236], Loss: 2.4452, Perplexity: 11.5334
Epoch [1/3], Step [1899/3236], Loss: 2.3475, Perplexity: 10.4590
Epoch [1/3], Step [1900/3236], Loss: 2.2075, Perplexity: 9.0933
Epoch [1/3], Step [1901/3236], Loss: 2.1296, Perplexity: 8.4115
Epoch [1/3], Step [1902/3236], Loss: 2.2064, Perplexity: 9.0831
Epoch [1/3], Step [1903/3236], Loss: 2.2168, Perplexity: 9.1781
Epoch [1/3], Step [1904/3236], Loss: 2.3426, Perplexity: 10.4081
Epoch [1/3], Step [1905/3236], Loss: 2.3724, Perplexity: 10.7233
Epoch [1/3], Step [1906/3236], Loss: 2.4909, Perplexity: 12.0722
Epoch [1/3], Step [1907/3236], Loss: 2.3419, Perplexity: 10.4007
Epoch [1/3], Step [1908/3236], Loss: 2.3364, Perplexity: 10.3437
Epoch [1/3], Step [1909/3236], Loss: 2.2101, Perplexity: 9.1170
Epoch [1/3], Step [1910/3236], Loss: 2.8438, Perplexity: 17.1817
Epoch [1/3], Step [1911/3236], Loss: 2.2810, Perplexity: 9.7866
Epoch [1/3], Step [1912/3236], Loss: 2.3516, Perplexity: 10.5019
Epoch [1/3], Step [1913/3236], Loss: 2.3285, Perplexity: 10.2628
Epoch [1/3], Step [1914/3236], Loss: 2.4120, Perplexity: 11.1557
Epoch [1/3], Step [1915/3236], Loss: 2.5090, Perplexity: 12.2922
Epoch [1/3], Step [1916/3236], Loss: 2.2554, Perplexity: 9.5392
Epoch [1/3], Step [1917/3236], Loss: 2.9507, Perplexity: 19.1189
Epoch [1/3], Step [1918/3236], Loss: 2.1981, Perplexity: 9.0078
Epoch [1/3], Step [1919/3236], Loss: 2.3977, Perplexity: 10.9976
Epoch [1/3], Step [1920/3236], Loss: 2.2281, Perplexity: 9.2818
Epoch [1/3], Step [1921/3236], Loss: 2.3813, Perplexity: 10.8194
Epoch [1/3], Step [1922/3236], Loss: 2.3871, Perplexity: 10.8824
Epoch [1/3], Step [1923/3236], Loss: 2.1691, Perplexity: 8.7504
Epoch [1/3], Step [1924/3236], Loss: 2.3290, Perplexity: 10.2676
Epoch [1/3], Step [1925/3236], Loss: 2.4215, Perplexity: 11.2624
Epoch [1/3], Step [1926/3236], Loss: 2.2769, Perplexity: 9.7460
Epoch [1/3], Step [1927/3236], Loss: 2.2647, Perplexity: 9.6282
Epoch [1/3], Step [1928/3236], Loss: 2.3855, Perplexity: 10.8640
Epoch [1/3], Step [1929/3236], Loss: 2.2563, Perplexity: 9.5480
Epoch [1/3], Step [1930/3236], Loss: 2.3615, Perplexity: 10.6068
Epoch [1/3], Step [1931/3236], Loss: 2.3094, Perplexity: 10.0685
Epoch [1/3], Step [1932/3236], Loss: 2.2374, Perplexity: 9.3687
Epoch [1/3], Step [1933/3236], Loss: 2.3545, Perplexity: 10.5326
Epoch [1/3], Step [1934/3236], Loss: 2.3460, Perplexity: 10.4434
Epoch [1/3], Step [1935/3236], Loss: 2.3772, Perplexity: 10.7751
Epoch [1/3], Step [1936/3236], Loss: 3.2350, Perplexity: 25.4070
Epoch [1/3], Step [1937/3236], Loss: 2.4763, Perplexity: 11.8968
Epoch [1/3], Step [1938/3236], Loss: 2.2661, Perplexity: 9.6414
Epoch [1/3], Step [1939/3236], Loss: 2.3787, Perplexity: 10.7907
Epoch [1/3], Step [1940/3236], Loss: 2.3084, Perplexity: 10.0578
Epoch [1/3], Step [1941/3236], Loss: 2.2701, Perplexity: 9.6806
Epoch [1/3], Step [1942/3236], Loss: 2.4459, Perplexity: 11.5407
Epoch [1/3], Step [1943/3236], Loss: 2.3155, Perplexity: 10.1303
Epoch [1/3], Step [1944/3236], Loss: 2.2717, Perplexity: 9.6955
Epoch [1/3], Step [1945/3236], Loss: 2.3054, Perplexity: 10.0281
Epoch [1/3], Step [1946/3236], Loss: 2.4229, Perplexity: 11.2783
Epoch [1/3], Step [1947/3236], Loss: 2.2599, Perplexity: 9.5823
Epoch [1/3], Step [1948/3236], Loss: 2.1967, Perplexity: 8.9957
Epoch [1/3], Step [1949/3236], Loss: 3.0343, Perplexity: 20.7863
Epoch [1/3], Step [1950/3236], Loss: 2.3883, Perplexity: 10.8954
Epoch [1/3], Step [1951/3236], Loss: 2.4296, Perplexity: 11.3545
Epoch [1/3], Step [1952/3236], Loss: 2.1664, Perplexity: 8.7269
Epoch [1/3], Step [1953/3236], Loss: 2.3510, Perplexity: 10.4963
Epoch [1/3], Step [1954/3236], Loss: 2.4108, Perplexity: 11.1428
Epoch [1/3], Step [1955/3236], Loss: 2.4175, Perplexity: 11.2174
Epoch [1/3], Step [1956/3236], Loss: 2.7281, Perplexity: 15.3040
Epoch [1/3], Step [1957/3236], Loss: 2.3175, Perplexity: 10.1499
Epoch [1/3], Step [1958/3236], Loss: 2.6811, Perplexity: 14.6004
Epoch [1/3], Step [1959/3236], Loss: 2.4171, Perplexity: 11.2135
Epoch [1/3], Step [1960/3236], Loss: 2.2799, Perplexity: 9.7760
Epoch [1/3], Step [1961/3236], Loss: 2.3880, Perplexity: 10.8919
Epoch [1/3], Step [1962/3236], Loss: 2.3351, Perplexity: 10.3301
Epoch [1/3], Step [1963/3236], Loss: 2.2470, Perplexity: 9.4594
Epoch [1/3], Step [1964/3236], Loss: 2.3264, Perplexity: 10.2413
Epoch [1/3], Step [1965/3236], Loss: 2.8019, Perplexity: 16.4753
Epoch [1/3], Step [1966/3236], Loss: 2.7667, Perplexity: 15.9061
Epoch [1/3], Step [1967/3236], Loss: 2.4540, Perplexity: 11.6348
Epoch [1/3], Step [1968/3236], Loss: 2.3148, Perplexity: 10.1233
Epoch [1/3], Step [1969/3236], Loss: 2.1389, Perplexity: 8.4902
Epoch [1/3], Step [1970/3236], Loss: 2.1896, Perplexity: 8.9321
Epoch [1/3], Step [1971/3236], Loss: 2.2152, Perplexity: 9.1628
Epoch [1/3], Step [1972/3236], Loss: 2.2927, Perplexity: 9.9015
Epoch [1/3], Step [1973/3236], Loss: 2.2206, Perplexity: 9.2125
Epoch [1/3], Step [1974/3236], Loss: 2.2665, Perplexity: 9.6460
Epoch [1/3], Step [1975/3236], Loss: 2.4091, Perplexity: 11.1242
Epoch [1/3], Step [1976/3236], Loss: 2.3996, Perplexity: 11.0191
Epoch [1/3], Step [1977/3236], Loss: 2.4217, Perplexity: 11.2654
Epoch [1/3], Step [1978/3236], Loss: 3.1469, Perplexity: 23.2633
Epoch [1/3], Step [1979/3236], Loss: 2.2092, Perplexity: 9.1081
Epoch [1/3], Step [1980/3236], Loss: 2.7128, Perplexity: 15.0717
Epoch [1/3], Step [1981/3236], Loss: 2.1343, Perplexity: 8.4514
Epoch [1/3], Step [1982/3236], Loss: 2.6170, Perplexity: 13.6945
Epoch [1/3], Step [1983/3236], Loss: 2.3941, Perplexity: 10.9583
Epoch [1/3], Step [1984/3236], Loss: 2.3026, Perplexity: 10.0005
Epoch [1/3], Step [1985/3236], Loss: 2.2749, Perplexity: 9.7267
Epoch [1/3], Step [1986/3236], Loss: 2.4003, Perplexity: 11.0270
Epoch [1/3], Step [1987/3236], Loss: 2.1207, Perplexity: 8.3367
Epoch [1/3], Step [1988/3236], Loss: 2.8253, Perplexity: 16.8666
Epoch [1/3], Step [1989/3236], Loss: 2.0550, Perplexity: 7.8067
Epoch [1/3], Step [1990/3236], Loss: 2.5239, Perplexity: 12.4770
Epoch [1/3], Step [1991/3236], Loss: 2.5043, Perplexity: 12.2353
Epoch [1/3], Step [1992/3236], Loss: 2.4771, Perplexity: 11.9067
Epoch [1/3], Step [1993/3236], Loss: 2.2077, Perplexity: 9.0945
Epoch [1/3], Step [1994/3236], Loss: 2.8171, Perplexity: 16.7283
Epoch [1/3], Step [1995/3236], Loss: 2.2225, Perplexity: 9.2305
Epoch [1/3], Step [1996/3236], Loss: 2.4548, Perplexity: 11.6442
Epoch [1/3], Step [1997/3236], Loss: 2.2498, Perplexity: 9.4855
Epoch [1/3], Step [1998/3236], Loss: 2.8126, Perplexity: 16.6535
Epoch [1/3], Step [1999/3236], Loss: 2.2166, Perplexity: 9.1756
Epoch [1/3], Step [2000/3236], Loss: 2.3227, Perplexity: 10.2030
Epoch [1/3], Step [2001/3236], Loss: 2.1712, Perplexity: 8.7684
Epoch [1/3], Step [2002/3236], Loss: 2.5483, Perplexity: 12.7851
Epoch [1/3], Step [2003/3236], Loss: 2.3708, Perplexity: 10.7060
Epoch [1/3], Step [2004/3236], Loss: 2.4370, Perplexity: 11.4392
Epoch [1/3], Step [2005/3236], Loss: 2.3099, Perplexity: 10.0737
Epoch [1/3], Step [2006/3236], Loss: 2.3053, Perplexity: 10.0270
Epoch [1/3], Step [2007/3236], Loss: 2.2506, Perplexity: 9.4935
Epoch [1/3], Step [2008/3236], Loss: 2.6207, Perplexity: 13.7451
Epoch [1/3], Step [2009/3236], Loss: 2.2421, Perplexity: 9.4135
Epoch [1/3], Step [2010/3236], Loss: 2.3597, Perplexity: 10.5880
Epoch [1/3], Step [2011/3236], Loss: 2.6587, Perplexity: 14.2777
Epoch [1/3], Step [2012/3236], Loss: 2.4121, Perplexity: 11.1572
Epoch [1/3], Step [2013/3236], Loss: 2.2570, Perplexity: 9.5542
Epoch [1/3], Step [2014/3236], Loss: 2.3930, Perplexity: 10.9467
Epoch [1/3], Step [2015/3236], Loss: 2.2211, Perplexity: 9.2174
Epoch [1/3], Step [2016/3236], Loss: 2.2083, Perplexity: 9.0998
Epoch [1/3], Step [2017/3236], Loss: 2.1469, Perplexity: 8.5579
Epoch [1/3], Step [2018/3236], Loss: 2.2659, Perplexity: 9.6393
Epoch [1/3], Step [2019/3236], Loss: 2.2981, Perplexity: 9.9555
Epoch [1/3], Step [2020/3236], Loss: 2.1747, Perplexity: 8.7998
Epoch [1/3], Step [2021/3236], Loss: 2.2298, Perplexity: 9.2984
Epoch [1/3], Step [2022/3236], Loss: 3.4525, Perplexity: 31.5794
Epoch [1/3], Step [2023/3236], Loss: 2.2001, Perplexity: 9.0257
Epoch [1/3], Step [2024/3236], Loss: 2.3495, Perplexity: 10.4802
Epoch [1/3], Step [2025/3236], Loss: 2.4942, Perplexity: 12.1124
Epoch [1/3], Step [2026/3236], Loss: 2.3406, Perplexity: 10.3879
Epoch [1/3], Step [2027/3236], Loss: 2.3015, Perplexity: 9.9893
Epoch [1/3], Step [2028/3236], Loss: 2.1871, Perplexity: 8.9091
Epoch [1/3], Step [2029/3236], Loss: 2.2716, Perplexity: 9.6952
Epoch [1/3], Step [2030/3236], Loss: 2.2671, Perplexity: 9.6511
Epoch [1/3], Step [2031/3236], Loss: 2.1331, Perplexity: 8.4412
Epoch [1/3], Step [2032/3236], Loss: 2.2540, Perplexity: 9.5259
Epoch [1/3], Step [2033/3236], Loss: 2.3645, Perplexity: 10.6388
Epoch [1/3], Step [2034/3236], Loss: 2.2116, Perplexity: 9.1306
Epoch [1/3], Step [2035/3236], Loss: 2.3400, Perplexity: 10.3813
Epoch [1/3], Step [2036/3236], Loss: 2.2363, Perplexity: 9.3583
Epoch [1/3], Step [2037/3236], Loss: 2.2079, Perplexity: 9.0966
Epoch [1/3], Step [2038/3236], Loss: 2.4180, Perplexity: 11.2239
Epoch [1/3], Step [2039/3236], Loss: 2.1696, Perplexity: 8.7544
Epoch [1/3], Step [2040/3236], Loss: 2.3425, Perplexity: 10.4072
Epoch [1/3], Step [2041/3236], Loss: 2.3844, Perplexity: 10.8521
Epoch [1/3], Step [2042/3236], Loss: 2.3628, Perplexity: 10.6202
Epoch [1/3], Step [2043/3236], Loss: 2.2355, Perplexity: 9.3508
Epoch [1/3], Step [2044/3236], Loss: 2.6630, Perplexity: 14.3392
Epoch [1/3], Step [2045/3236], Loss: 2.2595, Perplexity: 9.5783
Epoch [1/3], Step [2046/3236], Loss: 2.2306, Perplexity: 9.3052
Epoch [1/3], Step [2047/3236], Loss: 2.5236, Perplexity: 12.4738
Epoch [1/3], Step [2048/3236], Loss: 2.3047, Perplexity: 10.0212
Epoch [1/3], Step [2049/3236], Loss: 2.3251, Perplexity: 10.2275
Epoch [1/3], Step [2050/3236], Loss: 2.5005, Perplexity: 12.1883
Epoch [1/3], Step [2051/3236], Loss: 2.2616, Perplexity: 9.5980
Epoch [1/3], Step [2052/3236], Loss: 2.6718, Perplexity: 14.4664
Epoch [1/3], Step [2053/3236], Loss: 2.2053, Perplexity: 9.0731
Epoch [1/3], Step [2054/3236], Loss: 2.2205, Perplexity: 9.2121
Epoch [1/3], Step [2055/3236], Loss: 3.1932, Perplexity: 24.3652
Epoch [1/3], Step [2056/3236], Loss: 3.0081, Perplexity: 20.2484
Epoch [1/3], Step [2057/3236], Loss: 2.2806, Perplexity: 9.7829
Epoch [1/3], Step [2058/3236], Loss: 2.5223, Perplexity: 12.4573
Epoch [1/3], Step [2059/3236], Loss: 2.2793, Perplexity: 9.7701
Epoch [1/3], Step [2060/3236], Loss: 2.1385, Perplexity: 8.4870
Epoch [1/3], Step [2061/3236], Loss: 2.3734, Perplexity: 10.7339
Epoch [1/3], Step [2062/3236], Loss: 2.3302, Perplexity: 10.2797
Epoch [1/3], Step [2063/3236], Loss: 2.2363, Perplexity: 9.3588
Epoch [1/3], Step [2064/3236], Loss: 2.1868, Perplexity: 8.9070
Epoch [1/3], Step [2065/3236], Loss: 2.4104, Perplexity: 11.1387
Epoch [1/3], Step [2066/3236], Loss: 2.5627, Perplexity: 12.9711
Epoch [1/3], Step [2067/3236], Loss: 2.2055, Perplexity: 9.0750
Epoch [1/3], Step [2068/3236], Loss: 2.4008, Perplexity: 11.0320
Epoch [1/3], Step [2069/3236], Loss: 2.2896, Perplexity: 9.8706
Epoch [1/3], Step [2070/3236], Loss: 2.1809, Perplexity: 8.8544
Epoch [1/3], Step [2071/3236], Loss: 2.2867, Perplexity: 9.8420
Epoch [1/3], Step [2072/3236], Loss: 2.2276, Perplexity: 9.2775
Epoch [1/3], Step [2073/3236], Loss: 2.2173, Perplexity: 9.1827
Epoch [1/3], Step [2074/3236], Loss: 2.2176, Perplexity: 9.1851
Epoch [1/3], Step [2075/3236], Loss: 2.8197, Perplexity: 16.7712
Epoch [1/3], Step [2076/3236], Loss: 2.4227, Perplexity: 11.2757
Epoch [1/3], Step [2077/3236], Loss: 2.2306, Perplexity: 9.3053
Epoch [1/3], Step [2078/3236], Loss: 2.5772, Perplexity: 13.1608
Epoch [1/3], Step [2079/3236], Loss: 2.3002, Perplexity: 9.9759
Epoch [1/3], Step [2080/3236], Loss: 2.3646, Perplexity: 10.6398
Epoch [1/3], Step [2081/3236], Loss: 2.2038, Perplexity: 9.0592
Epoch [1/3], Step [2082/3236], Loss: 2.3535, Perplexity: 10.5226
Epoch [1/3], Step [2083/3236], Loss: 2.2700, Perplexity: 9.6795
Epoch [1/3], Step [2084/3236], Loss: 2.1730, Perplexity: 8.7844
Epoch [1/3], Step [2085/3236], Loss: 2.3670, Perplexity: 10.6658
Epoch [1/3], Step [2086/3236], Loss: 2.2290, Perplexity: 9.2902
Epoch [1/3], Step [2087/3236], Loss: 2.4456, Perplexity: 11.5376
Epoch [1/3], Step [2088/3236], Loss: 2.2503, Perplexity: 9.4904
Epoch [1/3], Step [2089/3236], Loss: 2.2713, Perplexity: 9.6921
Epoch [1/3], Step [2090/3236], Loss: 2.2909, Perplexity: 9.8835
Epoch [1/3], Step [2091/3236], Loss: 2.3038, Perplexity: 10.0126
Epoch [1/3], Step [2092/3236], Loss: 2.4046, Perplexity: 11.0735
Epoch [1/3], Step [2093/3236], Loss: 2.3026, Perplexity: 10.0002
Epoch [1/3], Step [2094/3236], Loss: 2.2535, Perplexity: 9.5212
Epoch [1/3], Step [2095/3236], Loss: 2.2870, Perplexity: 9.8453
Epoch [1/3], Step [2096/3236], Loss: 2.3184, Perplexity: 10.1596
Epoch [1/3], Step [2097/3236], Loss: 2.2644, Perplexity: 9.6254
Epoch [1/3], Step [2098/3236], Loss: 2.3356, Perplexity: 10.3361
Epoch [1/3], Step [2099/3236], Loss: 2.3634, Perplexity: 10.6267
Epoch [1/3], Step [2100/3236], Loss: 2.3397, Perplexity: 10.3777
Epoch [1/3], Step [2101/3236], Loss: 2.2687, Perplexity: 9.6667
Epoch [1/3], Step [2102/3236], Loss: 2.3111, Perplexity: 10.0860
Epoch [1/3], Step [2103/3236], Loss: 2.7085, Perplexity: 15.0067
Epoch [1/3], Step [2104/3236], Loss: 2.2651, Perplexity: 9.6323
Epoch [1/3], Step [2105/3236], Loss: 2.3452, Perplexity: 10.4352
Epoch [1/3], Step [2106/3236], Loss: 2.1553, Perplexity: 8.6309
Epoch [1/3], Step [2107/3236], Loss: 2.3440, Perplexity: 10.4226
Epoch [1/3], Step [2108/3236], Loss: 2.4522, Perplexity: 11.6133
Epoch [1/3], Step [2109/3236], Loss: 2.2185, Perplexity: 9.1936
Epoch [1/3], Step [2110/3236], Loss: 2.1371, Perplexity: 8.4751
Epoch [1/3], Step [2111/3236], Loss: 2.2166, Perplexity: 9.1758
Epoch [1/3], Step [2112/3236], Loss: 2.1470, Perplexity: 8.5594
Epoch [1/3], Step [2113/3236], Loss: 2.7615, Perplexity: 15.8236
Epoch [1/3], Step [2114/3236], Loss: 2.2893, Perplexity: 9.8685
Epoch [1/3], Step [2115/3236], Loss: 2.2330, Perplexity: 9.3278
Epoch [1/3], Step [2116/3236], Loss: 2.3125, Perplexity: 10.0993
Epoch [1/3], Step [2117/3236], Loss: 2.1984, Perplexity: 9.0106
Epoch [1/3], Step [2118/3236], Loss: 2.1994, Perplexity: 9.0200
Epoch [1/3], Step [2119/3236], Loss: 2.2473, Perplexity: 9.4622
Epoch [1/3], Step [2120/3236], Loss: 2.3303, Perplexity: 10.2811
Epoch [1/3], Step [2121/3236], Loss: 2.3538, Perplexity: 10.5256
Epoch [1/3], Step [2122/3236], Loss: 2.2822, Perplexity: 9.7985
Epoch [1/3], Step [2123/3236], Loss: 2.3008, Perplexity: 9.9821
Epoch [1/3], Step [2124/3236], Loss: 2.1781, Perplexity: 8.8292
Epoch [1/3], Step [2125/3236], Loss: 2.9780, Perplexity: 19.6486
Epoch [1/3], Step [2126/3236], Loss: 2.2632, Perplexity: 9.6139
Epoch [1/3], Step [2127/3236], Loss: 2.2264, Perplexity: 9.2663
Epoch [1/3], Step [2128/3236], Loss: 2.3792, Perplexity: 10.7960
Epoch [1/3], Step [2129/3236], Loss: 2.3072, Perplexity: 10.0465
Epoch [1/3], Step [2130/3236], Loss: 2.2842, Perplexity: 9.8179
Epoch [1/3], Step [2131/3236], Loss: 2.3224, Perplexity: 10.1998
Epoch [1/3], Step [2132/3236], Loss: 2.2959, Perplexity: 9.9335
Epoch [1/3], Step [2133/3236], Loss: 2.3148, Perplexity: 10.1226
Epoch [1/3], Step [2134/3236], Loss: 2.3006, Perplexity: 9.9798
Epoch [1/3], Step [2135/3236], Loss: 2.1451, Perplexity: 8.5430
Epoch [1/3], Step [2136/3236], Loss: 2.1754, Perplexity: 8.8061
Epoch [1/3], Step [2137/3236], Loss: 2.4972, Perplexity: 12.1479
Epoch [1/3], Step [2138/3236], Loss: 2.1990, Perplexity: 9.0156
Epoch [1/3], Step [2139/3236], Loss: 2.1902, Perplexity: 8.9371
Epoch [1/3], Step [2140/3236], Loss: 2.1213, Perplexity: 8.3416
Epoch [1/3], Step [2141/3236], Loss: 2.9102, Perplexity: 18.3601
Epoch [1/3], Step [2142/3236], Loss: 2.4442, Perplexity: 11.5215
Epoch [1/3], Step [2143/3236], Loss: 2.1366, Perplexity: 8.4708
Epoch [1/3], Step [2144/3236], Loss: 2.2955, Perplexity: 9.9294
Epoch [1/3], Step [2145/3236], Loss: 2.2812, Perplexity: 9.7885
Epoch [1/3], Step [2146/3236], Loss: 2.3060, Perplexity: 10.0341
Epoch [1/3], Step [2147/3236], Loss: 2.1591, Perplexity: 8.6632
Epoch [1/3], Step [2148/3236], Loss: 2.2074, Perplexity: 9.0920
Epoch [1/3], Step [2149/3236], Loss: 2.3887, Perplexity: 10.8991
Epoch [1/3], Step [2150/3236], Loss: 2.4303, Perplexity: 11.3623
Epoch [1/3], Step [2151/3236], Loss: 2.3598, Perplexity: 10.5891
Epoch [1/3], Step [2152/3236], Loss: 2.3573, Perplexity: 10.5628
Epoch [1/3], Step [2153/3236], Loss: 2.1502, Perplexity: 8.5866
Epoch [1/3], Step [2154/3236], Loss: 2.6299, Perplexity: 13.8720
Epoch [1/3], Step [2155/3236], Loss: 2.5783, Perplexity: 13.1750
Epoch [1/3], Step [2156/3236], Loss: 2.3308, Perplexity: 10.2864
Epoch [1/3], Step [2157/3236], Loss: 2.1310, Perplexity: 8.4230
Epoch [1/3], Step [2158/3236], Loss: 2.3374, Perplexity: 10.3538
Epoch [1/3], Step [2159/3236], Loss: 2.2354, Perplexity: 9.3505
Epoch [1/3], Step [2160/3236], Loss: 2.5346, Perplexity: 12.6112
Epoch [1/3], Step [2161/3236], Loss: 2.1675, Perplexity: 8.7364
Epoch [1/3], Step [2162/3236], Loss: 3.6291, Perplexity: 37.6779
Epoch [1/3], Step [2163/3236], Loss: 2.4529, Perplexity: 11.6215
Epoch [1/3], Step [2164/3236], Loss: 2.3654, Perplexity: 10.6485
Epoch [1/3], Step [2165/3236], Loss: 2.3471, Perplexity: 10.4553
Epoch [1/3], Step [2166/3236], Loss: 2.2826, Perplexity: 9.8017
Epoch [1/3], Step [2167/3236], Loss: 2.2729, Perplexity: 9.7079
Epoch [1/3], Step [2168/3236], Loss: 2.1225, Perplexity: 8.3517
Epoch [1/3], Step [2169/3236], Loss: 2.1820, Perplexity: 8.8644
Epoch [1/3], Step [2170/3236], Loss: 2.2435, Perplexity: 9.4265
Epoch [1/3], Step [2171/3236], Loss: 2.1315, Perplexity: 8.4275
Epoch [1/3], Step [2172/3236], Loss: 2.2873, Perplexity: 9.8486
Epoch [1/3], Step [2173/3236], Loss: 2.2072, Perplexity: 9.0905
Epoch [1/3], Step [2174/3236], Loss: 2.2255, Perplexity: 9.2583
Epoch [1/3], Step [2175/3236], Loss: 2.2550, Perplexity: 9.5354
Epoch [1/3], Step [2176/3236], Loss: 2.4027, Perplexity: 11.0529
Epoch [1/3], Step [2177/3236], Loss: 2.3652, Perplexity: 10.6465
Epoch [1/3], Step [2178/3236], Loss: 2.1516, Perplexity: 8.5985
Epoch [1/3], Step [2179/3236], Loss: 2.3649, Perplexity: 10.6427
Epoch [1/3], Step [2180/3236], Loss: 2.3196, Perplexity: 10.1713
Epoch [1/3], Step [2181/3236], Loss: 2.0773, Perplexity: 7.9831
Epoch [1/3], Step [2182/3236], Loss: 2.3385, Perplexity: 10.3656
Epoch [1/3], Step [2183/3236], Loss: 2.1031, Perplexity: 8.1918
Epoch [1/3], Step [2184/3236], Loss: 2.3832, Perplexity: 10.8399
Epoch [1/3], Step [2185/3236], Loss: 2.1455, Perplexity: 8.5464
Epoch [1/3], Step [2186/3236], Loss: 2.4271, Perplexity: 11.3258
Epoch [1/3], Step [2187/3236], Loss: 2.1557, Perplexity: 8.6341
Epoch [1/3], Step [2188/3236], Loss: 2.5714, Perplexity: 13.0835
Epoch [1/3], Step [2189/3236], Loss: 2.3155, Perplexity: 10.1304
Epoch [1/3], Step [2190/3236], Loss: 2.3039, Perplexity: 10.0136
Epoch [1/3], Step [2191/3236], Loss: 2.2110, Perplexity: 9.1246
Epoch [1/3], Step [2192/3236], Loss: 2.6046, Perplexity: 13.5254
Epoch [1/3], Step [2193/3236], Loss: 2.2990, Perplexity: 9.9647
Epoch [1/3], Step [2194/3236], Loss: 2.3007, Perplexity: 9.9810
Epoch [1/3], Step [2195/3236], Loss: 2.3156, Perplexity: 10.1311
Epoch [1/3], Step [2196/3236], Loss: 2.3421, Perplexity: 10.4030
Epoch [1/3], Step [2197/3236], Loss: 2.2934, Perplexity: 9.9081
Epoch [1/3], Step [2198/3236], Loss: 2.2742, Perplexity: 9.7200
Epoch [1/3], Step [2199/3236], Loss: 2.1947, Perplexity: 8.9773
Epoch [1/3], Step [2200/3236], Loss: 2.2940, Perplexity: 9.9147
Epoch [1/3], Step [2201/3236], Loss: 2.2364, Perplexity: 9.3598
Epoch [1/3], Step [2202/3236], Loss: 2.6201, Perplexity: 13.7375
Epoch [1/3], Step [2203/3236], Loss: 2.1659, Perplexity: 8.7221
Epoch [1/3], Step [2204/3236], Loss: 2.0366, Perplexity: 7.6641
Epoch [1/3], Step [2205/3236], Loss: 3.3361, Perplexity: 28.1086
Epoch [1/3], Step [2206/3236], Loss: 2.1453, Perplexity: 8.5445
Epoch [1/3], Step [2207/3236], Loss: 2.1385, Perplexity: 8.4865
Epoch [1/3], Step [2208/3236], Loss: 2.2890, Perplexity: 9.8652
Epoch [1/3], Step [2209/3236], Loss: 2.8908, Perplexity: 18.0083
Epoch [1/3], Step [2210/3236], Loss: 2.4784, Perplexity: 11.9226
Epoch [1/3], Step [2211/3236], Loss: 2.1862, Perplexity: 8.9015
Epoch [1/3], Step [2212/3236], Loss: 2.9515, Perplexity: 19.1349
Epoch [1/3], Step [2213/3236], Loss: 2.2263, Perplexity: 9.2657
Epoch [1/3], Step [2214/3236], Loss: 2.5661, Perplexity: 13.0148
Epoch [1/3], Step [2215/3236], Loss: 2.3488, Perplexity: 10.4729
Epoch [1/3], Step [2216/3236], Loss: 2.3317, Perplexity: 10.2953
Epoch [1/3], Step [2217/3236], Loss: 2.1744, Perplexity: 8.7969
Epoch [1/3], Step [2218/3236], Loss: 2.2089, Perplexity: 9.1060
Epoch [1/3], Step [2219/3236], Loss: 2.3108, Perplexity: 10.0824
Epoch [1/3], Step [2220/3236], Loss: 2.1691, Perplexity: 8.7506
Epoch [1/3], Step [2221/3236], Loss: 2.3146, Perplexity: 10.1208
Epoch [1/3], Step [2222/3236], Loss: 2.2556, Perplexity: 9.5415
Epoch [1/3], Step [2223/3236], Loss: 2.2084, Perplexity: 9.1015
Epoch [1/3], Step [2224/3236], Loss: 2.1336, Perplexity: 8.4451
Epoch [1/3], Step [2225/3236], Loss: 2.1277, Perplexity: 8.3956
Epoch [1/3], Step [2226/3236], Loss: 2.1541, Perplexity: 8.6198
Epoch [1/3], Step [2227/3236], Loss: 2.1906, Perplexity: 8.9406
Epoch [1/3], Step [2228/3236], Loss: 2.5305, Perplexity: 12.5593
Epoch [1/3], Step [2229/3236], Loss: 2.2983, Perplexity: 9.9569
Epoch [1/3], Step [2230/3236], Loss: 2.2894, Perplexity: 9.8686
Epoch [1/3], Step [2231/3236], Loss: 2.2996, Perplexity: 9.9704
Epoch [1/3], Step [2232/3236], Loss: 2.2688, Perplexity: 9.6681
Epoch [1/3], Step [2233/3236], Loss: 2.2381, Perplexity: 9.3754
Epoch [1/3], Step [2234/3236], Loss: 2.0968, Perplexity: 8.1398
Epoch [1/3], Step [2235/3236], Loss: 2.1916, Perplexity: 8.9491
Epoch [1/3], Step [2236/3236], Loss: 2.2686, Perplexity: 9.6658
Epoch [1/3], Step [2237/3236], Loss: 2.1312, Perplexity: 8.4246
Epoch [1/3], Step [2238/3236], Loss: 2.3767, Perplexity: 10.7697
Epoch [1/3], Step [2239/3236], Loss: 2.2517, Perplexity: 9.5041
Epoch [1/3], Step [2240/3236], Loss: 2.1922, Perplexity: 8.9545
Epoch [1/3], Step [2241/3236], Loss: 2.2321, Perplexity: 9.3197
Epoch [1/3], Step [2242/3236], Loss: 2.1438, Perplexity: 8.5318
Epoch [1/3], Step [2243/3236], Loss: 2.2894, Perplexity: 9.8694
Epoch [1/3], Step [2244/3236], Loss: 2.2623, Perplexity: 9.6053
Epoch [1/3], Step [2245/3236], Loss: 2.2001, Perplexity: 9.0259
Epoch [1/3], Step [2246/3236], Loss: 2.4321, Perplexity: 11.3826
Epoch [1/3], Step [2247/3236], Loss: 2.1798, Perplexity: 8.8449
Epoch [1/3], Step [2248/3236], Loss: 2.1362, Perplexity: 8.4673
Epoch [1/3], Step [2249/3236], Loss: 2.4646, Perplexity: 11.7588
Epoch [1/3], Step [2250/3236], Loss: 2.1873, Perplexity: 8.9111
Epoch [1/3], Step [2251/3236], Loss: 2.0661, Perplexity: 7.8939
Epoch [1/3], Step [2252/3236], Loss: 2.1924, Perplexity: 8.9566
Epoch [1/3], Step [2253/3236], Loss: 2.3382, Perplexity: 10.3629
Epoch [1/3], Step [2254/3236], Loss: 2.2574, Perplexity: 9.5581
Epoch [1/3], Step [2255/3236], Loss: 2.1638, Perplexity: 8.7041
Epoch [1/3], Step [2256/3236], Loss: 2.2705, Perplexity: 9.6843
Epoch [1/3], Step [2257/3236], Loss: 2.3086, Perplexity: 10.0608
Epoch [1/3], Step [2258/3236], Loss: 2.2450, Perplexity: 9.4408
Epoch [1/3], Step [2259/3236], Loss: 2.2499, Perplexity: 9.4869
Epoch [1/3], Step [2260/3236], Loss: 2.2251, Perplexity: 9.2540
Epoch [1/3], Step [2261/3236], Loss: 2.2628, Perplexity: 9.6096
Epoch [1/3], Step [2262/3236], Loss: 2.1752, Perplexity: 8.8037
Epoch [1/3], Step [2263/3236], Loss: 4.1290, Perplexity: 62.1134
Epoch [1/3], Step [2264/3236], Loss: 2.5551, Perplexity: 12.8728
Epoch [1/3], Step [2265/3236], Loss: 2.3833, Perplexity: 10.8401
Epoch [1/3], Step [2266/3236], Loss: 2.4342, Perplexity: 11.4071
Epoch [1/3], Step [2267/3236], Loss: 2.1960, Perplexity: 8.9890
Epoch [1/3], Step [2268/3236], Loss: 2.2843, Perplexity: 9.8184
Epoch [1/3], Step [2269/3236], Loss: 2.5777, Perplexity: 13.1667
Epoch [1/3], Step [2270/3236], Loss: 2.5363, Perplexity: 12.6325
Epoch [1/3], Step [2271/3236], Loss: 2.3904, Perplexity: 10.9182
Epoch [1/3], Step [2272/3236], Loss: 2.3826, Perplexity: 10.8327
Epoch [1/3], Step [2273/3236], Loss: 2.2689, Perplexity: 9.6691
Epoch [1/3], Step [2274/3236], Loss: 2.2595, Perplexity: 9.5778
Epoch [1/3], Step [2275/3236], Loss: 2.2438, Perplexity: 9.4292
Epoch [1/3], Step [2276/3236], Loss: 2.4505, Perplexity: 11.5940
Epoch [1/3], Step [2277/3236], Loss: 2.2378, Perplexity: 9.3731
Epoch [1/3], Step [2278/3236], Loss: 2.4829, Perplexity: 11.9759
Epoch [1/3], Step [2279/3236], Loss: 2.1023, Perplexity: 8.1846
Epoch [1/3], Step [2280/3236], Loss: 2.9911, Perplexity: 19.9084
Epoch [1/3], Step [2281/3236], Loss: 2.3176, Perplexity: 10.1517
Epoch [1/3], Step [2282/3236], Loss: 2.9720, Perplexity: 19.5311
Epoch [1/3], Step [2283/3236], Loss: 2.5869, Perplexity: 13.2889
Epoch [1/3], Step [2284/3236], Loss: 2.3077, Perplexity: 10.0517
Epoch [1/3], Step [2285/3236], Loss: 2.1080, Perplexity: 8.2317
Epoch [1/3], Step [2286/3236], Loss: 2.5690, Perplexity: 13.0522
Epoch [1/3], Step [2287/3236], Loss: 2.2280, Perplexity: 9.2809
Epoch [1/3], Step [2288/3236], Loss: 2.1825, Perplexity: 8.8680
Epoch [1/3], Step [2289/3236], Loss: 2.4025, Perplexity: 11.0511
Epoch [1/3], Step [2290/3236], Loss: 3.1386, Perplexity: 23.0712
Epoch [1/3], Step [2291/3236], Loss: 2.3255, Perplexity: 10.2314
Epoch [1/3], Step [2292/3236], Loss: 2.5033, Perplexity: 12.2224
Epoch [1/3], Step [2293/3236], Loss: 2.5162, Perplexity: 12.3820
Epoch [1/3], Step [2294/3236], Loss: 2.2608, Perplexity: 9.5910
Epoch [1/3], Step [2295/3236], Loss: 2.6230, Perplexity: 13.7769
Epoch [1/3], Step [2296/3236], Loss: 2.2604, Perplexity: 9.5869
Epoch [1/3], Step [2297/3236], Loss: 2.3336, Perplexity: 10.3155
Epoch [1/3], Step [2298/3236], Loss: 2.1850, Perplexity: 8.8907
Epoch [1/3], Step [2299/3236], Loss: 2.4993, Perplexity: 12.1736
Epoch [1/3], Step [2300/3236], Loss: 2.1715, Perplexity: 8.7715
Epoch [1/3], Step [2301/3236], Loss: 2.2715, Perplexity: 9.6940
Epoch [1/3], Step [2302/3236], Loss: 2.3985, Perplexity: 11.0062
Epoch [1/3], Step [2303/3236], Loss: 2.3947, Perplexity: 10.9650
Epoch [1/3], Step [2304/3236], Loss: 2.3351, Perplexity: 10.3307
Epoch [1/3], Step [2305/3236], Loss: 2.2357, Perplexity: 9.3529
Epoch [1/3], Step [2306/3236], Loss: 2.3548, Perplexity: 10.5360
Epoch [1/3], Step [2307/3236], Loss: 2.2537, Perplexity: 9.5226
Epoch [1/3], Step [2308/3236], Loss: 2.3105, Perplexity: 10.0797
Epoch [1/3], Step [2309/3236], Loss: 2.3658, Perplexity: 10.6528
Epoch [1/3], Step [2310/3236], Loss: 2.3176, Perplexity: 10.1510
Epoch [1/3], Step [2311/3236], Loss: 2.9295, Perplexity: 18.7188
Epoch [1/3], Step [2312/3236], Loss: 2.3629, Perplexity: 10.6214
Epoch [1/3], Step [2313/3236], Loss: 2.3295, Perplexity: 10.2732
Epoch [1/3], Step [2314/3236], Loss: 2.2641, Perplexity: 9.6226
Epoch [1/3], Step [2315/3236], Loss: 2.2471, Perplexity: 9.4601
Epoch [1/3], Step [2316/3236], Loss: 2.3056, Perplexity: 10.0304
Epoch [1/3], Step [2317/3236], Loss: 2.2480, Perplexity: 9.4684
Epoch [1/3], Step [2318/3236], Loss: 2.2439, Perplexity: 9.4299
Epoch [1/3], Step [2319/3236], Loss: 2.2233, Perplexity: 9.2381
Epoch [1/3], Step [2320/3236], Loss: 2.2585, Perplexity: 9.5691
Epoch [1/3], Step [2321/3236], Loss: 2.6056, Perplexity: 13.5394
Epoch [1/3], Step [2322/3236], Loss: 2.9471, Perplexity: 19.0499
Epoch [1/3], Step [2323/3236], Loss: 2.6474, Perplexity: 14.1178
Epoch [1/3], Step [2324/3236], Loss: 2.3454, Perplexity: 10.4370
Epoch [1/3], Step [2325/3236], Loss: 2.5520, Perplexity: 12.8331
Epoch [1/3], Step [2326/3236], Loss: 2.4438, Perplexity: 11.5172
Epoch [1/3], Step [2327/3236], Loss: 2.1891, Perplexity: 8.9270
Epoch [1/3], Step [2328/3236], Loss: 2.4056, Perplexity: 11.0846
Epoch [1/3], Step [2329/3236], Loss: 2.1191, Perplexity: 8.3240
Epoch [1/3], Step [2330/3236], Loss: 2.6066, Perplexity: 13.5530
Epoch [1/3], Step [2331/3236], Loss: 2.6524, Perplexity: 14.1875
Epoch [1/3], Step [2332/3236], Loss: 2.5686, Perplexity: 13.0474
Epoch [1/3], Step [2333/3236], Loss: 2.2882, Perplexity: 9.8572
Epoch [1/3], Step [2334/3236], Loss: 2.2344, Perplexity: 9.3408
Epoch [1/3], Step [2335/3236], Loss: 2.0914, Perplexity: 8.0964
Epoch [1/3], Step [2336/3236], Loss: 2.1683, Perplexity: 8.7435
Epoch [1/3], Step [2337/3236], Loss: 2.3311, Perplexity: 10.2893
Epoch [1/3], Step [2338/3236], Loss: 2.2148, Perplexity: 9.1595
Epoch [1/3], Step [2339/3236], Loss: 2.2140, Perplexity: 9.1524
Epoch [1/3], Step [2340/3236], Loss: 2.4146, Perplexity: 11.1851
Epoch [1/3], Step [2341/3236], Loss: 2.2657, Perplexity: 9.6378
Epoch [1/3], Step [2342/3236], Loss: 2.4289, Perplexity: 11.3462
Epoch [1/3], Step [2343/3236], Loss: 2.5814, Perplexity: 13.2158
Epoch [1/3], Step [2344/3236], Loss: 2.7081, Perplexity: 15.0003
Epoch [1/3], Step [2345/3236], Loss: 2.2189, Perplexity: 9.1976
Epoch [1/3], Step [2346/3236], Loss: 2.3167, Perplexity: 10.1424
Epoch [1/3], Step [2347/3236], Loss: 2.2591, Perplexity: 9.5740
Epoch [1/3], Step [2348/3236], Loss: 2.2662, Perplexity: 9.6431
Epoch [1/3], Step [2349/3236], Loss: 2.4354, Perplexity: 11.4208
Epoch [1/3], Step [2350/3236], Loss: 2.3351, Perplexity: 10.3308
Epoch [1/3], Step [2351/3236], Loss: 2.2084, Perplexity: 9.1015
Epoch [1/3], Step [2352/3236], Loss: 2.2303, Perplexity: 9.3023
Epoch [1/3], Step [2353/3236], Loss: 2.1919, Perplexity: 8.9524
Epoch [1/3], Step [2354/3236], Loss: 2.4887, Perplexity: 12.0460
Epoch [1/3], Step [2355/3236], Loss: 2.2798, Perplexity: 9.7746
Epoch [1/3], Step [2356/3236], Loss: 2.9857, Perplexity: 19.8003
Epoch [1/3], Step [2357/3236], Loss: 2.2720, Perplexity: 9.6992
Epoch [1/3], Step [2358/3236], Loss: 2.5665, Perplexity: 13.0201
Epoch [1/3], Step [2359/3236], Loss: 2.0367, Perplexity: 7.6655
Epoch [1/3], Step [2360/3236], Loss: 2.4505, Perplexity: 11.5940
Epoch [1/3], Step [2361/3236], Loss: 2.2704, Perplexity: 9.6832
Epoch [1/3], Step [2362/3236], Loss: 2.8083, Perplexity: 16.5810
Epoch [1/3], Step [2363/3236], Loss: 2.1989, Perplexity: 9.0153
Epoch [1/3], Step [2364/3236], Loss: 2.1368, Perplexity: 8.4725
Epoch [1/3], Step [2365/3236], Loss: 2.3151, Perplexity: 10.1255
Epoch [1/3], Step [2366/3236], Loss: 2.3649, Perplexity: 10.6432
Epoch [1/3], Step [2367/3236], Loss: 2.1277, Perplexity: 8.3959
Epoch [1/3], Step [2368/3236], Loss: 2.2236, Perplexity: 9.2408
Epoch [1/3], Step [2369/3236], Loss: 2.1118, Perplexity: 8.2630
Epoch [1/3], Step [2370/3236], Loss: 2.2384, Perplexity: 9.3787
Epoch [1/3], Step [2371/3236], Loss: 2.3649, Perplexity: 10.6430
Epoch [1/3], Step [2372/3236], Loss: 2.1402, Perplexity: 8.5010
Epoch [1/3], Step [2373/3236], Loss: 2.3101, Perplexity: 10.0759
Epoch [1/3], Step [2374/3236], Loss: 2.6544, Perplexity: 14.2167
Epoch [1/3], Step [2375/3236], Loss: 2.1538, Perplexity: 8.6179
Epoch [1/3], Step [2376/3236], Loss: 2.2292, Perplexity: 9.2928
Epoch [1/3], Step [2377/3236], Loss: 2.1642, Perplexity: 8.7074
Epoch [1/3], Step [2378/3236], Loss: 2.2068, Perplexity: 9.0866
Epoch [1/3], Step [2379/3236], Loss: 2.1581, Perplexity: 8.6550
Epoch [1/3], Step [2380/3236], Loss: 2.3505, Perplexity: 10.4909
Epoch [1/3], Step [2381/3236], Loss: 2.3556, Perplexity: 10.5440
Epoch [1/3], Step [2382/3236], Loss: 2.1215, Perplexity: 8.3435
Epoch [1/3], Step [2383/3236], Loss: 2.3714, Perplexity: 10.7121
Epoch [1/3], Step [2384/3236], Loss: 2.4302, Perplexity: 11.3608
Epoch [1/3], Step [2385/3236], Loss: 2.6413, Perplexity: 14.0321
Epoch [1/3], Step [2386/3236], Loss: 2.1723, Perplexity: 8.7781
Epoch [1/3], Step [2387/3236], Loss: 2.2053, Perplexity: 9.0728
Epoch [1/3], Step [2388/3236], Loss: 2.2218, Perplexity: 9.2238
Epoch [1/3], Step [2389/3236], Loss: 2.6552, Perplexity: 14.2279
Epoch [1/3], Step [2390/3236], Loss: 2.2297, Perplexity: 9.2974
Epoch [1/3], Step [2391/3236], Loss: 2.9871, Perplexity: 19.8275
Epoch [1/3], Step [2392/3236], Loss: 2.4586, Perplexity: 11.6887
Epoch [1/3], Step [2393/3236], Loss: 2.2273, Perplexity: 9.2750
Epoch [1/3], Step [2394/3236], Loss: 2.2606, Perplexity: 9.5891
Epoch [1/3], Step [2395/3236], Loss: 2.3390, Perplexity: 10.3711
Epoch [1/3], Step [2396/3236], Loss: 2.1352, Perplexity: 8.4587
Epoch [1/3], Step [2397/3236], Loss: 2.2975, Perplexity: 9.9489
Epoch [1/3], Step [2398/3236], Loss: 2.2582, Perplexity: 9.5659
Epoch [1/3], Step [2399/3236], Loss: 2.1415, Perplexity: 8.5120
Epoch [1/3], Step [2400/3236], Loss: 2.2878, Perplexity: 9.8533
Epoch [1/3], Step [2401/3236], Loss: 2.0149, Perplexity: 7.5002
Epoch [1/3], Step [2402/3236], Loss: 2.1090, Perplexity: 8.2398
Epoch [1/3], Step [2403/3236], Loss: 2.2125, Perplexity: 9.1388
Epoch [1/3], Step [2404/3236], Loss: 2.2524, Perplexity: 9.5104
Epoch [1/3], Step [2405/3236], Loss: 2.3503, Perplexity: 10.4889
Epoch [1/3], Step [2406/3236], Loss: 2.3511, Perplexity: 10.4975
Epoch [1/3], Step [2407/3236], Loss: 2.1571, Perplexity: 8.6460
Epoch [1/3], Step [2408/3236], Loss: 2.1038, Perplexity: 8.1974
Epoch [1/3], Step [2409/3236], Loss: 3.4829, Perplexity: 32.5547
Epoch [1/3], Step [2410/3236], Loss: 2.5441, Perplexity: 12.7315
Epoch [1/3], Step [2411/3236], Loss: 2.3014, Perplexity: 9.9881
Epoch [1/3], Step [2412/3236], Loss: 2.2454, Perplexity: 9.4441
Epoch [1/3], Step [2413/3236], Loss: 2.1335, Perplexity: 8.4443
Epoch [1/3], Step [2414/3236], Loss: 2.3193, Perplexity: 10.1681
Epoch [1/3], Step [2415/3236], Loss: 2.1186, Perplexity: 8.3192
Epoch [1/3], Step [2416/3236], Loss: 2.2616, Perplexity: 9.5983
Epoch [1/3], Step [2417/3236], Loss: 2.3408, Perplexity: 10.3897
Epoch [1/3], Step [2418/3236], Loss: 2.3140, Perplexity: 10.1152
Epoch [1/3], Step [2419/3236], Loss: 2.4761, Perplexity: 11.8945
Epoch [1/3], Step [2420/3236], Loss: 2.0800, Perplexity: 8.0042
Epoch [1/3], Step [2421/3236], Loss: 2.2836, Perplexity: 9.8121
Epoch [1/3], Step [2422/3236], Loss: 2.8166, Perplexity: 16.7200
Epoch [1/3], Step [2423/3236], Loss: 2.9035, Perplexity: 18.2370
Epoch [1/3], Step [2424/3236], Loss: 2.4280, Perplexity: 11.3367
Epoch [1/3], Step [2425/3236], Loss: 2.2048, Perplexity: 9.0681
Epoch [1/3], Step [2426/3236], Loss: 2.2316, Perplexity: 9.3152
Epoch [1/3], Step [2427/3236], Loss: 2.5355, Perplexity: 12.6229
Epoch [1/3], Step [2428/3236], Loss: 2.2881, Perplexity: 9.8567
Epoch [1/3], Step [2429/3236], Loss: 2.2844, Perplexity: 9.8196
Epoch [1/3], Step [2430/3236], Loss: 2.2867, Perplexity: 9.8427
Epoch [1/3], Step [2431/3236], Loss: 2.1840, Perplexity: 8.8821
Epoch [1/3], Step [2432/3236], Loss: 2.2771, Perplexity: 9.7482
Epoch [1/3], Step [2433/3236], Loss: 2.2793, Perplexity: 9.7697
Epoch [1/3], Step [2434/3236], Loss: 2.2579, Perplexity: 9.5626
Epoch [1/3], Step [2435/3236], Loss: 2.4357, Perplexity: 11.4243
Epoch [1/3], Step [2436/3236], Loss: 2.4883, Perplexity: 12.0406
Epoch [1/3], Step [2437/3236], Loss: 2.3087, Perplexity: 10.0617
Epoch [1/3], Step [2438/3236], Loss: 3.1880, Perplexity: 24.2400
Epoch [1/3], Step [2439/3236], Loss: 2.2391, Perplexity: 9.3847
Epoch [1/3], Step [2440/3236], Loss: 2.5096, Perplexity: 12.2998
Epoch [1/3], Step [2441/3236], Loss: 2.3136, Perplexity: 10.1108
Epoch [1/3], Step [2442/3236], Loss: 2.3149, Perplexity: 10.1238
Epoch [1/3], Step [2443/3236], Loss: 2.4037, Perplexity: 11.0643
Epoch [1/3], Step [2444/3236], Loss: 2.2619, Perplexity: 9.6013
Epoch [1/3], Step [2445/3236], Loss: 2.3005, Perplexity: 9.9791
Epoch [1/3], Step [2446/3236], Loss: 2.2782, Perplexity: 9.7592
Epoch [1/3], Step [2447/3236], Loss: 2.3999, Perplexity: 11.0220
Epoch [1/3], Step [2448/3236], Loss: 2.5310, Perplexity: 12.5655
Epoch [1/3], Step [2449/3236], Loss: 2.5204, Perplexity: 12.4339
Epoch [1/3], Step [2450/3236], Loss: 2.6040, Perplexity: 13.5183
Epoch [1/3], Step [2451/3236], Loss: 2.0322, Perplexity: 7.6310
Epoch [1/3], Step [2452/3236], Loss: 2.3146, Perplexity: 10.1206
Epoch [1/3], Step [2453/3236], Loss: 2.2952, Perplexity: 9.9260
Epoch [1/3], Step [2454/3236], Loss: 2.2009, Perplexity: 9.0334
Epoch [1/3], Step [2455/3236], Loss: 2.3423, Perplexity: 10.4054
Epoch [1/3], Step [2456/3236], Loss: 2.0386, Perplexity: 7.6796
Epoch [1/3], Step [2457/3236], Loss: 2.3903, Perplexity: 10.9166
Epoch [1/3], Step [2458/3236], Loss: 2.3057, Perplexity: 10.0313
Epoch [1/3], Step [2459/3236], Loss: 2.3770, Perplexity: 10.7730
Epoch [1/3], Step [2460/3236], Loss: 2.2788, Perplexity: 9.7648
Epoch [1/3], Step [2461/3236], Loss: 2.2547, Perplexity: 9.5326
Epoch [1/3], Step [2462/3236], Loss: 2.1170, Perplexity: 8.3065
Epoch [1/3], Step [2463/3236], Loss: 2.3127, Perplexity: 10.1021
Epoch [1/3], Step [2464/3236], Loss: 2.1884, Perplexity: 8.9208
Epoch [1/3], Step [2465/3236], Loss: 2.1633, Perplexity: 8.7000
Epoch [1/3], Step [2466/3236], Loss: 2.4117, Perplexity: 11.1525
Epoch [1/3], Step [2467/3236], Loss: 2.1819, Perplexity: 8.8630
Epoch [1/3], Step [2468/3236], Loss: 2.5672, Perplexity: 13.0299
Epoch [1/3], Step [2469/3236], Loss: 2.5807, Perplexity: 13.2061
Epoch [1/3], Step [2470/3236], Loss: 2.1543, Perplexity: 8.6216
Epoch [1/3], Step [2471/3236], Loss: 2.2793, Perplexity: 9.7698
Epoch [1/3], Step [2472/3236], Loss: 2.4675, Perplexity: 11.7931
Epoch [1/3], Step [2473/3236], Loss: 2.2892, Perplexity: 9.8672
Epoch [1/3], Step [2474/3236], Loss: 2.2085, Perplexity: 9.1021
Epoch [1/3], Step [2475/3236], Loss: 2.2940, Perplexity: 9.9147
Epoch [1/3], Step [2476/3236], Loss: 2.0637, Perplexity: 7.8751
Epoch [1/3], Step [2477/3236], Loss: 2.1692, Perplexity: 8.7514
Epoch [1/3], Step [2478/3236], Loss: 2.4103, Perplexity: 11.1371
Epoch [1/3], Step [2479/3236], Loss: 2.0753, Perplexity: 7.9665
Epoch [1/3], Step [2480/3236], Loss: 2.1748, Perplexity: 8.8002
Epoch [1/3], Step [2481/3236], Loss: 2.4001, Perplexity: 11.0245
Epoch [1/3], Step [2482/3236], Loss: 2.3158, Perplexity: 10.1328
Epoch [1/3], Step [2483/3236], Loss: 2.3995, Perplexity: 11.0176
Epoch [1/3], Step [2484/3236], Loss: 2.1703, Perplexity: 8.7610
Epoch [1/3], Step [2485/3236], Loss: 2.2228, Perplexity: 9.2331
Epoch [1/3], Step [2486/3236], Loss: 2.3204, Perplexity: 10.1796
Epoch [1/3], Step [2487/3236], Loss: 2.2629, Perplexity: 9.6111
Epoch [1/3], Step [2488/3236], Loss: 2.3940, Perplexity: 10.9568
Epoch [1/3], Step [2489/3236], Loss: 2.2832, Perplexity: 9.8083
Epoch [1/3], Step [2490/3236], Loss: 2.2141, Perplexity: 9.1529
Epoch [1/3], Step [2491/3236], Loss: 2.3126, Perplexity: 10.1010
Epoch [1/3], Step [2492/3236], Loss: 2.2198, Perplexity: 9.2055
Epoch [1/3], Step [2493/3236], Loss: 2.0752, Perplexity: 7.9658
Epoch [1/3], Step [2494/3236], Loss: 2.0255, Perplexity: 7.5796
Epoch [1/3], Step [2495/3236], Loss: 2.2970, Perplexity: 9.9444
Epoch [1/3], Step [2496/3236], Loss: 2.3985, Perplexity: 11.0063
Epoch [1/3], Step [2497/3236], Loss: 2.2694, Perplexity: 9.6736
Epoch [1/3], Step [2498/3236], Loss: 2.4918, Perplexity: 12.0833
Epoch [1/3], Step [2499/3236], Loss: 2.6229, Perplexity: 13.7750
Epoch [1/3], Step [2500/3236], Loss: 2.0440, Perplexity: 7.7211
Epoch [1/3], Step [2501/3236], Loss: 2.2127, Perplexity: 9.1408
Epoch [1/3], Step [2502/3236], Loss: 2.2544, Perplexity: 9.5297
Epoch [1/3], Step [2503/3236], Loss: 2.4123, Perplexity: 11.1601
Epoch [1/3], Step [2504/3236], Loss: 2.3059, Perplexity: 10.0333
Epoch [1/3], Step [2505/3236], Loss: 2.5990, Perplexity: 13.4509
Epoch [1/3], Step [2506/3236], Loss: 2.3486, Perplexity: 10.4704
Epoch [1/3], Step [2507/3236], Loss: 2.1005, Perplexity: 8.1704
Epoch [1/3], Step [2508/3236], Loss: 2.4969, Perplexity: 12.1448
Epoch [1/3], Step [2509/3236], Loss: 2.2592, Perplexity: 9.5759
Epoch [1/3], Step [2510/3236], Loss: 2.3274, Perplexity: 10.2513
Epoch [1/3], Step [2511/3236], Loss: 2.2842, Perplexity: 9.8175
Epoch [1/3], Step [2512/3236], Loss: 2.4109, Perplexity: 11.1436
Epoch [1/3], Step [2513/3236], Loss: 2.0573, Perplexity: 7.8247
Epoch [1/3], Step [2514/3236], Loss: 3.2009, Perplexity: 24.5545
Epoch [1/3], Step [2515/3236], Loss: 2.1648, Perplexity: 8.7129
Epoch [1/3], Step [2516/3236], Loss: 2.2680, Perplexity: 9.6603
Epoch [1/3], Step [2517/3236], Loss: 2.3789, Perplexity: 10.7928
Epoch [1/3], Step [2518/3236], Loss: 2.3648, Perplexity: 10.6421
Epoch [1/3], Step [2519/3236], Loss: 2.5255, Perplexity: 12.4967
Epoch [1/3], Step [2520/3236], Loss: 2.1270, Perplexity: 8.3895
Epoch [1/3], Step [2521/3236], Loss: 2.1791, Perplexity: 8.8385
Epoch [1/3], Step [2522/3236], Loss: 2.3089, Perplexity: 10.0630
Epoch [1/3], Step [2523/3236], Loss: 2.3788, Perplexity: 10.7923
Epoch [1/3], Step [2524/3236], Loss: 2.2175, Perplexity: 9.1845
Epoch [1/3], Step [2525/3236], Loss: 2.2631, Perplexity: 9.6127
Epoch [1/3], Step [2526/3236], Loss: 2.2623, Perplexity: 9.6055
Epoch [1/3], Step [2527/3236], Loss: 2.3565, Perplexity: 10.5542
Epoch [1/3], Step [2528/3236], Loss: 2.2571, Perplexity: 9.5557
Epoch [1/3], Step [2529/3236], Loss: 2.1564, Perplexity: 8.6399
Epoch [1/3], Step [2530/3236], Loss: 2.2181, Perplexity: 9.1897
Epoch [1/3], Step [2531/3236], Loss: 2.1532, Perplexity: 8.6127
Epoch [1/3], Step [2532/3236], Loss: 2.2499, Perplexity: 9.4870
Epoch [1/3], Step [2533/3236], Loss: 2.1914, Perplexity: 8.9479
Epoch [1/3], Step [2534/3236], Loss: 2.1627, Perplexity: 8.6950
Epoch [1/3], Step [2535/3236], Loss: 2.1153, Perplexity: 8.2925
Epoch [1/3], Step [2536/3236], Loss: 2.3181, Perplexity: 10.1567
Epoch [1/3], Step [2537/3236], Loss: 2.1067, Perplexity: 8.2210
Epoch [1/3], Step [2538/3236], Loss: 2.5330, Perplexity: 12.5910
Epoch [1/3], Step [2539/3236], Loss: 2.1423, Perplexity: 8.5192
Epoch [1/3], Step [2540/3236], Loss: 2.3514, Perplexity: 10.5001
Epoch [1/3], Step [2541/3236], Loss: 2.2543, Perplexity: 9.5285
Epoch [1/3], Step [2542/3236], Loss: 2.1719, Perplexity: 8.7750
Epoch [1/3], Step [2543/3236], Loss: 2.0694, Perplexity: 7.9199
Epoch [1/3], Step [2544/3236], Loss: 2.2554, Perplexity: 9.5393
Epoch [1/3], Step [2545/3236], Loss: 2.1086, Perplexity: 8.2366
Epoch [1/3], Step [2546/3236], Loss: 2.5221, Perplexity: 12.4543
Epoch [1/3], Step [2547/3236], Loss: 2.4046, Perplexity: 11.0737
Epoch [1/3], Step [2548/3236], Loss: 2.2881, Perplexity: 9.8558
Epoch [1/3], Step [2549/3236], Loss: 2.1661, Perplexity: 8.7242
Epoch [1/3], Step [2550/3236], Loss: 2.1943, Perplexity: 8.9734
Epoch [1/3], Step [2551/3236], Loss: 2.1992, Perplexity: 9.0180
Epoch [1/3], Step [2552/3236], Loss: 2.1544, Perplexity: 8.6227
Epoch [1/3], Step [2553/3236], Loss: 2.2281, Perplexity: 9.2818
Epoch [1/3], Step [2554/3236], Loss: 2.2246, Perplexity: 9.2499
Epoch [1/3], Step [2555/3236], Loss: 2.1514, Perplexity: 8.5972
Epoch [1/3], Step [2556/3236], Loss: 2.0886, Perplexity: 8.0738
Epoch [1/3], Step [2557/3236], Loss: 2.2857, Perplexity: 9.8327
Epoch [1/3], Step [2558/3236], Loss: 2.8806, Perplexity: 17.8250
Epoch [1/3], Step [2559/3236], Loss: 2.2581, Perplexity: 9.5649
Epoch [1/3], Step [2560/3236], Loss: 2.3632, Perplexity: 10.6247
Epoch [1/3], Step [2561/3236], Loss: 2.2006, Perplexity: 9.0308
Epoch [1/3], Step [2562/3236], Loss: 2.2991, Perplexity: 9.9652
Epoch [1/3], Step [2563/3236], Loss: 2.3007, Perplexity: 9.9810
Epoch [1/3], Step [2564/3236], Loss: 2.1087, Perplexity: 8.2377
Epoch [1/3], Step [2565/3236], Loss: 2.3597, Perplexity: 10.5878
Epoch [1/3], Step [2566/3236], Loss: 2.2743, Perplexity: 9.7207
Epoch [1/3], Step [2567/3236], Loss: 2.0565, Perplexity: 7.8186
Epoch [1/3], Step [2568/3236], Loss: 2.5184, Perplexity: 12.4093
Epoch [1/3], Step [2569/3236], Loss: 2.2411, Perplexity: 9.4034
Epoch [1/3], Step [2570/3236], Loss: 2.2852, Perplexity: 9.8280
Epoch [1/3], Step [2571/3236], Loss: 2.1740, Perplexity: 8.7938
Epoch [1/3], Step [2572/3236], Loss: 2.6605, Perplexity: 14.3039
Epoch [1/3], Step [2573/3236], Loss: 2.2625, Perplexity: 9.6072
Epoch [1/3], Step [2574/3236], Loss: 2.7577, Perplexity: 15.7634
Epoch [1/3], Step [2575/3236], Loss: 2.1955, Perplexity: 8.9849
Epoch [1/3], Step [2576/3236], Loss: 2.1281, Perplexity: 8.3991
Epoch [1/3], Step [2577/3236], Loss: 2.1848, Perplexity: 8.8886
Epoch [1/3], Step [2578/3236], Loss: 2.8307, Perplexity: 16.9576
Epoch [1/3], Step [2579/3236], Loss: 2.6710, Perplexity: 14.4540
Epoch [1/3], Step [2580/3236], Loss: 2.1716, Perplexity: 8.7721
Epoch [1/3], Step [2581/3236], Loss: 2.0300, Perplexity: 7.6137
Epoch [1/3], Step [2582/3236], Loss: 2.2001, Perplexity: 9.0263
Epoch [1/3], Step [2583/3236], Loss: 2.1631, Perplexity: 8.6977
Epoch [1/3], Step [2584/3236], Loss: 2.1527, Perplexity: 8.6081
Epoch [1/3], Step [2585/3236], Loss: 2.1659, Perplexity: 8.7223
Epoch [1/3], Step [2586/3236], Loss: 2.4562, Perplexity: 11.6602
Epoch [1/3], Step [2587/3236], Loss: 2.4500, Perplexity: 11.5884
Epoch [1/3], Step [2588/3236], Loss: 2.0955, Perplexity: 8.1298
Epoch [1/3], Step [2589/3236], Loss: 2.2509, Perplexity: 9.4967
Epoch [1/3], Step [2590/3236], Loss: 2.2712, Perplexity: 9.6906
Epoch [1/3], Step [2591/3236], Loss: 2.2475, Perplexity: 9.4644
Epoch [1/3], Step [2592/3236], Loss: 2.6743, Perplexity: 14.5019
Epoch [1/3], Step [2593/3236], Loss: 2.1791, Perplexity: 8.8383
Epoch [1/3], Step [2594/3236], Loss: 2.0289, Perplexity: 7.6054
Epoch [1/3], Step [2595/3236], Loss: 2.4499, Perplexity: 11.5875
Epoch [1/3], Step [2596/3236], Loss: 2.2553, Perplexity: 9.5382
Epoch [1/3], Step [2597/3236], Loss: 2.1272, Perplexity: 8.3911
Epoch [1/3], Step [2598/3236], Loss: 2.2908, Perplexity: 9.8830
Epoch [1/3], Step [2599/3236], Loss: 2.2041, Perplexity: 9.0620
Epoch [1/3], Step [2600/3236], Loss: 2.2758, Perplexity: 9.7355
Epoch [1/3], Step [2601/3236], Loss: 2.0820, Perplexity: 8.0207
Epoch [1/3], Step [2602/3236], Loss: 2.3955, Perplexity: 10.9734
Epoch [1/3], Step [2603/3236], Loss: 2.2992, Perplexity: 9.9660
Epoch [1/3], Step [2604/3236], Loss: 2.5056, Perplexity: 12.2504
Epoch [1/3], Step [2605/3236], Loss: 2.8177, Perplexity: 16.7380
Epoch [1/3], Step [2606/3236], Loss: 2.2041, Perplexity: 9.0618
Epoch [1/3], Step [2607/3236], Loss: 2.0308, Perplexity: 7.6202
Epoch [1/3], Step [2608/3236], Loss: 3.0669, Perplexity: 21.4750
Epoch [1/3], Step [2609/3236], Loss: 2.1251, Perplexity: 8.3741
Epoch [1/3], Step [2610/3236], Loss: 2.5378, Perplexity: 12.6522
Epoch [1/3], Step [2611/3236], Loss: 2.2750, Perplexity: 9.7283
Epoch [1/3], Step [2612/3236], Loss: 2.2012, Perplexity: 9.0359
Epoch [1/3], Step [2613/3236], Loss: 2.1088, Perplexity: 8.2387
Epoch [1/3], Step [2614/3236], Loss: 2.4441, Perplexity: 11.5198
Epoch [1/3], Step [2615/3236], Loss: 2.1139, Perplexity: 8.2809
Epoch [1/3], Step [2616/3236], Loss: 2.2018, Perplexity: 9.0412
Epoch [1/3], Step [2617/3236], Loss: 2.1150, Perplexity: 8.2894
Epoch [1/3], Step [2618/3236], Loss: 2.4606, Perplexity: 11.7120
Epoch [1/3], Step [2619/3236], Loss: 2.1588, Perplexity: 8.6610
Epoch [1/3], Step [2620/3236], Loss: 2.2505, Perplexity: 9.4927
Epoch [1/3], Step [2621/3236], Loss: 2.1733, Perplexity: 8.7868
Epoch [1/3], Step [2622/3236], Loss: 2.3026, Perplexity: 9.9999
Epoch [1/3], Step [2623/3236], Loss: 2.1993, Perplexity: 9.0191
Epoch [1/3], Step [2624/3236], Loss: 2.3721, Perplexity: 10.7203
Epoch [1/3], Step [2625/3236], Loss: 2.1539, Perplexity: 8.6183
Epoch [1/3], Step [2626/3236], Loss: 2.2817, Perplexity: 9.7937
Epoch [1/3], Step [2627/3236], Loss: 2.3208, Perplexity: 10.1834
Epoch [1/3], Step [2628/3236], Loss: 2.6830, Perplexity: 14.6286
Epoch [1/3], Step [2629/3236], Loss: 2.0845, Perplexity: 8.0405
Epoch [1/3], Step [2630/3236], Loss: 2.2069, Perplexity: 9.0876
Epoch [1/3], Step [2631/3236], Loss: 2.1033, Perplexity: 8.1931
Epoch [1/3], Step [2632/3236], Loss: 2.0446, Perplexity: 7.7264
Epoch [1/3], Step [2633/3236], Loss: 2.2771, Perplexity: 9.7483
Epoch [1/3], Step [2634/3236], Loss: 2.1291, Perplexity: 8.4070
Epoch [1/3], Step [2635/3236], Loss: 2.7553, Perplexity: 15.7264
Epoch [1/3], Step [2636/3236], Loss: 2.1386, Perplexity: 8.4872
Epoch [1/3], Step [2637/3236], Loss: 2.1406, Perplexity: 8.5043
Epoch [1/3], Step [2638/3236], Loss: 2.1177, Perplexity: 8.3119
Epoch [1/3], Step [2639/3236], Loss: 2.7801, Perplexity: 16.1213
Epoch [1/3], Step [2640/3236], Loss: 2.2381, Perplexity: 9.3751
Epoch [1/3], Step [2641/3236], Loss: 2.3572, Perplexity: 10.5616
Epoch [1/3], Step [2642/3236], Loss: 2.4992, Perplexity: 12.1733
Epoch [1/3], Step [2643/3236], Loss: 2.0341, Perplexity: 7.6455
Epoch [1/3], Step [2644/3236], Loss: 2.2130, Perplexity: 9.1427
Epoch [1/3], Step [2645/3236], Loss: 2.1512, Perplexity: 8.5951
Epoch [1/3], Step [2646/3236], Loss: 2.2698, Perplexity: 9.6777
Epoch [1/3], Step [2647/3236], Loss: 2.3748, Perplexity: 10.7492
Epoch [1/3], Step [2648/3236], Loss: 2.1463, Perplexity: 8.5527
Epoch [1/3], Step [2649/3236], Loss: 2.2690, Perplexity: 9.6697
Epoch [1/3], Step [2650/3236], Loss: 2.7293, Perplexity: 15.3227
Epoch [1/3], Step [2651/3236], Loss: 2.1344, Perplexity: 8.4516
Epoch [1/3], Step [2652/3236], Loss: 2.0404, Perplexity: 7.6938
Epoch [1/3], Step [2653/3236], Loss: 2.2041, Perplexity: 9.0622
Epoch [1/3], Step [2654/3236], Loss: 2.1239, Perplexity: 8.3641
Epoch [1/3], Step [2655/3236], Loss: 2.2123, Perplexity: 9.1368
Epoch [1/3], Step [2656/3236], Loss: 2.2332, Perplexity: 9.3300
Epoch [1/3], Step [2657/3236], Loss: 2.2189, Perplexity: 9.1971
Epoch [1/3], Step [2658/3236], Loss: 2.0761, Perplexity: 7.9734
Epoch [1/3], Step [2659/3236], Loss: 2.1952, Perplexity: 8.9815
Epoch [1/3], Step [2660/3236], Loss: 2.0380, Perplexity: 7.6750
Epoch [1/3], Step [2661/3236], Loss: 2.1468, Perplexity: 8.5574
Epoch [1/3], Step [2662/3236], Loss: 2.0068, Perplexity: 7.4395
Epoch [1/3], Step [2663/3236], Loss: 2.6465, Perplexity: 14.1043
Epoch [1/3], Step [2664/3236], Loss: 2.7077, Perplexity: 14.9954
Epoch [1/3], Step [2665/3236], Loss: 2.0280, Perplexity: 7.5987
Epoch [1/3], Step [2666/3236], Loss: 2.1229, Perplexity: 8.3557
Epoch [1/3], Step [2667/3236], Loss: 2.5890, Perplexity: 13.3165
Epoch [1/3], Step [2668/3236], Loss: 2.1652, Perplexity: 8.7162
Epoch [1/3], Step [2669/3236], Loss: 2.3074, Perplexity: 10.0484
Epoch [1/3], Step [2670/3236], Loss: 2.2606, Perplexity: 9.5884
Epoch [1/3], Step [2671/3236], Loss: 2.5038, Perplexity: 12.2295
Epoch [1/3], Step [2672/3236], Loss: 2.1608, Perplexity: 8.6784
Epoch [1/3], Step [2673/3236], Loss: 2.2240, Perplexity: 9.2446
Epoch [1/3], Step [2674/3236], Loss: 2.2914, Perplexity: 9.8883
Epoch [1/3], Step [2675/3236], Loss: 2.1245, Perplexity: 8.3687
Epoch [1/3], Step [2676/3236], Loss: 2.6520, Perplexity: 14.1827
Epoch [1/3], Step [2677/3236], Loss: 2.1304, Perplexity: 8.4179
Epoch [1/3], Step [2678/3236], Loss: 2.2182, Perplexity: 9.1908
Epoch [1/3], Step [2679/3236], Loss: 2.1105, Perplexity: 8.2522
Epoch [1/3], Step [2680/3236], Loss: 2.1648, Perplexity: 8.7126
Epoch [1/3], Step [2681/3236], Loss: 2.0770, Perplexity: 7.9807
Epoch [1/3], Step [2682/3236], Loss: 2.0658, Perplexity: 7.8912
Epoch [1/3], Step [2683/3236], Loss: 2.2465, Perplexity: 9.4547
Epoch [1/3], Step [2684/3236], Loss: 2.2341, Perplexity: 9.3377
Epoch [1/3], Step [2685/3236], Loss: 2.2560, Perplexity: 9.5448
Epoch [1/3], Step [2686/3236], Loss: 2.3435, Perplexity: 10.4179
Epoch [1/3], Step [2687/3236], Loss: 2.2338, Perplexity: 9.3352
Epoch [1/3], Step [2688/3236], Loss: 2.3727, Perplexity: 10.7259
Epoch [1/3], Step [2689/3236], Loss: 2.3008, Perplexity: 9.9825
Epoch [1/3], Step [2690/3236], Loss: 2.3590, Perplexity: 10.5801
Epoch [1/3], Step [2691/3236], Loss: 2.0792, Perplexity: 7.9984
Epoch [1/3], Step [2692/3236], Loss: 2.3244, Perplexity: 10.2209
Epoch [1/3], Step [2693/3236], Loss: 2.1457, Perplexity: 8.5476
Epoch [1/3], Step [2694/3236], Loss: 2.2783, Perplexity: 9.7601
Epoch [1/3], Step [2695/3236], Loss: 2.3451, Perplexity: 10.4342
Epoch [1/3], Step [2696/3236], Loss: 2.1111, Perplexity: 8.2575
Epoch [1/3], Step [2697/3236], Loss: 2.4383, Perplexity: 11.4530
Epoch [1/3], Step [2698/3236], Loss: 2.0146, Perplexity: 7.4974
Epoch [1/3], Step [2699/3236], Loss: 2.2673, Perplexity: 9.6533
Epoch [1/3], Step [2700/3236], Loss: 2.0896, Perplexity: 8.0815
Epoch [1/3], Step [2701/3236], Loss: 2.1534, Perplexity: 8.6141
Epoch [1/3], Step [2702/3236], Loss: 2.0749, Perplexity: 7.9637
Epoch [1/3], Step [2703/3236], Loss: 2.5274, Perplexity: 12.5206
Epoch [1/3], Step [2704/3236], Loss: 2.5805, Perplexity: 13.2037
Epoch [1/3], Step [2705/3236], Loss: 2.1497, Perplexity: 8.5820
Epoch [1/3], Step [2706/3236], Loss: 2.3222, Perplexity: 10.1982
Epoch [1/3], Step [2707/3236], Loss: 2.1674, Perplexity: 8.7353
Epoch [1/3], Step [2708/3236], Loss: 2.3880, Perplexity: 10.8917
Epoch [1/3], Step [2709/3236], Loss: 2.1528, Perplexity: 8.6086
Epoch [1/3], Step [2710/3236], Loss: 2.2334, Perplexity: 9.3318
Epoch [1/3], Step [2711/3236], Loss: 2.5461, Perplexity: 12.7569
Epoch [1/3], Step [2712/3236], Loss: 2.5192, Perplexity: 12.4184
Epoch [1/3], Step [2713/3236], Loss: 2.2201, Perplexity: 9.2083
Epoch [1/3], Step [2714/3236], Loss: 2.0435, Perplexity: 7.7174
Epoch [1/3], Step [2715/3236], Loss: 2.1643, Perplexity: 8.7081
Epoch [1/3], Step [2716/3236], Loss: 2.1937, Perplexity: 8.9679
Epoch [1/3], Step [2717/3236], Loss: 2.3630, Perplexity: 10.6233
Epoch [1/3], Step [2718/3236], Loss: 1.9916, Perplexity: 7.3271
Epoch [1/3], Step [2719/3236], Loss: 2.0857, Perplexity: 8.0499
Epoch [1/3], Step [2720/3236], Loss: 2.1779, Perplexity: 8.8280
Epoch [1/3], Step [2721/3236], Loss: 2.2212, Perplexity: 9.2182
Epoch [1/3], Step [2722/3236], Loss: 2.5070, Perplexity: 12.2686
Epoch [1/3], Step [2723/3236], Loss: 2.5439, Perplexity: 12.7294
Epoch [1/3], Step [2724/3236], Loss: 2.1373, Perplexity: 8.4769
Epoch [1/3], Step [2725/3236], Loss: 3.3741, Perplexity: 29.1993
Epoch [1/3], Step [2726/3236], Loss: 2.2068, Perplexity: 9.0865
Epoch [1/3], Step [2727/3236], Loss: 2.1723, Perplexity: 8.7784
Epoch [1/3], Step [2728/3236], Loss: 2.2014, Perplexity: 9.0378
Epoch [1/3], Step [2729/3236], Loss: 2.1291, Perplexity: 8.4073
Epoch [1/3], Step [2730/3236], Loss: 2.0956, Perplexity: 8.1301
Epoch [1/3], Step [2731/3236], Loss: 2.2004, Perplexity: 9.0282
Epoch [1/3], Step [2732/3236], Loss: 2.2455, Perplexity: 9.4450
Epoch [1/3], Step [2733/3236], Loss: 2.1890, Perplexity: 8.9266
Epoch [1/3], Step [2734/3236], Loss: 2.1111, Perplexity: 8.2572
Epoch [1/3], Step [2735/3236], Loss: 2.1155, Perplexity: 8.2941
Epoch [1/3], Step [2736/3236], Loss: 2.2495, Perplexity: 9.4830
Epoch [1/3], Step [2737/3236], Loss: 2.2225, Perplexity: 9.2307
Epoch [1/3], Step [2738/3236], Loss: 2.2090, Perplexity: 9.1068
Epoch [1/3], Step [2739/3236], Loss: 2.3015, Perplexity: 9.9889
Epoch [1/3], Step [2740/3236], Loss: 3.1587, Perplexity: 23.5400
Epoch [1/3], Step [2741/3236], Loss: 2.3092, Perplexity: 10.0666
Epoch [1/3], Step [2742/3236], Loss: 2.1771, Perplexity: 8.8208
Epoch [1/3], Step [2743/3236], Loss: 2.2653, Perplexity: 9.6336
Epoch [1/3], Step [2744/3236], Loss: 2.2835, Perplexity: 9.8108
Epoch [1/3], Step [2745/3236], Loss: 2.5710, Perplexity: 13.0789
Epoch [1/3], Step [2746/3236], Loss: 2.1466, Perplexity: 8.5561
Epoch [1/3], Step [2747/3236], Loss: 2.1840, Perplexity: 8.8819
Epoch [1/3], Step [2748/3236], Loss: 2.2404, Perplexity: 9.3975
Epoch [1/3], Step [2749/3236], Loss: 2.0527, Perplexity: 7.7893
Epoch [1/3], Step [2750/3236], Loss: 2.1932, Perplexity: 8.9641
Epoch [1/3], Step [2751/3236], Loss: 2.4082, Perplexity: 11.1139
Epoch [1/3], Step [2752/3236], Loss: 2.2332, Perplexity: 9.3299
Epoch [1/3], Step [2753/3236], Loss: 2.1735, Perplexity: 8.7889
Epoch [1/3], Step [2754/3236], Loss: 2.1185, Perplexity: 8.3189
Epoch [1/3], Step [2755/3236], Loss: 2.3842, Perplexity: 10.8506
Epoch [1/3], Step [2756/3236], Loss: 2.2319, Perplexity: 9.3173
Epoch [1/3], Step [2757/3236], Loss: 2.2448, Perplexity: 9.4382
Epoch [1/3], Step [2758/3236], Loss: 2.2129, Perplexity: 9.1420
Epoch [1/3], Step [2759/3236], Loss: 2.6231, Perplexity: 13.7778
Epoch [1/3], Step [2760/3236], Loss: 2.4163, Perplexity: 11.2048
Epoch [1/3], Step [2761/3236], Loss: 2.1653, Perplexity: 8.7170
Epoch [1/3], Step [2762/3236], Loss: 2.4657, Perplexity: 11.7721
Epoch [1/3], Step [2763/3236], Loss: 2.4016, Perplexity: 11.0413
Epoch [1/3], Step [2764/3236], Loss: 2.1054, Perplexity: 8.2105
Epoch [1/3], Step [2765/3236], Loss: 2.1405, Perplexity: 8.5040
Epoch [1/3], Step [2766/3236], Loss: 2.2034, Perplexity: 9.0559
Epoch [1/3], Step [2767/3236], Loss: 2.2142, Perplexity: 9.1536
Epoch [1/3], Step [2768/3236], Loss: 2.2221, Perplexity: 9.2269
Epoch [1/3], Step [2769/3236], Loss: 2.6118, Perplexity: 13.6233
Epoch [1/3], Step [2770/3236], Loss: 2.1522, Perplexity: 8.6038
Epoch [1/3], Step [2771/3236], Loss: 2.2545, Perplexity: 9.5304
Epoch [1/3], Step [2772/3236], Loss: 2.1312, Perplexity: 8.4246
Epoch [1/3], Step [2773/3236], Loss: 2.5436, Perplexity: 12.7258
Epoch [1/3], Step [2774/3236], Loss: 2.0904, Perplexity: 8.0884
Epoch [1/3], Step [2775/3236], Loss: 2.7290, Perplexity: 15.3169
Epoch [1/3], Step [2776/3236], Loss: 3.9940, Perplexity: 54.2724
Epoch [1/3], Step [2777/3236], Loss: 2.0388, Perplexity: 7.6811
Epoch [1/3], Step [2778/3236], Loss: 2.5140, Perplexity: 12.3543
Epoch [1/3], Step [2779/3236], Loss: 2.1697, Perplexity: 8.7552
Epoch [1/3], Step [2780/3236], Loss: 2.2219, Perplexity: 9.2247
Epoch [1/3], Step [2781/3236], Loss: 2.3397, Perplexity: 10.3776
Epoch [1/3], Step [2782/3236], Loss: 2.0806, Perplexity: 8.0092
Epoch [1/3], Step [2783/3236], Loss: 2.3528, Perplexity: 10.5148
Epoch [1/3], Step [2784/3236], Loss: 2.3729, Perplexity: 10.7280
Epoch [1/3], Step [2785/3236], Loss: 2.2182, Perplexity: 9.1908
Epoch [1/3], Step [2786/3236], Loss: 2.1348, Perplexity: 8.4556
Epoch [1/3], Step [2787/3236], Loss: 2.3340, Perplexity: 10.3192
Epoch [1/3], Step [2788/3236], Loss: 2.1850, Perplexity: 8.8902
Epoch [1/3], Step [2789/3236], Loss: 2.1781, Perplexity: 8.8292
Epoch [1/3], Step [2790/3236], Loss: 2.4251, Perplexity: 11.3038
Epoch [1/3], Step [2791/3236], Loss: 2.2715, Perplexity: 9.6937
Epoch [1/3], Step [2792/3236], Loss: 2.0884, Perplexity: 8.0723
Epoch [1/3], Step [2793/3236], Loss: 2.1127, Perplexity: 8.2704
Epoch [1/3], Step [2794/3236], Loss: 2.1233, Perplexity: 8.3585
Epoch [1/3], Step [2795/3236], Loss: 2.2140, Perplexity: 9.1521
Epoch [1/3], Step [2796/3236], Loss: 2.0583, Perplexity: 7.8327
Epoch [1/3], Step [2797/3236], Loss: 2.1463, Perplexity: 8.5532
Epoch [1/3], Step [2798/3236], Loss: 2.2784, Perplexity: 9.7611
Epoch [1/3], Step [2799/3236], Loss: 2.2441, Perplexity: 9.4324
Epoch [1/3], Step [2800/3236], Loss: 2.2776, Perplexity: 9.7533
Epoch [1/3], Step [2801/3236], Loss: 2.0263, Perplexity: 7.5862
Epoch [1/3], Step [2802/3236], Loss: 2.1234, Perplexity: 8.3596
Epoch [1/3], Step [2803/3236], Loss: 2.2433, Perplexity: 9.4241
Epoch [1/3], Step [2804/3236], Loss: 2.1490, Perplexity: 8.5763
Epoch [1/3], Step [2805/3236], Loss: 2.1095, Perplexity: 8.2438
Epoch [1/3], Step [2806/3236], Loss: 2.4982, Perplexity: 12.1609
Epoch [1/3], Step [2807/3236], Loss: 2.1398, Perplexity: 8.4974
Epoch [1/3], Step [2808/3236], Loss: 2.1990, Perplexity: 9.0164
Epoch [1/3], Step [2809/3236], Loss: 2.7533, Perplexity: 15.6942
Epoch [1/3], Step [2810/3236], Loss: 2.2637, Perplexity: 9.6186
Epoch [1/3], Step [2811/3236], Loss: 2.3195, Perplexity: 10.1709
Epoch [1/3], Step [2812/3236], Loss: 2.2406, Perplexity: 9.3990
Epoch [1/3], Step [2813/3236], Loss: 2.1098, Perplexity: 8.2468
Epoch [1/3], Step [2814/3236], Loss: 2.3485, Perplexity: 10.4699
Epoch [1/3], Step [2815/3236], Loss: 2.3971, Perplexity: 10.9911
Epoch [1/3], Step [2816/3236], Loss: 2.4481, Perplexity: 11.5669
Epoch [1/3], Step [2817/3236], Loss: 2.3616, Perplexity: 10.6081
Epoch [1/3], Step [2818/3236], Loss: 2.1055, Perplexity: 8.2114
Epoch [1/3], Step [2819/3236], Loss: 2.3660, Perplexity: 10.6551
Epoch [1/3], Step [2820/3236], Loss: 2.1866, Perplexity: 8.9053
Epoch [1/3], Step [2821/3236], Loss: 2.4860, Perplexity: 12.0135
Epoch [1/3], Step [2822/3236], Loss: 2.1577, Perplexity: 8.6513
Epoch [1/3], Step [2823/3236], Loss: 2.3051, Perplexity: 10.0253
Epoch [1/3], Step [2824/3236], Loss: 2.1655, Perplexity: 8.7189
Epoch [1/3], Step [2825/3236], Loss: 2.2538, Perplexity: 9.5235
Epoch [1/3], Step [2826/3236], Loss: 2.2823, Perplexity: 9.7995
Epoch [1/3], Step [2827/3236], Loss: 2.2102, Perplexity: 9.1180
Epoch [1/3], Step [2828/3236], Loss: 2.2537, Perplexity: 9.5232
Epoch [1/3], Step [2829/3236], Loss: 2.1279, Perplexity: 8.3971
Epoch [1/3], Step [2830/3236], Loss: 2.1950, Perplexity: 8.9802
Epoch [1/3], Step [2831/3236], Loss: 2.2985, Perplexity: 9.9591
Epoch [1/3], Step [2832/3236], Loss: 2.5220, Perplexity: 12.4537
Epoch [1/3], Step [2833/3236], Loss: 2.0610, Perplexity: 7.8536
Epoch [1/3], Step [2834/3236], Loss: 2.3422, Perplexity: 10.4041
Epoch [1/3], Step [2835/3236], Loss: 2.4005, Perplexity: 11.0292
Epoch [1/3], Step [2836/3236], Loss: 2.2908, Perplexity: 9.8831
Epoch [1/3], Step [2837/3236], Loss: 2.2827, Perplexity: 9.8033
Epoch [1/3], Step [2838/3236], Loss: 2.2101, Perplexity: 9.1163
Epoch [1/3], Step [2839/3236], Loss: 2.6207, Perplexity: 13.7451
Epoch [1/3], Step [2840/3236], Loss: 2.4225, Perplexity: 11.2735
Epoch [1/3], Step [2841/3236], Loss: 2.1589, Perplexity: 8.6617
Epoch [1/3], Step [2842/3236], Loss: 2.2365, Perplexity: 9.3601
Epoch [1/3], Step [2843/3236], Loss: 2.1741, Perplexity: 8.7945
Epoch [1/3], Step [2844/3236], Loss: 2.3072, Perplexity: 10.0460
Epoch [1/3], Step [2845/3236], Loss: 2.1256, Perplexity: 8.3782
Epoch [1/3], Step [2846/3236], Loss: 2.1643, Perplexity: 8.7087
Epoch [1/3], Step [2847/3236], Loss: 2.2811, Perplexity: 9.7871
Epoch [1/3], Step [2848/3236], Loss: 2.2829, Perplexity: 9.8051
Epoch [1/3], Step [2849/3236], Loss: 2.2731, Perplexity: 9.7096
Epoch [1/3], Step [2850/3236], Loss: 2.0846, Perplexity: 8.0410
Epoch [1/3], Step [2851/3236], Loss: 2.1346, Perplexity: 8.4541
Epoch [1/3], Step [2852/3236], Loss: 2.3081, Perplexity: 10.0551
Epoch [1/3], Step [2853/3236], Loss: 2.3898, Perplexity: 10.9116
Epoch [1/3], Step [2854/3236], Loss: 2.1542, Perplexity: 8.6206
Epoch [1/3], Step [2855/3236], Loss: 2.2817, Perplexity: 9.7936
Epoch [1/3], Step [2856/3236], Loss: 2.4111, Perplexity: 11.1458
Epoch [1/3], Step [2857/3236], Loss: 2.2131, Perplexity: 9.1443
Epoch [1/3], Step [2858/3236], Loss: 2.6045, Perplexity: 13.5244
Epoch [1/3], Step [2859/3236], Loss: 2.1140, Perplexity: 8.2811
Epoch [1/3], Step [2860/3236], Loss: 2.2878, Perplexity: 9.8534
Epoch [1/3], Step [2861/3236], Loss: 2.0815, Perplexity: 8.0162
Epoch [1/3], Step [2862/3236], Loss: 2.5003, Perplexity: 12.1856
Epoch [1/3], Step [2863/3236], Loss: 2.1701, Perplexity: 8.7594
Epoch [1/3], Step [2864/3236], Loss: 2.1274, Perplexity: 8.3929
Epoch [1/3], Step [2865/3236], Loss: 2.1739, Perplexity: 8.7922
Epoch [1/3], Step [2866/3236], Loss: 2.1565, Perplexity: 8.6405
Epoch [1/3], Step [2867/3236], Loss: 2.1893, Perplexity: 8.9287
Epoch [1/3], Step [2868/3236], Loss: 2.2084, Perplexity: 9.1014
Epoch [1/3], Step [2869/3236], Loss: 2.3286, Perplexity: 10.2636
Epoch [1/3], Step [2870/3236], Loss: 2.1796, Perplexity: 8.8429
Epoch [1/3], Step [2871/3236], Loss: 2.2760, Perplexity: 9.7375
Epoch [1/3], Step [2872/3236], Loss: 2.1321, Perplexity: 8.4327
Epoch [1/3], Step [2873/3236], Loss: 2.1794, Perplexity: 8.8409
Epoch [1/3], Step [2874/3236], Loss: 2.2664, Perplexity: 9.6446
Epoch [1/3], Step [2875/3236], Loss: 2.2578, Perplexity: 9.5621
Epoch [1/3], Step [2876/3236], Loss: 2.3496, Perplexity: 10.4813
Epoch [1/3], Step [2877/3236], Loss: 2.1061, Perplexity: 8.2160
Epoch [1/3], Step [2878/3236], Loss: 2.0821, Perplexity: 8.0211
Epoch [1/3], Step [2879/3236], Loss: 2.2334, Perplexity: 9.3319
Epoch [1/3], Step [2880/3236], Loss: 2.1065, Perplexity: 8.2193
Epoch [1/3], Step [2881/3236], Loss: 2.6514, Perplexity: 14.1739
Epoch [1/3], Step [2882/3236], Loss: 2.1712, Perplexity: 8.7687
Epoch [1/3], Step [2883/3236], Loss: 2.1539, Perplexity: 8.6183
Epoch [1/3], Step [2884/3236], Loss: 2.1964, Perplexity: 8.9922
Epoch [1/3], Step [2885/3236], Loss: 2.7414, Perplexity: 15.5083
Epoch [1/3], Step [2886/3236], Loss: 2.2732, Perplexity: 9.7104
Epoch [1/3], Step [2887/3236], Loss: 2.3449, Perplexity: 10.4320
Epoch [1/3], Step [2888/3236], Loss: 2.3301, Perplexity: 10.2791
Epoch [1/3], Step [2889/3236], Loss: 2.5581, Perplexity: 12.9108
Epoch [1/3], Step [2890/3236], Loss: 2.3275, Perplexity: 10.2520
Epoch [1/3], Step [2891/3236], Loss: 2.1302, Perplexity: 8.4169
Epoch [1/3], Step [2892/3236], Loss: 2.1749, Perplexity: 8.8015
Epoch [1/3], Step [2893/3236], Loss: 2.2878, Perplexity: 9.8533
Epoch [1/3], Step [2894/3236], Loss: 2.4102, Perplexity: 11.1360
Epoch [1/3], Step [2895/3236], Loss: 2.1659, Perplexity: 8.7226
Epoch [1/3], Step [2896/3236], Loss: 2.0791, Perplexity: 7.9970
Epoch [1/3], Step [2897/3236], Loss: 2.0854, Perplexity: 8.0480
Epoch [1/3], Step [2898/3236], Loss: 2.1942, Perplexity: 8.9731
Epoch [1/3], Step [2899/3236], Loss: 2.0038, Perplexity: 7.4169
Epoch [1/3], Step [2900/3236], Loss: 2.3013, Perplexity: 9.9873
Epoch [1/3], Step [2901/3236], Loss: 2.1980, Perplexity: 9.0068
Epoch [1/3], Step [2902/3236], Loss: 2.2326, Perplexity: 9.3237
Epoch [1/3], Step [2903/3236], Loss: 2.1740, Perplexity: 8.7934
Epoch [1/3], Step [2904/3236], Loss: 2.2424, Perplexity: 9.4156
Epoch [1/3], Step [2905/3236], Loss: 2.2769, Perplexity: 9.7464
Epoch [1/3], Step [2906/3236], Loss: 2.1114, Perplexity: 8.2596
Epoch [1/3], Step [2907/3236], Loss: 2.2006, Perplexity: 9.0303
Epoch [1/3], Step [2908/3236], Loss: 2.1002, Perplexity: 8.1679
Epoch [1/3], Step [2909/3236], Loss: 2.1635, Perplexity: 8.7017
Epoch [1/3], Step [2910/3236], Loss: 2.2033, Perplexity: 9.0549
Epoch [1/3], Step [2911/3236], Loss: 2.2464, Perplexity: 9.4533
Epoch [1/3], Step [2912/3236], Loss: 2.3978, Perplexity: 10.9990
Epoch [1/3], Step [2913/3236], Loss: 2.2539, Perplexity: 9.5251
Epoch [1/3], Step [2914/3236], Loss: 2.2826, Perplexity: 9.8017
Epoch [1/3], Step [2915/3236], Loss: 2.1480, Perplexity: 8.5678
Epoch [1/3], Step [2916/3236], Loss: 2.2011, Perplexity: 9.0345
Epoch [1/3], Step [2917/3236], Loss: 2.2306, Perplexity: 9.3055
Epoch [1/3], Step [2918/3236], Loss: 2.1763, Perplexity: 8.8139
Epoch [1/3], Step [2919/3236], Loss: 2.8689, Perplexity: 17.6174
Epoch [1/3], Step [2920/3236], Loss: 2.1009, Perplexity: 8.1732
Epoch [1/3], Step [2921/3236], Loss: 2.2182, Perplexity: 9.1904
Epoch [1/3], Step [2922/3236], Loss: 2.2656, Perplexity: 9.6367
Epoch [1/3], Step [2923/3236], Loss: 2.1592, Perplexity: 8.6639
Epoch [1/3], Step [2924/3236], Loss: 2.2805, Perplexity: 9.7811
Epoch [1/3], Step [2925/3236], Loss: 2.3428, Perplexity: 10.4103
Epoch [1/3], Step [2926/3236], Loss: 2.4893, Perplexity: 12.0527
Epoch [1/3], Step [2927/3236], Loss: 2.1487, Perplexity: 8.5735
Epoch [1/3], Step [2928/3236], Loss: 2.1162, Perplexity: 8.2992
Epoch [1/3], Step [2929/3236], Loss: 2.3721, Perplexity: 10.7194
Epoch [1/3], Step [2930/3236], Loss: 2.0243, Perplexity: 7.5710
Epoch [1/3], Step [2931/3236], Loss: 2.5230, Perplexity: 12.4657
Epoch [1/3], Step [2932/3236], Loss: 2.1311, Perplexity: 8.4239
Epoch [1/3], Step [2933/3236], Loss: 2.1474, Perplexity: 8.5626
Epoch [1/3], Step [2934/3236], Loss: 2.2381, Perplexity: 9.3758
Epoch [1/3], Step [2935/3236], Loss: 2.7520, Perplexity: 15.6737
Epoch [1/3], Step [2936/3236], Loss: 2.1886, Perplexity: 8.9231
Epoch [1/3], Step [2937/3236], Loss: 2.1854, Perplexity: 8.8941
Epoch [1/3], Step [2938/3236], Loss: 2.1748, Perplexity: 8.8004
Epoch [1/3], Step [2939/3236], Loss: 2.1977, Perplexity: 9.0043
Epoch [1/3], Step [2940/3236], Loss: 2.1183, Perplexity: 8.3169
Epoch [1/3], Step [2941/3236], Loss: 2.4660, Perplexity: 11.7753
Epoch [1/3], Step [2942/3236], Loss: 2.2781, Perplexity: 9.7578
Epoch [1/3], Step [2943/3236], Loss: 3.1522, Perplexity: 23.3883
Epoch [1/3], Step [2944/3236], Loss: 2.2404, Perplexity: 9.3967
Epoch [1/3], Step [2945/3236], Loss: 2.1632, Perplexity: 8.6993
Epoch [1/3], Step [2946/3236], Loss: 2.1087, Perplexity: 8.2377
Epoch [1/3], Step [2947/3236], Loss: 2.4807, Perplexity: 11.9502
Epoch [1/3], Step [2948/3236], Loss: 2.3303, Perplexity: 10.2805
Epoch [1/3], Step [2949/3236], Loss: 2.4301, Perplexity: 11.3600
Epoch [1/3], Step [2950/3236], Loss: 2.0927, Perplexity: 8.1071
Epoch [1/3], Step [2951/3236], Loss: 2.2209, Perplexity: 9.2156
Epoch [1/3], Step [2952/3236], Loss: 2.1291, Perplexity: 8.4073
Epoch [1/3], Step [2953/3236], Loss: 2.1588, Perplexity: 8.6605
Epoch [1/3], Step [2954/3236], Loss: 2.0847, Perplexity: 8.0420
Epoch [1/3], Step [2955/3236], Loss: 2.3261, Perplexity: 10.2377
Epoch [1/3], Step [2956/3236], Loss: 2.2330, Perplexity: 9.3278
Epoch [1/3], Step [2957/3236], Loss: 2.2006, Perplexity: 9.0301
Epoch [1/3], Step [2958/3236], Loss: 2.2903, Perplexity: 9.8779
Epoch [1/3], Step [2959/3236], Loss: 2.1378, Perplexity: 8.4812
Epoch [1/3], Step [2960/3236], Loss: 2.4220, Perplexity: 11.2688
Epoch [1/3], Step [2961/3236], Loss: 2.7337, Perplexity: 15.3905
Epoch [1/3], Step [2962/3236], Loss: 2.2277, Perplexity: 9.2783
Epoch [1/3], Step [2963/3236], Loss: 2.4966, Perplexity: 12.1407
Epoch [1/3], Step [2964/3236], Loss: 2.1098, Perplexity: 8.2468
Epoch [1/3], Step [2965/3236], Loss: 2.2971, Perplexity: 9.9456
Epoch [1/3], Step [2966/3236], Loss: 2.1901, Perplexity: 8.9357
Epoch [1/3], Step [2967/3236], Loss: 2.0437, Perplexity: 7.7193
Epoch [1/3], Step [2968/3236], Loss: 2.2935, Perplexity: 9.9100
Epoch [1/3], Step [2969/3236], Loss: 2.1468, Perplexity: 8.5573
Epoch [1/3], Step [2970/3236], Loss: 2.4657, Perplexity: 11.7714
Epoch [1/3], Step [2971/3236], Loss: 2.2089, Perplexity: 9.1061
Epoch [1/3], Step [2972/3236], Loss: 2.4965, Perplexity: 12.1398
Epoch [1/3], Step [2973/3236], Loss: 2.3139, Perplexity: 10.1138
Epoch [1/3], Step [2974/3236], Loss: 2.0772, Perplexity: 7.9818
Epoch [1/3], Step [2975/3236], Loss: 2.5460, Perplexity: 12.7557
Epoch [1/3], Step [2976/3236], Loss: 2.2865, Perplexity: 9.8404
Epoch [1/3], Step [2977/3236], Loss: 2.8041, Perplexity: 16.5121
Epoch [1/3], Step [2978/3236], Loss: 2.9236, Perplexity: 18.6073
Epoch [1/3], Step [2979/3236], Loss: 2.2918, Perplexity: 9.8928
Epoch [1/3], Step [2980/3236], Loss: 2.1445, Perplexity: 8.5380
Epoch [1/3], Step [2981/3236], Loss: 2.2728, Perplexity: 9.7069
Epoch [1/3], Step [2982/3236], Loss: 2.1227, Perplexity: 8.3534
Epoch [1/3], Step [2983/3236], Loss: 2.0336, Perplexity: 7.6418
Epoch [1/3], Step [2984/3236], Loss: 2.1364, Perplexity: 8.4692
Epoch [1/3], Step [2985/3236], Loss: 2.1542, Perplexity: 8.6206
Epoch [1/3], Step [2986/3236], Loss: 2.1029, Perplexity: 8.1895
Epoch [1/3], Step [2987/3236], Loss: 2.0891, Perplexity: 8.0780
Epoch [1/3], Step [2988/3236], Loss: 4.4849, Perplexity: 88.6676
Epoch [1/3], Step [2989/3236], Loss: 2.1318, Perplexity: 8.4304
Epoch [1/3], Step [2990/3236], Loss: 2.2761, Perplexity: 9.7390
Epoch [1/3], Step [2991/3236], Loss: 2.4864, Perplexity: 12.0176
Epoch [1/3], Step [2992/3236], Loss: 2.1873, Perplexity: 8.9115
Epoch [1/3], Step [2993/3236], Loss: 2.2293, Perplexity: 9.2935
Epoch [1/3], Step [2994/3236], Loss: 2.1074, Perplexity: 8.2272
Epoch [1/3], Step [2995/3236], Loss: 2.0130, Perplexity: 7.4857
Epoch [1/3], Step [2996/3236], Loss: 2.2623, Perplexity: 9.6056
Epoch [1/3], Step [2997/3236], Loss: 2.4426, Perplexity: 11.5033
Epoch [1/3], Step [2998/3236], Loss: 2.0160, Perplexity: 7.5084
Epoch [1/3], Step [2999/3236], Loss: 2.3914, Perplexity: 10.9284
Epoch [1/3], Step [3000/3236], Loss: 2.1413, Perplexity: 8.5104
Epoch [1/3], Step [3001/3236], Loss: 2.1241, Perplexity: 8.3653
Epoch [1/3], Step [3002/3236], Loss: 2.3944, Perplexity: 10.9612
Epoch [1/3], Step [3003/3236], Loss: 2.3527, Perplexity: 10.5142
Epoch [1/3], Step [3004/3236], Loss: 2.1707, Perplexity: 8.7648
Epoch [1/3], Step [3005/3236], Loss: 2.4950, Perplexity: 12.1217
Epoch [1/3], Step [3006/3236], Loss: 2.5326, Perplexity: 12.5861
Epoch [1/3], Step [3007/3236], Loss: 2.2440, Perplexity: 9.4314
Epoch [1/3], Step [3008/3236], Loss: 2.2469, Perplexity: 9.4586
Epoch [1/3], Step [3009/3236], Loss: 2.3107, Perplexity: 10.0811
Epoch [1/3], Step [3010/3236], Loss: 2.3422, Perplexity: 10.4041
Epoch [1/3], Step [3011/3236], Loss: 2.2216, Perplexity: 9.2220
Epoch [1/3], Step [3012/3236], Loss: 2.4835, Perplexity: 11.9836
Epoch [1/3], Step [3013/3236], Loss: 2.2752, Perplexity: 9.7295
Epoch [1/3], Step [3014/3236], Loss: 2.7149, Perplexity: 15.1030
Epoch [1/3], Step [3015/3236], Loss: 2.1035, Perplexity: 8.1951
Epoch [1/3], Step [3016/3236], Loss: 2.1262, Perplexity: 8.3828
Epoch [1/3], Step [3017/3236], Loss: 2.5369, Perplexity: 12.6408
Epoch [1/3], Step [3018/3236], Loss: 2.2742, Perplexity: 9.7198
Epoch [1/3], Step [3019/3236], Loss: 2.2658, Perplexity: 9.6388
Epoch [1/3], Step [3020/3236], Loss: 2.2075, Perplexity: 9.0932
Epoch [1/3], Step [3021/3236], Loss: 2.3030, Perplexity: 10.0046
Epoch [1/3], Step [3022/3236], Loss: 2.1957, Perplexity: 8.9865
Epoch [1/3], Step [3023/3236], Loss: 2.3366, Perplexity: 10.3465
Epoch [1/3], Step [3024/3236], Loss: 2.2311, Perplexity: 9.3102
Epoch [1/3], Step [3025/3236], Loss: 2.8314, Perplexity: 16.9688
Epoch [1/3], Step [3026/3236], Loss: 2.1297, Perplexity: 8.4122
Epoch [1/3], Step [3027/3236], Loss: 2.4571, Perplexity: 11.6712
Epoch [1/3], Step [3028/3236], Loss: 2.3187, Perplexity: 10.1626
Epoch [1/3], Step [3029/3236], Loss: 2.3125, Perplexity: 10.0994
Epoch [1/3], Step [3030/3236], Loss: 2.2749, Perplexity: 9.7266
Epoch [1/3], Step [3031/3236], Loss: 2.1646, Perplexity: 8.7114
Epoch [1/3], Step [3032/3236], Loss: 2.6261, Perplexity: 13.8194
Epoch [1/3], Step [3033/3236], Loss: 2.0553, Perplexity: 7.8092
Epoch [1/3], Step [3034/3236], Loss: 2.2747, Perplexity: 9.7247
Epoch [1/3], Step [3035/3236], Loss: 2.3987, Perplexity: 11.0083
Epoch [1/3], Step [3036/3236], Loss: 2.2016, Perplexity: 9.0396
Epoch [1/3], Step [3037/3236], Loss: 2.1929, Perplexity: 8.9615
Epoch [1/3], Step [3038/3236], Loss: 2.0585, Perplexity: 7.8342
Epoch [1/3], Step [3039/3236], Loss: 2.0896, Perplexity: 8.0818
Epoch [1/3], Step [3040/3236], Loss: 2.3558, Perplexity: 10.5467
Epoch [1/3], Step [3041/3236], Loss: 2.2439, Perplexity: 9.4302
Epoch [1/3], Step [3042/3236], Loss: 2.3483, Perplexity: 10.4674
Epoch [1/3], Step [3043/3236], Loss: 2.1828, Perplexity: 8.8714
Epoch [1/3], Step [3044/3236], Loss: 2.2427, Perplexity: 9.4184
Epoch [1/3], Step [3045/3236], Loss: 2.1298, Perplexity: 8.4129
Epoch [1/3], Step [3046/3236], Loss: 2.2280, Perplexity: 9.2812
Epoch [1/3], Step [3047/3236], Loss: 2.2543, Perplexity: 9.5282
Epoch [1/3], Step [3048/3236], Loss: 2.6565, Perplexity: 14.2464
Epoch [1/3], Step [3049/3236], Loss: 2.2827, Perplexity: 9.8030
Epoch [1/3], Step [3050/3236], Loss: 2.3491, Perplexity: 10.4765
Epoch [1/3], Step [3051/3236], Loss: 2.2814, Perplexity: 9.7905
Epoch [1/3], Step [3052/3236], Loss: 2.0795, Perplexity: 8.0002
Epoch [1/3], Step [3053/3236], Loss: 2.4886, Perplexity: 12.0439
Epoch [1/3], Step [3054/3236], Loss: 2.5573, Perplexity: 12.9016
Epoch [1/3], Step [3055/3236], Loss: 2.8940, Perplexity: 18.0654
Epoch [1/3], Step [3056/3236], Loss: 2.1767, Perplexity: 8.8174
Epoch [1/3], Step [3057/3236], Loss: 2.3793, Perplexity: 10.7970
Epoch [1/3], Step [3058/3236], Loss: 2.1661, Perplexity: 8.7246
Epoch [1/3], Step [3059/3236], Loss: 2.1608, Perplexity: 8.6777
Epoch [1/3], Step [3060/3236], Loss: 2.1071, Perplexity: 8.2241
Epoch [1/3], Step [3061/3236], Loss: 2.0772, Perplexity: 7.9817
Epoch [1/3], Step [3062/3236], Loss: 2.0177, Perplexity: 7.5208
Epoch [1/3], Step [3063/3236], Loss: 2.1284, Perplexity: 8.4014
Epoch [1/3], Step [3064/3236], Loss: 2.1428, Perplexity: 8.5235
Epoch [1/3], Step [3065/3236], Loss: 2.5282, Perplexity: 12.5309
Epoch [1/3], Step [3066/3236], Loss: 2.2302, Perplexity: 9.3014
Epoch [1/3], Step [3067/3236], Loss: 2.1590, Perplexity: 8.6628
Epoch [1/3], Step [3068/3236], Loss: 2.2927, Perplexity: 9.9014
Epoch [1/3], Step [3069/3236], Loss: 2.1921, Perplexity: 8.9537
Epoch [1/3], Step [3070/3236], Loss: 2.0614, Perplexity: 7.8567
Epoch [1/3], Step [3071/3236], Loss: 2.1111, Perplexity: 8.2573
Epoch [1/3], Step [3072/3236], Loss: 2.2399, Perplexity: 9.3928
Epoch [1/3], Step [3073/3236], Loss: 2.5129, Perplexity: 12.3410
Epoch [1/3], Step [3074/3236], Loss: 2.5826, Perplexity: 13.2318
Epoch [1/3], Step [3075/3236], Loss: 2.3232, Perplexity: 10.2083
Epoch [1/3], Step [3076/3236], Loss: 2.1625, Perplexity: 8.6931
Epoch [1/3], Step [3077/3236], Loss: 2.2888, Perplexity: 9.8634
Epoch [1/3], Step [3078/3236], Loss: 2.1713, Perplexity: 8.7693
Epoch [1/3], Step [3079/3236], Loss: 2.4015, Perplexity: 11.0393
Epoch [1/3], Step [3080/3236], Loss: 2.3380, Perplexity: 10.3610
Epoch [1/3], Step [3081/3236], Loss: 2.6029, Perplexity: 13.5026
Epoch [1/3], Step [3082/3236], Loss: 2.1974, Perplexity: 9.0013
Epoch [1/3], Step [3083/3236], Loss: 2.2889, Perplexity: 9.8637
Epoch [1/3], Step [3084/3236], Loss: 2.1528, Perplexity: 8.6090
Epoch [1/3], Step [3085/3236], Loss: 2.1831, Perplexity: 8.8735
Epoch [1/3], Step [3086/3236], Loss: 2.2929, Perplexity: 9.9035
Epoch [1/3], Step [3087/3236], Loss: 2.6714, Perplexity: 14.4600
Epoch [1/3], Step [3088/3236], Loss: 2.1022, Perplexity: 8.1839
Epoch [1/3], Step [3089/3236], Loss: 2.2895, Perplexity: 9.8701
Epoch [1/3], Step [3090/3236], Loss: 2.6292, Perplexity: 13.8630
Epoch [1/3], Step [3091/3236], Loss: 2.4001, Perplexity: 11.0246
Epoch [1/3], Step [3092/3236], Loss: 2.1419, Perplexity: 8.5157
Epoch [1/3], Step [3093/3236], Loss: 2.1356, Perplexity: 8.4622
Epoch [1/3], Step [3094/3236], Loss: 2.2124, Perplexity: 9.1374
Epoch [1/3], Step [3095/3236], Loss: 2.1183, Perplexity: 8.3166
Epoch [1/3], Step [3096/3236], Loss: 1.9845, Perplexity: 7.2751
Epoch [1/3], Step [3097/3236], Loss: 2.1436, Perplexity: 8.5299
Epoch [1/3], Step [3098/3236], Loss: 2.2095, Perplexity: 9.1115
Epoch [1/3], Step [3099/3236], Loss: 4.5154, Perplexity: 91.4176
Epoch [1/3], Step [3100/3236], Loss: 2.5740, Perplexity: 13.1179
Epoch [1/3], Step [3101/3236], Loss: 2.1858, Perplexity: 8.8977
Epoch [1/3], Step [3102/3236], Loss: 2.4984, Perplexity: 12.1627
Epoch [1/3], Step [3103/3236], Loss: 2.3267, Perplexity: 10.2443
Epoch [1/3], Step [3104/3236], Loss: 2.1923, Perplexity: 8.9557
Epoch [1/3], Step [3105/3236], Loss: 2.1140, Perplexity: 8.2812
Epoch [1/3], Step [3106/3236], Loss: 2.1441, Perplexity: 8.5344
Epoch [1/3], Step [3107/3236], Loss: 2.2904, Perplexity: 9.8785
Epoch [1/3], Step [3108/3236], Loss: 2.1870, Perplexity: 8.9084
Epoch [1/3], Step [3109/3236], Loss: 2.5966, Perplexity: 13.4176
Epoch [1/3], Step [3110/3236], Loss: 2.2283, Perplexity: 9.2840
Epoch [1/3], Step [3111/3236], Loss: 2.2269, Perplexity: 9.2712
Epoch [1/3], Step [3112/3236], Loss: 2.2496, Perplexity: 9.4838
Epoch [1/3], Step [3113/3236], Loss: 2.2773, Perplexity: 9.7504
Epoch [1/3], Step [3114/3236], Loss: 2.1334, Perplexity: 8.4437
Epoch [1/3], Step [3115/3236], Loss: 1.9850, Perplexity: 7.2787
Epoch [1/3], Step [3116/3236], Loss: 2.2138, Perplexity: 9.1508
Epoch [1/3], Step [3117/3236], Loss: 2.1866, Perplexity: 8.9047
Epoch [1/3], Step [3118/3236], Loss: 2.5651, Perplexity: 13.0016
Epoch [1/3], Step [3119/3236], Loss: 2.7432, Perplexity: 15.5360
Epoch [1/3], Step [3120/3236], Loss: 2.2085, Perplexity: 9.1017
Epoch [1/3], Step [3121/3236], Loss: 1.9160, Perplexity: 6.7935
Epoch [1/3], Step [3122/3236], Loss: 2.1369, Perplexity: 8.4730
Epoch [1/3], Step [3123/3236], Loss: 2.1789, Perplexity: 8.8366
Epoch [1/3], Step [3124/3236], Loss: 2.0816, Perplexity: 8.0176
Epoch [1/3], Step [3125/3236], Loss: 2.1696, Perplexity: 8.7546
Epoch [1/3], Step [3126/3236], Loss: 2.5221, Perplexity: 12.4552
Epoch [1/3], Step [3127/3236], Loss: 2.3586, Perplexity: 10.5764
Epoch [1/3], Step [3128/3236], Loss: 2.2403, Perplexity: 9.3965
Epoch [1/3], Step [3129/3236], Loss: 2.0737, Perplexity: 7.9542
Epoch [1/3], Step [3130/3236], Loss: 2.6770, Perplexity: 14.5420
Epoch [1/3], Step [3131/3236], Loss: 2.2402, Perplexity: 9.3953
Epoch [1/3], Step [3132/3236], Loss: 2.4833, Perplexity: 11.9813
Epoch [1/3], Step [3133/3236], Loss: 2.5595, Perplexity: 12.9298
Epoch [1/3], Step [3134/3236], Loss: 2.5607, Perplexity: 12.9446
Epoch [1/3], Step [3135/3236], Loss: 2.2737, Perplexity: 9.7149
Epoch [1/3], Step [3136/3236], Loss: 2.1432, Perplexity: 8.5263
Epoch [1/3], Step [3137/3236], Loss: 2.1595, Perplexity: 8.6670
Epoch [1/3], Step [3138/3236], Loss: 2.2411, Perplexity: 9.4033
Epoch [1/3], Step [3139/3236], Loss: 2.2196, Perplexity: 9.2037
Epoch [1/3], Step [3140/3236], Loss: 2.3843, Perplexity: 10.8518
Epoch [1/3], Step [3141/3236], Loss: 2.1286, Perplexity: 8.4027
Epoch [1/3], Step [3142/3236], Loss: 2.2948, Perplexity: 9.9224
Epoch [1/3], Step [3143/3236], Loss: 2.1610, Perplexity: 8.6799
Epoch [1/3], Step [3144/3236], Loss: 2.1385, Perplexity: 8.4869
Epoch [1/3], Step [3145/3236], Loss: 2.1969, Perplexity: 8.9968
Epoch [1/3], Step [3146/3236], Loss: 2.3953, Perplexity: 10.9715
Epoch [1/3], Step [3147/3236], Loss: 2.2289, Perplexity: 9.2895
Epoch [1/3], Step [3148/3236], Loss: 2.1937, Perplexity: 8.9687
Epoch [1/3], Step [3149/3236], Loss: 2.2266, Perplexity: 9.2687
Epoch [1/3], Step [3150/3236], Loss: 2.1532, Perplexity: 8.6127
Epoch [1/3], Step [3151/3236], Loss: 2.0904, Perplexity: 8.0879
Epoch [1/3], Step [3152/3236], Loss: 2.2535, Perplexity: 9.5212
Epoch [1/3], Step [3153/3236], Loss: 2.1364, Perplexity: 8.4687
Epoch [1/3], Step [3154/3236], Loss: 2.0390, Perplexity: 7.6826
Epoch [1/3], Step [3155/3236], Loss: 2.2295, Perplexity: 9.2948
Epoch [1/3], Step [3156/3236], Loss: 2.1486, Perplexity: 8.5730
Epoch [1/3], Step [3157/3236], Loss: 2.3202, Perplexity: 10.1778
Epoch [1/3], Step [3158/3236], Loss: 2.1716, Perplexity: 8.7722
Epoch [1/3], Step [3159/3236], Loss: 2.0977, Perplexity: 8.1472
Epoch [1/3], Step [3160/3236], Loss: 2.2732, Perplexity: 9.7107
Epoch [1/3], Step [3161/3236], Loss: 2.2505, Perplexity: 9.4927
Epoch [1/3], Step [3162/3236], Loss: 2.1831, Perplexity: 8.8734
Epoch [1/3], Step [3163/3236], Loss: 2.1586, Perplexity: 8.6592
Epoch [1/3], Step [3164/3236], Loss: 2.2943, Perplexity: 9.9178
Epoch [1/3], Step [3165/3236], Loss: 2.1569, Perplexity: 8.6446
Epoch [1/3], Step [3166/3236], Loss: 2.1815, Perplexity: 8.8594
Epoch [1/3], Step [3167/3236], Loss: 2.3286, Perplexity: 10.2640
Epoch [1/3], Step [3168/3236], Loss: 2.1099, Perplexity: 8.2475
Epoch [1/3], Step [3169/3236], Loss: 2.1686, Perplexity: 8.7458
Epoch [1/3], Step [3170/3236], Loss: 2.1248, Perplexity: 8.3713
Epoch [1/3], Step [3171/3236], Loss: 2.1914, Perplexity: 8.9478
Epoch [1/3], Step [3172/3236], Loss: 2.0920, Perplexity: 8.1011
Epoch [1/3], Step [3173/3236], Loss: 2.3620, Perplexity: 10.6126
Epoch [1/3], Step [3174/3236], Loss: 2.0850, Perplexity: 8.0447
Epoch [1/3], Step [3175/3236], Loss: 2.2333, Perplexity: 9.3306
Epoch [1/3], Step [3176/3236], Loss: 2.1842, Perplexity: 8.8837
Epoch [1/3], Step [3177/3236], Loss: 2.2167, Perplexity: 9.1770
Epoch [1/3], Step [3178/3236], Loss: 2.2566, Perplexity: 9.5506
Epoch [1/3], Step [3179/3236], Loss: 2.2514, Perplexity: 9.5013
Epoch [1/3], Step [3180/3236], Loss: 2.1191, Perplexity: 8.3240
Epoch [1/3], Step [3181/3236], Loss: 2.3573, Perplexity: 10.5620
Epoch [1/3], Step [3182/3236], Loss: 2.0444, Perplexity: 7.7241
Epoch [1/3], Step [3183/3236], Loss: 2.1214, Perplexity: 8.3431
Epoch [1/3], Step [3184/3236], Loss: 2.1118, Perplexity: 8.2629
Epoch [1/3], Step [3185/3236], Loss: 2.4354, Perplexity: 11.4207
Epoch [1/3], Step [3186/3236], Loss: 2.1996, Perplexity: 9.0215
Epoch [1/3], Step [3187/3236], Loss: 2.2155, Perplexity: 9.1661
Epoch [1/3], Step [3188/3236], Loss: 2.4135, Perplexity: 11.1733
Epoch [1/3], Step [3189/3236], Loss: 2.1837, Perplexity: 8.8788
Epoch [1/3], Step [3190/3236], Loss: 2.2493, Perplexity: 9.4810
Epoch [1/3], Step [3191/3236], Loss: 2.5368, Perplexity: 12.6394
Epoch [1/3], Step [3192/3236], Loss: 2.0803, Perplexity: 8.0067
Epoch [1/3], Step [3193/3236], Loss: 2.4560, Perplexity: 11.6575
Epoch [1/3], Step [3194/3236], Loss: 2.1634, Perplexity: 8.7010
Epoch [1/3], Step [3195/3236], Loss: 2.1834, Perplexity: 8.8762
Epoch [1/3], Step [3196/3236], Loss: 2.5313, Perplexity: 12.5696
Epoch [1/3], Step [3197/3236], Loss: 2.1837, Perplexity: 8.8789
Epoch [1/3], Step [3198/3236], Loss: 2.3192, Perplexity: 10.1675
Epoch [1/3], Step [3199/3236], Loss: 2.1323, Perplexity: 8.4344
Epoch [1/3], Step [3200/3236], Loss: 2.5715, Perplexity: 13.0860
Epoch [1/3], Step [3201/3236], Loss: 2.1436, Perplexity: 8.5300
Epoch [1/3], Step [3202/3236], Loss: 2.2196, Perplexity: 9.2034
Epoch [1/3], Step [3203/3236], Loss: 2.2249, Perplexity: 9.2527
Epoch [1/3], Step [3204/3236], Loss: 2.3100, Perplexity: 10.0744
Epoch [1/3], Step [3205/3236], Loss: 2.2662, Perplexity: 9.6424
Epoch [1/3], Step [3206/3236], Loss: 2.1784, Perplexity: 8.8323
Epoch [1/3], Step [3207/3236], Loss: 2.1818, Perplexity: 8.8623
Epoch [1/3], Step [3208/3236], Loss: 2.0846, Perplexity: 8.0411
Epoch [1/3], Step [3209/3236], Loss: 2.2620, Perplexity: 9.6020
Epoch [1/3], Step [3210/3236], Loss: 2.0938, Perplexity: 8.1155
Epoch [1/3], Step [3211/3236], Loss: 2.1409, Perplexity: 8.5068
Epoch [1/3], Step [3212/3236], Loss: 2.3262, Perplexity: 10.2394
Epoch [1/3], Step [3213/3236], Loss: 2.1969, Perplexity: 8.9968
Epoch [1/3], Step [3214/3236], Loss: 2.1995, Perplexity: 9.0201
Epoch [1/3], Step [3215/3236], Loss: 2.3766, Perplexity: 10.7683
Epoch [1/3], Step [3216/3236], Loss: 2.2372, Perplexity: 9.3668
Epoch [1/3], Step [3217/3236], Loss: 2.3318, Perplexity: 10.2964
Epoch [1/3], Step [3218/3236], Loss: 2.0116, Perplexity: 7.4750
Epoch [1/3], Step [3219/3236], Loss: 2.2958, Perplexity: 9.9319
Epoch [1/3], Step [3220/3236], Loss: 2.2428, Perplexity: 9.4194
Epoch [1/3], Step [3221/3236], Loss: 2.1744, Perplexity: 8.7969
Epoch [1/3], Step [3222/3236], Loss: 2.2045, Perplexity: 9.0659
Epoch [1/3], Step [3223/3236], Loss: 2.4722, Perplexity: 11.8490
Epoch [1/3], Step [3224/3236], Loss: 2.1925, Perplexity: 8.9579
Epoch [1/3], Step [3225/3236], Loss: 2.2395, Perplexity: 9.3891
Epoch [1/3], Step [3226/3236], Loss: 1.9623, Perplexity: 7.1156
Epoch [1/3], Step [3227/3236], Loss: 2.3994, Perplexity: 11.0170
Epoch [1/3], Step [3228/3236], Loss: 2.1144, Perplexity: 8.2850
Epoch [1/3], Step [3229/3236], Loss: 2.3642, Perplexity: 10.6359
Epoch [1/3], Step [3230/3236], Loss: 2.0303, Perplexity: 7.6165
Epoch [1/3], Step [3231/3236], Loss: 2.2420, Perplexity: 9.4125
Epoch [1/3], Step [3232/3236], Loss: 2.2498, Perplexity: 9.4855
Epoch [1/3], Step [3233/3236], Loss: 2.0935, Perplexity: 8.1134
Epoch [1/3], Step [3234/3236], Loss: 2.2719, Perplexity: 9.6974
Epoch [1/3], Step [3235/3236], Loss: 2.1217, Perplexity: 8.3457
Epoch [1/3], Step [3236/3236], Loss: 2.2740, Perplexity: 9.7181
Epoch [2/3], Step [1/3236], Loss: 2.1486, Perplexity: 8.5733
Epoch [2/3], Step [2/3236], Loss: 2.2902, Perplexity: 9.8773
Epoch [2/3], Step [3/3236], Loss: 2.6248, Perplexity: 13.8016
Epoch [2/3], Step [4/3236], Loss: 2.2473, Perplexity: 9.4620
Epoch [2/3], Step [5/3236], Loss: 2.1189, Perplexity: 8.3217
Epoch [2/3], Step [6/3236], Loss: 2.0953, Perplexity: 8.1281
Epoch [2/3], Step [7/3236], Loss: 2.0713, Perplexity: 7.9354
Epoch [2/3], Step [8/3236], Loss: 2.4386, Perplexity: 11.4569
Epoch [2/3], Step [9/3236], Loss: 2.1298, Perplexity: 8.4134
Epoch [2/3], Step [10/3236], Loss: 2.1637, Perplexity: 8.7033
Epoch [2/3], Step [11/3236], Loss: 2.3518, Perplexity: 10.5045
Epoch [2/3], Step [12/3236], Loss: 2.1362, Perplexity: 8.4676
Epoch [2/3], Step [13/3236], Loss: 2.2742, Perplexity: 9.7198
Epoch [2/3], Step [14/3236], Loss: 2.7852, Perplexity: 16.2032
Epoch [2/3], Step [15/3236], Loss: 2.1742, Perplexity: 8.7948
Epoch [2/3], Step [16/3236], Loss: 2.3286, Perplexity: 10.2635
Epoch [2/3], Step [17/3236], Loss: 2.9096, Perplexity: 18.3490
Epoch [2/3], Step [18/3236], Loss: 2.1591, Perplexity: 8.6636
Epoch [2/3], Step [19/3236], Loss: 2.2356, Perplexity: 9.3523
Epoch [2/3], Step [20/3236], Loss: 2.3365, Perplexity: 10.3453
Epoch [2/3], Step [21/3236], Loss: 2.0930, Perplexity: 8.1095
Epoch [2/3], Step [22/3236], Loss: 2.1484, Perplexity: 8.5708
Epoch [2/3], Step [23/3236], Loss: 2.1247, Perplexity: 8.3704
Epoch [2/3], Step [24/3236], Loss: 2.2643, Perplexity: 9.6242
Epoch [2/3], Step [25/3236], Loss: 2.1138, Perplexity: 8.2800
Epoch [2/3], Step [26/3236], Loss: 2.0798, Perplexity: 8.0028
Epoch [2/3], Step [27/3236], Loss: 2.1757, Perplexity: 8.8081
Epoch [2/3], Step [28/3236], Loss: 2.1338, Perplexity: 8.4468
Epoch [2/3], Step [29/3236], Loss: 2.3452, Perplexity: 10.4358
Epoch [2/3], Step [30/3236], Loss: 2.3834, Perplexity: 10.8414
Epoch [2/3], Step [31/3236], Loss: 2.1249, Perplexity: 8.3721
Epoch [2/3], Step [32/3236], Loss: 2.0425, Perplexity: 7.7095
Epoch [2/3], Step [33/3236], Loss: 2.0840, Perplexity: 8.0369
Epoch [2/3], Step [34/3236], Loss: 2.1718, Perplexity: 8.7743
Epoch [2/3], Step [35/3236], Loss: 2.1812, Perplexity: 8.8570
Epoch [2/3], Step [36/3236], Loss: 2.1978, Perplexity: 9.0049
Epoch [2/3], Step [37/3236], Loss: 2.1354, Perplexity: 8.4608
Epoch [2/3], Step [38/3236], Loss: 2.1473, Perplexity: 8.5616
Epoch [2/3], Step [39/3236], Loss: 2.1615, Perplexity: 8.6840
Epoch [2/3], Step [40/3236], Loss: 2.1000, Perplexity: 8.1663
Epoch [2/3], Step [41/3236], Loss: 2.3526, Perplexity: 10.5131
Epoch [2/3], Step [42/3236], Loss: 3.0137, Perplexity: 20.3617
Epoch [2/3], Step [43/3236], Loss: 2.3064, Perplexity: 10.0378
Epoch [2/3], Step [44/3236], Loss: 2.2009, Perplexity: 9.0331
Epoch [2/3], Step [45/3236], Loss: 2.2235, Perplexity: 9.2396
Epoch [2/3], Step [46/3236], Loss: 2.2007, Perplexity: 9.0316
Epoch [2/3], Step [47/3236], Loss: 2.0496, Perplexity: 7.7646
Epoch [2/3], Step [48/3236], Loss: 2.1087, Perplexity: 8.2373
Epoch [2/3], Step [49/3236], Loss: 2.2565, Perplexity: 9.5500
Epoch [2/3], Step [50/3236], Loss: 2.2593, Perplexity: 9.5763
Epoch [2/3], Step [51/3236], Loss: 2.2476, Perplexity: 9.4651
Epoch [2/3], Step [52/3236], Loss: 2.3363, Perplexity: 10.3429
Epoch [2/3], Step [53/3236], Loss: 2.2429, Perplexity: 9.4203
Epoch [2/3], Step [54/3236], Loss: 2.0831, Perplexity: 8.0294
Epoch [2/3], Step [55/3236], Loss: 2.1567, Perplexity: 8.6428
Epoch [2/3], Step [56/3236], Loss: 2.1767, Perplexity: 8.8173
Epoch [2/3], Step [57/3236], Loss: 2.1927, Perplexity: 8.9591
Epoch [2/3], Step [58/3236], Loss: 2.2057, Perplexity: 9.0765
Epoch [2/3], Step [59/3236], Loss: 2.1597, Perplexity: 8.6685
Epoch [2/3], Step [60/3236], Loss: 2.3036, Perplexity: 10.0102
Epoch [2/3], Step [61/3236], Loss: 2.1255, Perplexity: 8.3775
Epoch [2/3], Step [62/3236], Loss: 3.7526, Perplexity: 42.6326
Epoch [2/3], Step [63/3236], Loss: 2.7062, Perplexity: 14.9717
Epoch [2/3], Step [64/3236], Loss: 2.2122, Perplexity: 9.1358
Epoch [2/3], Step [65/3236], Loss: 2.0754, Perplexity: 7.9681
Epoch [2/3], Step [66/3236], Loss: 2.0698, Perplexity: 7.9234
Epoch [2/3], Step [67/3236], Loss: 2.2348, Perplexity: 9.3445
Epoch [2/3], Step [68/3236], Loss: 2.0806, Perplexity: 8.0095
Epoch [2/3], Step [69/3236], Loss: 2.3985, Perplexity: 11.0062
Epoch [2/3], Step [70/3236], Loss: 2.1887, Perplexity: 8.9238
Epoch [2/3], Step [71/3236], Loss: 2.0839, Perplexity: 8.0358
Epoch [2/3], Step [72/3236], Loss: 2.0966, Perplexity: 8.1385
Epoch [2/3], Step [73/3236], Loss: 2.0666, Perplexity: 7.8983
Epoch [2/3], Step [74/3236], Loss: 2.1593, Perplexity: 8.6647
Epoch [2/3], Step [75/3236], Loss: 2.1513, Perplexity: 8.5963
Epoch [2/3], Step [76/3236], Loss: 3.3192, Perplexity: 27.6369
Epoch [2/3], Step [77/3236], Loss: 2.1290, Perplexity: 8.4063
Epoch [2/3], Step [78/3236], Loss: 2.2392, Perplexity: 9.3863
Epoch [2/3], Step [79/3236], Loss: 2.4210, Perplexity: 11.2571
Epoch [2/3], Step [80/3236], Loss: 2.0481, Perplexity: 7.7533
Epoch [2/3], Step [81/3236], Loss: 2.1380, Perplexity: 8.4826
Epoch [2/3], Step [82/3236], Loss: 2.1378, Perplexity: 8.4804
Epoch [2/3], Step [83/3236], Loss: 2.1985, Perplexity: 9.0113
Epoch [2/3], Step [84/3236], Loss: 2.0544, Perplexity: 7.8020
Epoch [2/3], Step [85/3236], Loss: 2.2964, Perplexity: 9.9383
Epoch [2/3], Step [86/3236], Loss: 2.2594, Perplexity: 9.5773
Epoch [2/3], Step [87/3236], Loss: 2.1850, Perplexity: 8.8910
Epoch [2/3], Step [88/3236], Loss: 2.4594, Perplexity: 11.6983
Epoch [2/3], Step [89/3236], Loss: 2.1732, Perplexity: 8.7864
Epoch [2/3], Step [90/3236], Loss: 3.4103, Perplexity: 30.2738
Epoch [2/3], Step [91/3236], Loss: 2.2353, Perplexity: 9.3494
Epoch [2/3], Step [92/3236], Loss: 2.1175, Perplexity: 8.3103
Epoch [2/3], Step [93/3236], Loss: 2.1009, Perplexity: 8.1738
Epoch [2/3], Step [94/3236], Loss: 2.2413, Perplexity: 9.4060
Epoch [2/3], Step [95/3236], Loss: 2.5552, Perplexity: 12.8739
Epoch [2/3], Step [96/3236], Loss: 2.4332, Perplexity: 11.3958
Epoch [2/3], Step [97/3236], Loss: 2.1406, Perplexity: 8.5043
Epoch [2/3], Step [98/3236], Loss: 2.2320, Perplexity: 9.3183
Epoch [2/3], Step [99/3236], Loss: 2.0741, Perplexity: 7.9575
Epoch [2/3], Step [100/3236], Loss: 2.1158, Perplexity: 8.2964
Epoch [2/3], Step [101/3236], Loss: 2.2345, Perplexity: 9.3422
Epoch [2/3], Step [102/3236], Loss: 2.4366, Perplexity: 11.4344
Epoch [2/3], Step [103/3236], Loss: 2.2413, Perplexity: 9.4059
Epoch [2/3], Step [104/3236], Loss: 2.5105, Perplexity: 12.3106
Epoch [2/3], Step [105/3236], Loss: 2.1758, Perplexity: 8.8091
Epoch [2/3], Step [106/3236], Loss: 2.1958, Perplexity: 8.9876
Epoch [2/3], Step [107/3236], Loss: 2.1497, Perplexity: 8.5826
Epoch [2/3], Step [108/3236], Loss: 2.0190, Perplexity: 7.5311
Epoch [2/3], Step [109/3236], Loss: 2.1801, Perplexity: 8.8468
Epoch [2/3], Step [110/3236], Loss: 2.1368, Perplexity: 8.4719
Epoch [2/3], Step [111/3236], Loss: 2.2100, Perplexity: 9.1161
Epoch [2/3], Step [112/3236], Loss: 2.1954, Perplexity: 8.9838
Epoch [2/3], Step [113/3236], Loss: 2.4202, Perplexity: 11.2478
Epoch [2/3], Step [114/3236], Loss: 2.4419, Perplexity: 11.4950
Epoch [2/3], Step [115/3236], Loss: 2.1295, Perplexity: 8.4108
Epoch [2/3], Step [116/3236], Loss: 2.4756, Perplexity: 11.8883
Epoch [2/3], Step [117/3236], Loss: 2.2266, Perplexity: 9.2683
Epoch [2/3], Step [118/3236], Loss: 2.9914, Perplexity: 19.9138
Epoch [2/3], Step [119/3236], Loss: 2.3479, Perplexity: 10.4637
Epoch [2/3], Step [120/3236], Loss: 2.1585, Perplexity: 8.6580
Epoch [2/3], Step [121/3236], Loss: 2.1264, Perplexity: 8.3845
Epoch [2/3], Step [122/3236], Loss: 2.1133, Perplexity: 8.2754
Epoch [2/3], Step [123/3236], Loss: 2.0386, Perplexity: 7.6795
Epoch [2/3], Step [124/3236], Loss: 2.2059, Perplexity: 9.0784
Epoch [2/3], Step [125/3236], Loss: 2.1851, Perplexity: 8.8920
Epoch [2/3], Step [126/3236], Loss: 2.1927, Perplexity: 8.9596
Epoch [2/3], Step [127/3236], Loss: 2.0457, Perplexity: 7.7347
Epoch [2/3], Step [128/3236], Loss: 2.2213, Perplexity: 9.2194
Epoch [2/3], Step [129/3236], Loss: 2.2307, Perplexity: 9.3065
Epoch [2/3], Step [130/3236], Loss: 2.0765, Perplexity: 7.9762
Epoch [2/3], Step [131/3236], Loss: 2.3719, Perplexity: 10.7179
Epoch [2/3], Step [132/3236], Loss: 2.1595, Perplexity: 8.6665
Epoch [2/3], Step [133/3236], Loss: 2.1397, Perplexity: 8.4970
Epoch [2/3], Step [134/3236], Loss: 2.1413, Perplexity: 8.5109
Epoch [2/3], Step [135/3236], Loss: 2.0951, Perplexity: 8.1263
Epoch [2/3], Step [136/3236], Loss: 2.2246, Perplexity: 9.2499
Epoch [2/3], Step [137/3236], Loss: 2.0814, Perplexity: 8.0157
Epoch [2/3], Step [138/3236], Loss: 2.0798, Perplexity: 8.0026
Epoch [2/3], Step [139/3236], Loss: 2.1370, Perplexity: 8.4739
Epoch [2/3], Step [140/3236], Loss: 2.1712, Perplexity: 8.7689
Epoch [2/3], Step [141/3236], Loss: 2.0443, Perplexity: 7.7235
Epoch [2/3], Step [142/3236], Loss: 2.0613, Perplexity: 7.8559
Epoch [2/3], Step [143/3236], Loss: 2.0623, Perplexity: 7.8644
Epoch [2/3], Step [144/3236], Loss: 2.1891, Perplexity: 8.9273
Epoch [2/3], Step [145/3236], Loss: 2.0925, Perplexity: 8.1055
Epoch [2/3], Step [146/3236], Loss: 2.3093, Perplexity: 10.0675
Epoch [2/3], Step [147/3236], Loss: 2.2265, Perplexity: 9.2678
Epoch [2/3], Step [148/3236], Loss: 2.2933, Perplexity: 9.9075
Epoch [2/3], Step [149/3236], Loss: 2.1629, Perplexity: 8.6964
Epoch [2/3], Step [150/3236], Loss: 2.1391, Perplexity: 8.4919
Epoch [2/3], Step [151/3236], Loss: 2.2281, Perplexity: 9.2825
Epoch [2/3], Step [152/3236], Loss: 2.3283, Perplexity: 10.2606
Epoch [2/3], Step [153/3236], Loss: 2.5776, Perplexity: 13.1654
Epoch [2/3], Step [154/3236], Loss: 2.2978, Perplexity: 9.9526
Epoch [2/3], Step [155/3236], Loss: 2.1356, Perplexity: 8.4622
Epoch [2/3], Step [156/3236], Loss: 2.1442, Perplexity: 8.5354
Epoch [2/3], Step [157/3236], Loss: 2.3369, Perplexity: 10.3491
Epoch [2/3], Step [158/3236], Loss: 2.6530, Perplexity: 14.1971
Epoch [2/3], Step [159/3236], Loss: 2.2896, Perplexity: 9.8713
Epoch [2/3], Step [160/3236], Loss: 2.8058, Perplexity: 16.5410
Epoch [2/3], Step [161/3236], Loss: 2.1559, Perplexity: 8.6354
Epoch [2/3], Step [162/3236], Loss: 2.5204, Perplexity: 12.4331
Epoch [2/3], Step [163/3236], Loss: 1.9982, Perplexity: 7.3760
Epoch [2/3], Step [164/3236], Loss: 2.3337, Perplexity: 10.3157
Epoch [2/3], Step [165/3236], Loss: 2.1940, Perplexity: 8.9711
Epoch [2/3], Step [166/3236], Loss: 2.2608, Perplexity: 9.5910
Epoch [2/3], Step [167/3236], Loss: 2.0277, Perplexity: 7.5968
Epoch [2/3], Step [168/3236], Loss: 2.5283, Perplexity: 12.5320
Epoch [2/3], Step [169/3236], Loss: 2.1079, Perplexity: 8.2306
Epoch [2/3], Step [170/3236], Loss: 2.2468, Perplexity: 9.4578
Epoch [2/3], Step [171/3236], Loss: 2.2616, Perplexity: 9.5986
Epoch [2/3], Step [172/3236], Loss: 2.8903, Perplexity: 17.9996
Epoch [2/3], Step [173/3236], Loss: 2.0049, Perplexity: 7.4250
Epoch [2/3], Step [174/3236], Loss: 2.2989, Perplexity: 9.9632
Epoch [2/3], Step [175/3236], Loss: 2.1045, Perplexity: 8.2026
Epoch [2/3], Step [176/3236], Loss: 2.3111, Perplexity: 10.0855
Epoch [2/3], Step [177/3236], Loss: 2.1987, Perplexity: 9.0135
Epoch [2/3], Step [178/3236], Loss: 2.3056, Perplexity: 10.0300
Epoch [2/3], Step [179/3236], Loss: 2.3126, Perplexity: 10.1002
Epoch [2/3], Step [180/3236], Loss: 2.7910, Perplexity: 16.2970
Epoch [2/3], Step [181/3236], Loss: 2.0878, Perplexity: 8.0668
Epoch [2/3], Step [182/3236], Loss: 2.1162, Perplexity: 8.2999
Epoch [2/3], Step [183/3236], Loss: 2.0665, Perplexity: 7.8973
Epoch [2/3], Step [184/3236], Loss: 2.2842, Perplexity: 9.8180
Epoch [2/3], Step [185/3236], Loss: 2.0790, Perplexity: 7.9966
Epoch [2/3], Step [186/3236], Loss: 2.1361, Perplexity: 8.4662
Epoch [2/3], Step [187/3236], Loss: 1.9970, Perplexity: 7.3670
Epoch [2/3], Step [188/3236], Loss: 2.0629, Perplexity: 7.8687
Epoch [2/3], Step [189/3236], Loss: 2.2138, Perplexity: 9.1505
Epoch [2/3], Step [190/3236], Loss: 2.0619, Perplexity: 7.8610
Epoch [2/3], Step [191/3236], Loss: 2.1188, Perplexity: 8.3214
Epoch [2/3], Step [192/3236], Loss: 2.2144, Perplexity: 9.1556
Epoch [2/3], Step [193/3236], Loss: 2.1225, Perplexity: 8.3519
Epoch [2/3], Step [194/3236], Loss: 2.3058, Perplexity: 10.0319
Epoch [2/3], Step [195/3236], Loss: 2.2548, Perplexity: 9.5330
Epoch [2/3], Step [196/3236], Loss: 1.9432, Perplexity: 6.9811
Epoch [2/3], Step [197/3236], Loss: 2.3132, Perplexity: 10.1066
Epoch [2/3], Step [198/3236], Loss: 2.5728, Perplexity: 13.1031
Epoch [2/3], Step [199/3236], Loss: 2.8066, Perplexity: 16.5544
Epoch [2/3], Step [200/3236], Loss: 2.1205, Perplexity: 8.3351
Epoch [2/3], Step [201/3236], Loss: 2.0995, Perplexity: 8.1620
Epoch [2/3], Step [202/3236], Loss: 2.4749, Perplexity: 11.8802
Epoch [2/3], Step [203/3236], Loss: 2.2402, Perplexity: 9.3956
Epoch [2/3], Step [204/3236], Loss: 2.0957, Perplexity: 8.1309
Epoch [2/3], Step [205/3236], Loss: 2.0834, Perplexity: 8.0318
Epoch [2/3], Step [206/3236], Loss: 2.0248, Perplexity: 7.5746
Epoch [2/3], Step [207/3236], Loss: 2.2663, Perplexity: 9.6432
Epoch [2/3], Step [208/3236], Loss: 2.2098, Perplexity: 9.1143
Epoch [2/3], Step [209/3236], Loss: 2.0459, Perplexity: 7.7359
Epoch [2/3], Step [210/3236], Loss: 2.1889, Perplexity: 8.9255
Epoch [2/3], Step [211/3236], Loss: 2.2054, Perplexity: 9.0735
Epoch [2/3], Step [212/3236], Loss: 1.9993, Perplexity: 7.3837
Epoch [2/3], Step [213/3236], Loss: 2.0867, Perplexity: 8.0584
Epoch [2/3], Step [214/3236], Loss: 2.5570, Perplexity: 12.8976
Epoch [2/3], Step [215/3236], Loss: 2.1635, Perplexity: 8.7013
Epoch [2/3], Step [216/3236], Loss: 2.0027, Perplexity: 7.4092
Epoch [2/3], Step [217/3236], Loss: 2.1909, Perplexity: 8.9430
Epoch [2/3], Step [218/3236], Loss: 2.4948, Perplexity: 12.1189
Epoch [2/3], Step [219/3236], Loss: 2.0970, Perplexity: 8.1417
Epoch [2/3], Step [220/3236], Loss: 2.1177, Perplexity: 8.3123
Epoch [2/3], Step [221/3236], Loss: 2.0142, Perplexity: 7.4950
Epoch [2/3], Step [222/3236], Loss: 2.1481, Perplexity: 8.5686
Epoch [2/3], Step [223/3236], Loss: 2.1537, Perplexity: 8.6163
Epoch [2/3], Step [224/3236], Loss: 2.2689, Perplexity: 9.6691
Epoch [2/3], Step [225/3236], Loss: 2.0913, Perplexity: 8.0958
Epoch [2/3], Step [226/3236], Loss: 2.2645, Perplexity: 9.6263
Epoch [2/3], Step [227/3236], Loss: 2.2566, Perplexity: 9.5509
Epoch [2/3], Step [228/3236], Loss: 2.3459, Perplexity: 10.4431
Epoch [2/3], Step [229/3236], Loss: 2.1642, Perplexity: 8.7079
Epoch [2/3], Step [230/3236], Loss: 2.5239, Perplexity: 12.4776
Epoch [2/3], Step [231/3236], Loss: 2.1365, Perplexity: 8.4700
Epoch [2/3], Step [232/3236], Loss: 2.0134, Perplexity: 7.4884
Epoch [2/3], Step [233/3236], Loss: 2.1911, Perplexity: 8.9454
Epoch [2/3], Step [234/3236], Loss: 2.0684, Perplexity: 7.9121
Epoch [2/3], Step [235/3236], Loss: 2.8130, Perplexity: 16.6591
Epoch [2/3], Step [236/3236], Loss: 2.2073, Perplexity: 9.0907
Epoch [2/3], Step [237/3236], Loss: 2.0249, Perplexity: 7.5754
Epoch [2/3], Step [238/3236], Loss: 2.0654, Perplexity: 7.8886
Epoch [2/3], Step [239/3236], Loss: 2.0359, Perplexity: 7.6589
Epoch [2/3], Step [240/3236], Loss: 2.1286, Perplexity: 8.4034
Epoch [2/3], Step [241/3236], Loss: 2.2130, Perplexity: 9.1435
Epoch [2/3], Step [242/3236], Loss: 2.5872, Perplexity: 13.2920
Epoch [2/3], Step [243/3236], Loss: 2.3212, Perplexity: 10.1881
Epoch [2/3], Step [244/3236], Loss: 2.1923, Perplexity: 8.9554
Epoch [2/3], Step [245/3236], Loss: 1.9823, Perplexity: 7.2591
Epoch [2/3], Step [246/3236], Loss: 2.1358, Perplexity: 8.4641
Epoch [2/3], Step [247/3236], Loss: 2.9419, Perplexity: 18.9521
Epoch [2/3], Step [248/3236], Loss: 2.2884, Perplexity: 9.8587
Epoch [2/3], Step [249/3236], Loss: 2.2423, Perplexity: 9.4151
Epoch [2/3], Step [250/3236], Loss: 2.1907, Perplexity: 8.9417
Epoch [2/3], Step [251/3236], Loss: 2.5259, Perplexity: 12.5023
Epoch [2/3], Step [252/3236], Loss: 2.4115, Perplexity: 11.1512
Epoch [2/3], Step [253/3236], Loss: 2.1441, Perplexity: 8.5342
Epoch [2/3], Step [254/3236], Loss: 2.1836, Perplexity: 8.8779
Epoch [2/3], Step [255/3236], Loss: 2.1822, Perplexity: 8.8657
Epoch [2/3], Step [256/3236], Loss: 2.0825, Perplexity: 8.0244
Epoch [2/3], Step [257/3236], Loss: 2.1412, Perplexity: 8.5098
Epoch [2/3], Step [258/3236], Loss: 2.1499, Perplexity: 8.5840
Epoch [2/3], Step [259/3236], Loss: 2.4210, Perplexity: 11.2573
Epoch [2/3], Step [260/3236], Loss: 2.1023, Perplexity: 8.1847
Epoch [2/3], Step [261/3236], Loss: 2.1585, Perplexity: 8.6585
Epoch [2/3], Step [262/3236], Loss: 2.1372, Perplexity: 8.4759
Epoch [2/3], Step [263/3236], Loss: 2.1200, Perplexity: 8.3315
Epoch [2/3], Step [264/3236], Loss: 2.2459, Perplexity: 9.4493
Epoch [2/3], Step [265/3236], Loss: 2.1253, Perplexity: 8.3756
Epoch [2/3], Step [266/3236], Loss: 2.1921, Perplexity: 8.9539
Epoch [2/3], Step [267/3236], Loss: 2.1416, Perplexity: 8.5132
Epoch [2/3], Step [268/3236], Loss: 2.2768, Perplexity: 9.7450
Epoch [2/3], Step [269/3236], Loss: 2.0025, Perplexity: 7.4078
Epoch [2/3], Step [270/3236], Loss: 3.0097, Perplexity: 20.2823
Epoch [2/3], Step [271/3236], Loss: 2.1489, Perplexity: 8.5757
Epoch [2/3], Step [272/3236], Loss: 2.4788, Perplexity: 11.9268
Epoch [2/3], Step [273/3236], Loss: 2.0131, Perplexity: 7.4864
Epoch [2/3], Step [274/3236], Loss: 1.8601, Perplexity: 6.4243
Epoch [2/3], Step [275/3236], Loss: 2.1199, Perplexity: 8.3300
Epoch [2/3], Step [276/3236], Loss: 2.2353, Perplexity: 9.3492
Epoch [2/3], Step [277/3236], Loss: 2.3001, Perplexity: 9.9755
Epoch [2/3], Step [278/3236], Loss: 2.1826, Perplexity: 8.8695
Epoch [2/3], Step [279/3236], Loss: 2.0210, Perplexity: 7.5457
Epoch [2/3], Step [280/3236], Loss: 2.1736, Perplexity: 8.7899
Epoch [2/3], Step [281/3236], Loss: 2.2398, Perplexity: 9.3914
Epoch [2/3], Step [282/3236], Loss: 2.4189, Perplexity: 11.2330
Epoch [2/3], Step [283/3236], Loss: 2.3149, Perplexity: 10.1240
Epoch [2/3], Step [284/3236], Loss: 2.2814, Perplexity: 9.7906
Epoch [2/3], Step [285/3236], Loss: 2.2145, Perplexity: 9.1565
Epoch [2/3], Step [286/3236], Loss: 2.1173, Perplexity: 8.3086
Epoch [2/3], Step [287/3236], Loss: 2.1447, Perplexity: 8.5393
Epoch [2/3], Step [288/3236], Loss: 2.1290, Perplexity: 8.4062
Epoch [2/3], Step [289/3236], Loss: 2.0387, Perplexity: 7.6810
Epoch [2/3], Step [290/3236], Loss: 2.0407, Perplexity: 7.6963
Epoch [2/3], Step [291/3236], Loss: 2.7105, Perplexity: 15.0369
Epoch [2/3], Step [292/3236], Loss: 2.1570, Perplexity: 8.6448
Epoch [2/3], Step [293/3236], Loss: 2.0525, Perplexity: 7.7874
Epoch [2/3], Step [294/3236], Loss: 2.3566, Perplexity: 10.5549
Epoch [2/3], Step [295/3236], Loss: 2.5338, Perplexity: 12.6017
Epoch [2/3], Step [296/3236], Loss: 2.2291, Perplexity: 9.2916
Epoch [2/3], Step [297/3236], Loss: 2.2682, Perplexity: 9.6623
Epoch [2/3], Step [298/3236], Loss: 2.2062, Perplexity: 9.0811
Epoch [2/3], Step [299/3236], Loss: 2.1773, Perplexity: 8.8224
Epoch [2/3], Step [300/3236], Loss: 2.4216, Perplexity: 11.2634
Epoch [2/3], Step [301/3236], Loss: 2.1973, Perplexity: 9.0003
Epoch [2/3], Step [302/3236], Loss: 1.9988, Perplexity: 7.3805
Epoch [2/3], Step [303/3236], Loss: 2.1666, Perplexity: 8.7285
Epoch [2/3], Step [304/3236], Loss: 2.1790, Perplexity: 8.8376
Epoch [2/3], Step [305/3236], Loss: 2.1969, Perplexity: 8.9969
Epoch [2/3], Step [306/3236], Loss: 2.1202, Perplexity: 8.3330
Epoch [2/3], Step [307/3236], Loss: 2.1671, Perplexity: 8.7332
Epoch [2/3], Step [308/3236], Loss: 2.0646, Perplexity: 7.8819
Epoch [2/3], Step [309/3236], Loss: 1.9789, Perplexity: 7.2345
Epoch [2/3], Step [310/3236], Loss: 2.4059, Perplexity: 11.0886
Epoch [2/3], Step [311/3236], Loss: 2.4259, Perplexity: 11.3125
Epoch [2/3], Step [312/3236], Loss: 2.1367, Perplexity: 8.4718
Epoch [2/3], Step [313/3236], Loss: 2.4552, Perplexity: 11.6487
Epoch [2/3], Step [314/3236], Loss: 2.4969, Perplexity: 12.1444
Epoch [2/3], Step [315/3236], Loss: 2.3002, Perplexity: 9.9761
Epoch [2/3], Step [316/3236], Loss: 2.2901, Perplexity: 9.8755
Epoch [2/3], Step [317/3236], Loss: 2.1979, Perplexity: 9.0060
Epoch [2/3], Step [318/3236], Loss: 2.3029, Perplexity: 10.0031
Epoch [2/3], Step [319/3236], Loss: 2.2221, Perplexity: 9.2263
Epoch [2/3], Step [320/3236], Loss: 2.2897, Perplexity: 9.8716
Epoch [2/3], Step [321/3236], Loss: 1.9864, Perplexity: 7.2890
Epoch [2/3], Step [322/3236], Loss: 2.0555, Perplexity: 7.8104
Epoch [2/3], Step [323/3236], Loss: 2.2230, Perplexity: 9.2351
Epoch [2/3], Step [324/3236], Loss: 2.0829, Perplexity: 8.0280
Epoch [2/3], Step [325/3236], Loss: 2.2284, Perplexity: 9.2851
Epoch [2/3], Step [326/3236], Loss: 2.2292, Perplexity: 9.2920
Epoch [2/3], Step [327/3236], Loss: 2.2319, Perplexity: 9.3174
Epoch [2/3], Step [328/3236], Loss: 2.3864, Perplexity: 10.8746
Epoch [2/3], Step [329/3236], Loss: 2.1824, Perplexity: 8.8676
Epoch [2/3], Step [330/3236], Loss: 2.0365, Perplexity: 7.6639
Epoch [2/3], Step [331/3236], Loss: 2.1902, Perplexity: 8.9370
Epoch [2/3], Step [332/3236], Loss: 2.5770, Perplexity: 13.1571
Epoch [2/3], Step [333/3236], Loss: 2.0374, Perplexity: 7.6707
Epoch [2/3], Step [334/3236], Loss: 2.7232, Perplexity: 15.2285
Epoch [2/3], Step [335/3236], Loss: 1.9239, Perplexity: 6.8473
Epoch [2/3], Step [336/3236], Loss: 2.1608, Perplexity: 8.6785
Epoch [2/3], Step [337/3236], Loss: 2.3119, Perplexity: 10.0932
Epoch [2/3], Step [338/3236], Loss: 2.2426, Perplexity: 9.4173
Epoch [2/3], Step [339/3236], Loss: 1.9900, Perplexity: 7.3158
Epoch [2/3], Step [340/3236], Loss: 2.4179, Perplexity: 11.2223
Epoch [2/3], Step [341/3236], Loss: 2.3442, Perplexity: 10.4253
Epoch [2/3], Step [342/3236], Loss: 2.1725, Perplexity: 8.7802
Epoch [2/3], Step [343/3236], Loss: 2.1565, Perplexity: 8.6406
Epoch [2/3], Step [344/3236], Loss: 2.2132, Perplexity: 9.1448
Epoch [2/3], Step [345/3236], Loss: 2.5067, Perplexity: 12.2645
Epoch [2/3], Step [346/3236], Loss: 2.1233, Perplexity: 8.3589
Epoch [2/3], Step [347/3236], Loss: 2.1523, Perplexity: 8.6047
Epoch [2/3], Step [348/3236], Loss: 2.2029, Perplexity: 9.0511
Epoch [2/3], Step [349/3236], Loss: 2.1637, Perplexity: 8.7036
Epoch [2/3], Step [350/3236], Loss: 2.1628, Perplexity: 8.6955
Epoch [2/3], Step [351/3236], Loss: 2.1531, Perplexity: 8.6112
Epoch [2/3], Step [352/3236], Loss: 2.0566, Perplexity: 7.8190
Epoch [2/3], Step [353/3236], Loss: 2.5633, Perplexity: 12.9788
Epoch [2/3], Step [354/3236], Loss: 2.1596, Perplexity: 8.6673
Epoch [2/3], Step [355/3236], Loss: 2.1786, Perplexity: 8.8335
Epoch [2/3], Step [356/3236], Loss: 2.1283, Perplexity: 8.4010
Epoch [2/3], Step [357/3236], Loss: 2.0498, Perplexity: 7.7664
Epoch [2/3], Step [358/3236], Loss: 2.0994, Perplexity: 8.1611
Epoch [2/3], Step [359/3236], Loss: 2.0923, Perplexity: 8.1039
Epoch [2/3], Step [360/3236], Loss: 2.2422, Perplexity: 9.4140
Epoch [2/3], Step [361/3236], Loss: 2.1792, Perplexity: 8.8395
Epoch [2/3], Step [362/3236], Loss: 2.3384, Perplexity: 10.3646
Epoch [2/3], Step [363/3236], Loss: 2.2196, Perplexity: 9.2041
Epoch [2/3], Step [364/3236], Loss: 2.3328, Perplexity: 10.3071
Epoch [2/3], Step [365/3236], Loss: 2.0176, Perplexity: 7.5200
Epoch [2/3], Step [366/3236], Loss: 2.0386, Perplexity: 7.6802
Epoch [2/3], Step [367/3236], Loss: 2.0834, Perplexity: 8.0320
Epoch [2/3], Step [368/3236], Loss: 2.0858, Perplexity: 8.0514
Epoch [2/3], Step [369/3236], Loss: 2.2183, Perplexity: 9.1920
Epoch [2/3], Step [370/3236], Loss: 2.1013, Perplexity: 8.1771
Epoch [2/3], Step [371/3236], Loss: 2.1695, Perplexity: 8.7539
Epoch [2/3], Step [372/3236], Loss: 2.2172, Perplexity: 9.1812
Epoch [2/3], Step [373/3236], Loss: 2.4541, Perplexity: 11.6358
Epoch [2/3], Step [374/3236], Loss: 2.5071, Perplexity: 12.2688
Epoch [2/3], Step [375/3236], Loss: 2.1354, Perplexity: 8.4603
Epoch [2/3], Step [376/3236], Loss: 2.0869, Perplexity: 8.0603
Epoch [2/3], Step [377/3236], Loss: 2.0642, Perplexity: 7.8787
Epoch [2/3], Step [378/3236], Loss: 2.2273, Perplexity: 9.2745
Epoch [2/3], Step [379/3236], Loss: 2.2531, Perplexity: 9.5176
Epoch [2/3], Step [380/3236], Loss: 2.0121, Perplexity: 7.4793
Epoch [2/3], Step [381/3236], Loss: 2.0939, Perplexity: 8.1165
Epoch [2/3], Step [382/3236], Loss: 1.9860, Perplexity: 7.2861
Epoch [2/3], Step [383/3236], Loss: 2.1571, Perplexity: 8.6457
Epoch [2/3], Step [384/3236], Loss: 2.1669, Perplexity: 8.7312
Epoch [2/3], Step [385/3236], Loss: 2.1207, Perplexity: 8.3369
Epoch [2/3], Step [386/3236], Loss: 2.3847, Perplexity: 10.8556
Epoch [2/3], Step [387/3236], Loss: 2.0985, Perplexity: 8.1537
Epoch [2/3], Step [388/3236], Loss: 2.1191, Perplexity: 8.3236
Epoch [2/3], Step [389/3236], Loss: 2.0794, Perplexity: 7.9998
Epoch [2/3], Step [390/3236], Loss: 2.1374, Perplexity: 8.4776
Epoch [2/3], Step [391/3236], Loss: 2.2047, Perplexity: 9.0675
Epoch [2/3], Step [392/3236], Loss: 2.1703, Perplexity: 8.7606
Epoch [2/3], Step [393/3236], Loss: 2.1635, Perplexity: 8.7019
Epoch [2/3], Step [394/3236], Loss: 3.2586, Perplexity: 26.0142
Epoch [2/3], Step [395/3236], Loss: 2.1348, Perplexity: 8.4553
Epoch [2/3], Step [396/3236], Loss: 2.0873, Perplexity: 8.0629
Epoch [2/3], Step [397/3236], Loss: 1.9721, Perplexity: 7.1857
Epoch [2/3], Step [398/3236], Loss: 2.3990, Perplexity: 11.0123
Epoch [2/3], Step [399/3236], Loss: 2.0097, Perplexity: 7.4607
Epoch [2/3], Step [400/3236], Loss: 2.0606, Perplexity: 7.8505
Epoch [2/3], Step [401/3236], Loss: 2.1164, Perplexity: 8.3012
Epoch [2/3], Step [402/3236], Loss: 2.1915, Perplexity: 8.9490
Epoch [2/3], Step [403/3236], Loss: 2.0689, Perplexity: 7.9163
Epoch [2/3], Step [404/3236], Loss: 2.0975, Perplexity: 8.1461
Epoch [2/3], Step [405/3236], Loss: 2.1545, Perplexity: 8.6237
Epoch [2/3], Step [406/3236], Loss: 2.6683, Perplexity: 14.4150
Epoch [2/3], Step [407/3236], Loss: 2.1181, Perplexity: 8.3155
Epoch [2/3], Step [408/3236], Loss: 1.9898, Perplexity: 7.3144
Epoch [2/3], Step [409/3236], Loss: 2.1153, Perplexity: 8.2921
Epoch [2/3], Step [410/3236], Loss: 2.2463, Perplexity: 9.4525
Epoch [2/3], Step [411/3236], Loss: 2.0587, Perplexity: 7.8359
Epoch [2/3], Step [412/3236], Loss: 2.2539, Perplexity: 9.5245
Epoch [2/3], Step [413/3236], Loss: 2.3421, Perplexity: 10.4032
Epoch [2/3], Step [414/3236], Loss: 2.2688, Perplexity: 9.6682
Epoch [2/3], Step [415/3236], Loss: 2.0346, Perplexity: 7.6495
Epoch [2/3], Step [416/3236], Loss: 2.0942, Perplexity: 8.1190
Epoch [2/3], Step [417/3236], Loss: 2.1334, Perplexity: 8.4436
Epoch [2/3], Step [418/3236], Loss: 2.6672, Perplexity: 14.3997
Epoch [2/3], Step [419/3236], Loss: 2.5119, Perplexity: 12.3289
Epoch [2/3], Step [420/3236], Loss: 2.3856, Perplexity: 10.8655
Epoch [2/3], Step [421/3236], Loss: 2.0561, Perplexity: 7.8158
Epoch [2/3], Step [422/3236], Loss: 2.0699, Perplexity: 7.9244
Epoch [2/3], Step [423/3236], Loss: 2.3897, Perplexity: 10.9101
Epoch [2/3], Step [424/3236], Loss: 2.3089, Perplexity: 10.0632
Epoch [2/3], Step [425/3236], Loss: 2.0943, Perplexity: 8.1195
Epoch [2/3], Step [426/3236], Loss: 2.9504, Perplexity: 19.1140
Epoch [2/3], Step [427/3236], Loss: 2.1867, Perplexity: 8.9056
Epoch [2/3], Step [428/3236], Loss: 2.4313, Perplexity: 11.3738
Epoch [2/3], Step [429/3236], Loss: 2.2176, Perplexity: 9.1855
Epoch [2/3], Step [430/3236], Loss: 2.1892, Perplexity: 8.9284
Epoch [2/3], Step [431/3236], Loss: 2.0715, Perplexity: 7.9365
Epoch [2/3], Step [432/3236], Loss: 2.7796, Perplexity: 16.1126
Epoch [2/3], Step [433/3236], Loss: 2.0255, Perplexity: 7.5799
Epoch [2/3], Step [434/3236], Loss: 2.1130, Perplexity: 8.2733
Epoch [2/3], Step [435/3236], Loss: 2.6483, Perplexity: 14.1295
Epoch [2/3], Step [436/3236], Loss: 2.0131, Perplexity: 7.4862
Epoch [2/3], Step [437/3236], Loss: 2.1053, Perplexity: 8.2099
Epoch [2/3], Step [438/3236], Loss: 2.4144, Perplexity: 11.1833
Epoch [2/3], Step [439/3236], Loss: 2.2463, Perplexity: 9.4524
Epoch [2/3], Step [440/3236], Loss: 2.1500, Perplexity: 8.5845
Epoch [2/3], Step [441/3236], Loss: 2.1695, Perplexity: 8.7541
Epoch [2/3], Step [442/3236], Loss: 2.1181, Perplexity: 8.3153
Epoch [2/3], Step [443/3236], Loss: 2.1306, Perplexity: 8.4199
Epoch [2/3], Step [444/3236], Loss: 2.1827, Perplexity: 8.8701
Epoch [2/3], Step [445/3236], Loss: 2.1342, Perplexity: 8.4501
Epoch [2/3], Step [446/3236], Loss: 2.2052, Perplexity: 9.0725
Epoch [2/3], Step [447/3236], Loss: 2.2201, Perplexity: 9.2087
Epoch [2/3], Step [448/3236], Loss: 1.9560, Perplexity: 7.0711
Epoch [2/3], Step [449/3236], Loss: 2.4306, Perplexity: 11.3654
Epoch [2/3], Step [450/3236], Loss: 2.0201, Perplexity: 7.5393
Epoch [2/3], Step [451/3236], Loss: 2.0101, Perplexity: 7.4644
Epoch [2/3], Step [452/3236], Loss: 2.1009, Perplexity: 8.1733
Epoch [2/3], Step [453/3236], Loss: 2.0976, Perplexity: 8.1469
Epoch [2/3], Step [454/3236], Loss: 2.0059, Perplexity: 7.4331
Epoch [2/3], Step [455/3236], Loss: 2.0458, Perplexity: 7.7353
Epoch [2/3], Step [456/3236], Loss: 2.3878, Perplexity: 10.8890
Epoch [2/3], Step [457/3236], Loss: 2.0157, Perplexity: 7.5061
Epoch [2/3], Step [458/3236], Loss: 2.3063, Perplexity: 10.0376
Epoch [2/3], Step [459/3236], Loss: 2.0142, Perplexity: 7.4945
Epoch [2/3], Step [460/3236], Loss: 2.0323, Perplexity: 7.6319
Epoch [2/3], Step [461/3236], Loss: 2.1723, Perplexity: 8.7780
Epoch [2/3], Step [462/3236], Loss: 2.4241, Perplexity: 11.2921
Epoch [2/3], Step [463/3236], Loss: 2.1427, Perplexity: 8.5223
Epoch [2/3], Step [464/3236], Loss: 2.1227, Perplexity: 8.3539
Epoch [2/3], Step [465/3236], Loss: 2.2053, Perplexity: 9.0729
Epoch [2/3], Step [466/3236], Loss: 2.1948, Perplexity: 8.9784
Epoch [2/3], Step [467/3236], Loss: 2.0659, Perplexity: 7.8921
Epoch [2/3], Step [468/3236], Loss: 1.9946, Perplexity: 7.3495
Epoch [2/3], Step [469/3236], Loss: 2.0201, Perplexity: 7.5391
Epoch [2/3], Step [470/3236], Loss: 2.1116, Perplexity: 8.2611
Epoch [2/3], Step [471/3236], Loss: 2.5741, Perplexity: 13.1190
Epoch [2/3], Step [472/3236], Loss: 2.0862, Perplexity: 8.0541
Epoch [2/3], Step [473/3236], Loss: 2.2950, Perplexity: 9.9245
Epoch [2/3], Step [474/3236], Loss: 2.1072, Perplexity: 8.2251
Epoch [2/3], Step [475/3236], Loss: 2.1708, Perplexity: 8.7650
Epoch [2/3], Step [476/3236], Loss: 2.1286, Perplexity: 8.4031
Epoch [2/3], Step [477/3236], Loss: 2.0552, Perplexity: 7.8081
Epoch [2/3], Step [478/3236], Loss: 2.0253, Perplexity: 7.5784
Epoch [2/3], Step [479/3236], Loss: 2.0844, Perplexity: 8.0398
Epoch [2/3], Step [480/3236], Loss: 2.2200, Perplexity: 9.2076
Epoch [2/3], Step [481/3236], Loss: 2.1647, Perplexity: 8.7116
Epoch [2/3], Step [482/3236], Loss: 2.5895, Perplexity: 13.3228
Epoch [2/3], Step [483/3236], Loss: 2.0479, Perplexity: 7.7519
Epoch [2/3], Step [484/3236], Loss: 2.2178, Perplexity: 9.1871
Epoch [2/3], Step [485/3236], Loss: 2.0962, Perplexity: 8.1349
Epoch [2/3], Step [486/3236], Loss: 2.3156, Perplexity: 10.1308
Epoch [2/3], Step [487/3236], Loss: 2.2082, Perplexity: 9.0996
Epoch [2/3], Step [488/3236], Loss: 2.0262, Perplexity: 7.5849
Epoch [2/3], Step [489/3236], Loss: 2.4628, Perplexity: 11.7380
Epoch [2/3], Step [490/3236], Loss: 2.1008, Perplexity: 8.1728
Epoch [2/3], Step [491/3236], Loss: 2.0905, Perplexity: 8.0893
Epoch [2/3], Step [492/3236], Loss: 2.0201, Perplexity: 7.5392
Epoch [2/3], Step [493/3236], Loss: 2.1139, Perplexity: 8.2805
Epoch [2/3], Step [494/3236], Loss: 2.2061, Perplexity: 9.0801
Epoch [2/3], Step [495/3236], Loss: 2.2706, Perplexity: 9.6851
Epoch [2/3], Step [496/3236], Loss: 2.2218, Perplexity: 9.2238
Epoch [2/3], Step [497/3236], Loss: 2.9291, Perplexity: 18.7100
Epoch [2/3], Step [498/3236], Loss: 2.0176, Perplexity: 7.5201
Epoch [2/3], Step [499/3236], Loss: 2.2655, Perplexity: 9.6356
Epoch [2/3], Step [500/3236], Loss: 2.1181, Perplexity: 8.3151
Epoch [2/3], Step [501/3236], Loss: 2.2055, Perplexity: 9.0752
Epoch [2/3], Step [502/3236], Loss: 2.0613, Perplexity: 7.8565
Epoch [2/3], Step [503/3236], Loss: 2.3209, Perplexity: 10.1845
Epoch [2/3], Step [504/3236], Loss: 2.2089, Perplexity: 9.1053
Epoch [2/3], Step [505/3236], Loss: 2.1384, Perplexity: 8.4862
Epoch [2/3], Step [506/3236], Loss: 2.3063, Perplexity: 10.0375
Epoch [2/3], Step [507/3236], Loss: 2.0111, Perplexity: 7.4716
Epoch [2/3], Step [508/3236], Loss: 2.1805, Perplexity: 8.8503
Epoch [2/3], Step [509/3236], Loss: 2.0904, Perplexity: 8.0883
Epoch [2/3], Step [510/3236], Loss: 2.0263, Perplexity: 7.5858
Epoch [2/3], Step [511/3236], Loss: 2.5107, Perplexity: 12.3140
Epoch [2/3], Step [512/3236], Loss: 2.8569, Perplexity: 17.4071
Epoch [2/3], Step [513/3236], Loss: 2.0908, Perplexity: 8.0910
Epoch [2/3], Step [514/3236], Loss: 2.0910, Perplexity: 8.0929
Epoch [2/3], Step [515/3236], Loss: 2.9081, Perplexity: 18.3217
Epoch [2/3], Step [516/3236], Loss: 2.2841, Perplexity: 9.8169
Epoch [2/3], Step [517/3236], Loss: 2.1704, Perplexity: 8.7615
Epoch [2/3], Step [518/3236], Loss: 2.2528, Perplexity: 9.5140
Epoch [2/3], Step [519/3236], Loss: 2.1544, Perplexity: 8.6225
Epoch [2/3], Step [520/3236], Loss: 2.0240, Perplexity: 7.5687
Epoch [2/3], Step [521/3236], Loss: 2.4709, Perplexity: 11.8331
Epoch [2/3], Step [522/3236], Loss: 2.2735, Perplexity: 9.7129
Epoch [2/3], Step [523/3236], Loss: 2.2219, Perplexity: 9.2253
Epoch [2/3], Step [524/3236], Loss: 2.1534, Perplexity: 8.6144
Epoch [2/3], Step [525/3236], Loss: 2.1412, Perplexity: 8.5095
Epoch [2/3], Step [526/3236], Loss: 3.1334, Perplexity: 22.9521
Epoch [2/3], Step [527/3236], Loss: 2.1392, Perplexity: 8.4926
Epoch [2/3], Step [528/3236], Loss: 2.8408, Perplexity: 17.1291
Epoch [2/3], Step [529/3236], Loss: 2.1048, Perplexity: 8.2051
Epoch [2/3], Step [530/3236], Loss: 2.3353, Perplexity: 10.3322
Epoch [2/3], Step [531/3236], Loss: 2.1667, Perplexity: 8.7294
Epoch [2/3], Step [532/3236], Loss: 1.9811, Perplexity: 7.2505
Epoch [2/3], Step [533/3236], Loss: 2.0826, Perplexity: 8.0257
Epoch [2/3], Step [534/3236], Loss: 2.0760, Perplexity: 7.9729
Epoch [2/3], Step [535/3236], Loss: 2.0426, Perplexity: 7.7107
Epoch [2/3], Step [536/3236], Loss: 2.7317, Perplexity: 15.3595
Epoch [2/3], Step [537/3236], Loss: 2.1925, Perplexity: 8.9578
Epoch [2/3], Step [538/3236], Loss: 2.0147, Perplexity: 7.4983
Epoch [2/3], Step [539/3236], Loss: 2.3566, Perplexity: 10.5555
Epoch [2/3], Step [540/3236], Loss: 2.1152, Perplexity: 8.2910
Epoch [2/3], Step [541/3236], Loss: 2.4200, Perplexity: 11.2462
Epoch [2/3], Step [542/3236], Loss: 2.1834, Perplexity: 8.8765
Epoch [2/3], Step [543/3236], Loss: 2.3616, Perplexity: 10.6077
Epoch [2/3], Step [544/3236], Loss: 1.9915, Perplexity: 7.3264
Epoch [2/3], Step [545/3236], Loss: 2.4054, Perplexity: 11.0831
Epoch [2/3], Step [546/3236], Loss: 2.0967, Perplexity: 8.1391
Epoch [2/3], Step [547/3236], Loss: 2.1277, Perplexity: 8.3957
Epoch [2/3], Step [548/3236], Loss: 2.7146, Perplexity: 15.0980
Epoch [2/3], Step [549/3236], Loss: 2.3871, Perplexity: 10.8819
Epoch [2/3], Step [550/3236], Loss: 2.1877, Perplexity: 8.9144
Epoch [2/3], Step [551/3236], Loss: 2.3752, Perplexity: 10.7532
Epoch [2/3], Step [552/3236], Loss: 2.2130, Perplexity: 9.1433
Epoch [2/3], Step [553/3236], Loss: 2.2471, Perplexity: 9.4604
Epoch [2/3], Step [554/3236], Loss: 2.2073, Perplexity: 9.0911
Epoch [2/3], Step [555/3236], Loss: 2.1881, Perplexity: 8.9178
Epoch [2/3], Step [556/3236], Loss: 2.2111, Perplexity: 9.1255
Epoch [2/3], Step [557/3236], Loss: 2.2512, Perplexity: 9.4994
Epoch [2/3], Step [558/3236], Loss: 2.0191, Perplexity: 7.5313
Epoch [2/3], Step [559/3236], Loss: 2.1739, Perplexity: 8.7929
Epoch [2/3], Step [560/3236], Loss: 2.0678, Perplexity: 7.9074
Epoch [2/3], Step [561/3236], Loss: 2.1799, Perplexity: 8.8452
Epoch [2/3], Step [562/3236], Loss: 2.0563, Perplexity: 7.8166
Epoch [2/3], Step [563/3236], Loss: 2.1752, Perplexity: 8.8039
Epoch [2/3], Step [564/3236], Loss: 2.1207, Perplexity: 8.3370
Epoch [2/3], Step [565/3236], Loss: 2.4154, Perplexity: 11.1944
Epoch [2/3], Step [566/3236], Loss: 2.3856, Perplexity: 10.8653
Epoch [2/3], Step [567/3236], Loss: 2.0761, Perplexity: 7.9734
Epoch [2/3], Step [568/3236], Loss: 2.3074, Perplexity: 10.0480
Epoch [2/3], Step [569/3236], Loss: 1.8976, Perplexity: 6.6701
Epoch [2/3], Step [570/3236], Loss: 2.1335, Perplexity: 8.4440
Epoch [2/3], Step [571/3236], Loss: 2.1644, Perplexity: 8.7097
Epoch [2/3], Step [572/3236], Loss: 2.3766, Perplexity: 10.7683
Epoch [2/3], Step [573/3236], Loss: 2.1648, Perplexity: 8.7125
Epoch [2/3], Step [574/3236], Loss: 2.2165, Perplexity: 9.1748
Epoch [2/3], Step [575/3236], Loss: 2.6022, Perplexity: 13.4935
Epoch [2/3], Step [576/3236], Loss: 2.2755, Perplexity: 9.7330
Epoch [2/3], Step [577/3236], Loss: 2.0622, Perplexity: 7.8633
Epoch [2/3], Step [578/3236], Loss: 2.1353, Perplexity: 8.4596
Epoch [2/3], Step [579/3236], Loss: 2.1626, Perplexity: 8.6937
Epoch [2/3], Step [580/3236], Loss: 2.1382, Perplexity: 8.4844
Epoch [2/3], Step [581/3236], Loss: 2.2420, Perplexity: 9.4119
Epoch [2/3], Step [582/3236], Loss: 2.1974, Perplexity: 9.0018
Epoch [2/3], Step [583/3236], Loss: 2.2319, Perplexity: 9.3180
Epoch [2/3], Step [584/3236], Loss: 1.9479, Perplexity: 7.0141
Epoch [2/3], Step [585/3236], Loss: 1.9421, Perplexity: 6.9736
Epoch [2/3], Step [586/3236], Loss: 2.1156, Perplexity: 8.2947
Epoch [2/3], Step [587/3236], Loss: 2.3775, Perplexity: 10.7781
Epoch [2/3], Step [588/3236], Loss: 2.4064, Perplexity: 11.0942
Epoch [2/3], Step [589/3236], Loss: 2.2566, Perplexity: 9.5505
Epoch [2/3], Step [590/3236], Loss: 2.2578, Perplexity: 9.5623
Epoch [2/3], Step [591/3236], Loss: 2.8027, Perplexity: 16.4899
Epoch [2/3], Step [592/3236], Loss: 2.8017, Perplexity: 16.4731
Epoch [2/3], Step [593/3236], Loss: 2.0857, Perplexity: 8.0498
Epoch [2/3], Step [594/3236], Loss: 2.1285, Perplexity: 8.4026
Epoch [2/3], Step [595/3236], Loss: 2.0678, Perplexity: 7.9075
Epoch [2/3], Step [596/3236], Loss: 2.0759, Perplexity: 7.9721
Epoch [2/3], Step [597/3236], Loss: 2.0807, Perplexity: 8.0097
Epoch [2/3], Step [598/3236], Loss: 2.1084, Perplexity: 8.2348
Epoch [2/3], Step [599/3236], Loss: 2.1286, Perplexity: 8.4030
Epoch [2/3], Step [600/3236], Loss: 2.1406, Perplexity: 8.5047
Epoch [2/3], Step [601/3236], Loss: 2.1793, Perplexity: 8.8403
Epoch [2/3], Step [602/3236], Loss: 2.1960, Perplexity: 8.9886
Epoch [2/3], Step [603/3236], Loss: 1.9834, Perplexity: 7.2674
Epoch [2/3], Step [604/3236], Loss: 1.9874, Perplexity: 7.2963
Epoch [2/3], Step [605/3236], Loss: 2.1979, Perplexity: 9.0065
Epoch [2/3], Step [606/3236], Loss: 2.0506, Perplexity: 7.7724
Epoch [2/3], Step [607/3236], Loss: 2.1993, Perplexity: 9.0189
Epoch [2/3], Step [608/3236], Loss: 2.0551, Perplexity: 7.8074
Epoch [2/3], Step [609/3236], Loss: 2.0999, Perplexity: 8.1650
Epoch [2/3], Step [610/3236], Loss: 2.0227, Perplexity: 7.5586
Epoch [2/3], Step [611/3236], Loss: 2.1617, Perplexity: 8.6858
Epoch [2/3], Step [612/3236], Loss: 2.0142, Perplexity: 7.4951
Epoch [2/3], Step [613/3236], Loss: 2.0075, Perplexity: 7.4449
Epoch [2/3], Step [614/3236], Loss: 2.6350, Perplexity: 13.9438
Epoch [2/3], Step [615/3236], Loss: 2.0338, Perplexity: 7.6432
Epoch [2/3], Step [616/3236], Loss: 1.9890, Perplexity: 7.3084
Epoch [2/3], Step [617/3236], Loss: 2.7947, Perplexity: 16.3573
Epoch [2/3], Step [618/3236], Loss: 2.0899, Perplexity: 8.0838
Epoch [2/3], Step [619/3236], Loss: 2.0412, Perplexity: 7.7001
Epoch [2/3], Step [620/3236], Loss: 2.0465, Perplexity: 7.7410
Epoch [2/3], Step [621/3236], Loss: 2.2411, Perplexity: 9.4039
Epoch [2/3], Step [622/3236], Loss: 2.0473, Perplexity: 7.7466
Epoch [2/3], Step [623/3236], Loss: 2.2265, Perplexity: 9.2677
Epoch [2/3], Step [624/3236], Loss: 2.2034, Perplexity: 9.0562
Epoch [2/3], Step [625/3236], Loss: 2.4764, Perplexity: 11.8984
Epoch [2/3], Step [626/3236], Loss: 2.0445, Perplexity: 7.7256
Epoch [2/3], Step [627/3236], Loss: 2.1479, Perplexity: 8.5668
Epoch [2/3], Step [628/3236], Loss: 2.1881, Perplexity: 8.9184
Epoch [2/3], Step [629/3236], Loss: 2.1878, Perplexity: 8.9158
Epoch [2/3], Step [630/3236], Loss: 2.1821, Perplexity: 8.8646
Epoch [2/3], Step [631/3236], Loss: 2.0888, Perplexity: 8.0755
Epoch [2/3], Step [632/3236], Loss: 1.9837, Perplexity: 7.2693
Epoch [2/3], Step [633/3236], Loss: 2.2299, Perplexity: 9.2992
Epoch [2/3], Step [634/3236], Loss: 2.0560, Perplexity: 7.8145
Epoch [2/3], Step [635/3236], Loss: 2.4218, Perplexity: 11.2660
Epoch [2/3], Step [636/3236], Loss: 2.0603, Perplexity: 7.8482
Epoch [2/3], Step [637/3236], Loss: 2.0111, Perplexity: 7.4718
Epoch [2/3], Step [638/3236], Loss: 2.0314, Perplexity: 7.6246
Epoch [2/3], Step [639/3236], Loss: 2.0409, Perplexity: 7.6974
Epoch [2/3], Step [640/3236], Loss: 2.0303, Perplexity: 7.6164
Epoch [2/3], Step [641/3236], Loss: 2.2703, Perplexity: 9.6827
Epoch [2/3], Step [642/3236], Loss: 2.1021, Perplexity: 8.1837
Epoch [2/3], Step [643/3236], Loss: 2.0964, Perplexity: 8.1366
Epoch [2/3], Step [644/3236], Loss: 2.1236, Perplexity: 8.3612
Epoch [2/3], Step [645/3236], Loss: 2.0874, Perplexity: 8.0642
Epoch [2/3], Step [646/3236], Loss: 2.1613, Perplexity: 8.6828
Epoch [2/3], Step [647/3236], Loss: 2.0640, Perplexity: 7.8778
Epoch [2/3], Step [648/3236], Loss: 2.0599, Perplexity: 7.8449
Epoch [2/3], Step [649/3236], Loss: 2.1598, Perplexity: 8.6690
Epoch [2/3], Step [650/3236], Loss: 2.0804, Perplexity: 8.0078
Epoch [2/3], Step [651/3236], Loss: 2.1269, Perplexity: 8.3888
Epoch [2/3], Step [652/3236], Loss: 2.1012, Perplexity: 8.1759
Epoch [2/3], Step [653/3236], Loss: 2.1795, Perplexity: 8.8421
Epoch [2/3], Step [654/3236], Loss: 2.0384, Perplexity: 7.6785
Epoch [2/3], Step [655/3236], Loss: 2.0930, Perplexity: 8.1094
Epoch [2/3], Step [656/3236], Loss: 2.7190, Perplexity: 15.1644
Epoch [2/3], Step [657/3236], Loss: 2.1330, Perplexity: 8.4405
Epoch [2/3], Step [658/3236], Loss: 2.1981, Perplexity: 9.0078
Epoch [2/3], Step [659/3236], Loss: 1.9755, Perplexity: 7.2102
Epoch [2/3], Step [660/3236], Loss: 2.1382, Perplexity: 8.4842
Epoch [2/3], Step [661/3236], Loss: 2.2972, Perplexity: 9.9460
Epoch [2/3], Step [662/3236], Loss: 2.1644, Perplexity: 8.7098
Epoch [2/3], Step [663/3236], Loss: 2.0480, Perplexity: 7.7525
Epoch [2/3], Step [664/3236], Loss: 2.1255, Perplexity: 8.3775
Epoch [2/3], Step [665/3236], Loss: 2.4296, Perplexity: 11.3541
Epoch [2/3], Step [666/3236], Loss: 2.0716, Perplexity: 7.9378
Epoch [2/3], Step [667/3236], Loss: 2.2421, Perplexity: 9.4129
Epoch [2/3], Step [668/3236], Loss: 2.1794, Perplexity: 8.8412
Epoch [2/3], Step [669/3236], Loss: 2.0779, Perplexity: 7.9875
Epoch [2/3], Step [670/3236], Loss: 2.0975, Perplexity: 8.1455
Epoch [2/3], Step [671/3236], Loss: 2.3543, Perplexity: 10.5309
Epoch [2/3], Step [672/3236], Loss: 2.0456, Perplexity: 7.7336
Epoch [2/3], Step [673/3236], Loss: 2.0443, Perplexity: 7.7235
Epoch [2/3], Step [674/3236], Loss: 2.1855, Perplexity: 8.8947
Epoch [2/3], Step [675/3236], Loss: 2.2861, Perplexity: 9.8361
Epoch [2/3], Step [676/3236], Loss: 2.0401, Perplexity: 7.6917
Epoch [2/3], Step [677/3236], Loss: 2.0132, Perplexity: 7.4875
Epoch [2/3], Step [678/3236], Loss: 2.3677, Perplexity: 10.6731
Epoch [2/3], Step [679/3236], Loss: 4.2590, Perplexity: 70.7426
Epoch [2/3], Step [680/3236], Loss: 2.0589, Perplexity: 7.8373
Epoch [2/3], Step [681/3236], Loss: 2.0095, Perplexity: 7.4598
Epoch [2/3], Step [682/3236], Loss: 2.0315, Perplexity: 7.6257
Epoch [2/3], Step [683/3236], Loss: 2.3282, Perplexity: 10.2597
Epoch [2/3], Step [684/3236], Loss: 2.0886, Perplexity: 8.0733
Epoch [2/3], Step [685/3236], Loss: 2.0520, Perplexity: 7.7836
Epoch [2/3], Step [686/3236], Loss: 2.0195, Perplexity: 7.5345
Epoch [2/3], Step [687/3236], Loss: 2.1791, Perplexity: 8.8380
Epoch [2/3], Step [688/3236], Loss: 2.2030, Perplexity: 9.0523
Epoch [2/3], Step [689/3236], Loss: 1.9727, Perplexity: 7.1899
Epoch [2/3], Step [690/3236], Loss: 1.9731, Perplexity: 7.1928
Epoch [2/3], Step [691/3236], Loss: 1.9814, Perplexity: 7.2527
Epoch [2/3], Step [692/3236], Loss: 2.1385, Perplexity: 8.4869
Epoch [2/3], Step [693/3236], Loss: 2.1656, Perplexity: 8.7200
Epoch [2/3], Step [694/3236], Loss: 2.1741, Perplexity: 8.7939
Epoch [2/3], Step [695/3236], Loss: 2.1062, Perplexity: 8.2166
Epoch [2/3], Step [696/3236], Loss: 2.1054, Perplexity: 8.2104
Epoch [2/3], Step [697/3236], Loss: 2.3931, Perplexity: 10.9474
Epoch [2/3], Step [698/3236], Loss: 2.0650, Perplexity: 7.8855
Epoch [2/3], Step [699/3236], Loss: 2.0751, Perplexity: 7.9656
Epoch [2/3], Step [700/3236], Loss: 2.1916, Perplexity: 8.9494
Epoch [2/3], Step [701/3236], Loss: 3.1365, Perplexity: 23.0243
Epoch [2/3], Step [702/3236], Loss: 2.0360, Perplexity: 7.6598
Epoch [2/3], Step [703/3236], Loss: 2.1495, Perplexity: 8.5807
Epoch [2/3], Step [704/3236], Loss: 2.1776, Perplexity: 8.8247
Epoch [2/3], Step [705/3236], Loss: 2.0214, Perplexity: 7.5486
Epoch [2/3], Step [706/3236], Loss: 2.1587, Perplexity: 8.6597
Epoch [2/3], Step [707/3236], Loss: 2.2912, Perplexity: 9.8869
Epoch [2/3], Step [708/3236], Loss: 2.2280, Perplexity: 9.2813
Epoch [2/3], Step [709/3236], Loss: 2.1413, Perplexity: 8.5103
Epoch [2/3], Step [710/3236], Loss: 2.0684, Perplexity: 7.9124
Epoch [2/3], Step [711/3236], Loss: 2.0554, Perplexity: 7.8099
Epoch [2/3], Step [712/3236], Loss: 2.2734, Perplexity: 9.7126
Epoch [2/3], Step [713/3236], Loss: 2.1487, Perplexity: 8.5733
Epoch [2/3], Step [714/3236], Loss: 2.0266, Perplexity: 7.5880
Epoch [2/3], Step [715/3236], Loss: 2.0200, Perplexity: 7.5383
Epoch [2/3], Step [716/3236], Loss: 2.6752, Perplexity: 14.5153
Epoch [2/3], Step [717/3236], Loss: 1.9411, Perplexity: 6.9662
Epoch [2/3], Step [718/3236], Loss: 2.0847, Perplexity: 8.0426
Epoch [2/3], Step [719/3236], Loss: 2.0313, Perplexity: 7.6237
Epoch [2/3], Step [720/3236], Loss: 2.2301, Perplexity: 9.3009
Epoch [2/3], Step [721/3236], Loss: 2.0004, Perplexity: 7.3923
Epoch [2/3], Step [722/3236], Loss: 2.2740, Perplexity: 9.7177
Epoch [2/3], Step [723/3236], Loss: 2.3299, Perplexity: 10.2764
Epoch [2/3], Step [724/3236], Loss: 2.1172, Perplexity: 8.3081
Epoch [2/3], Step [725/3236], Loss: 2.1030, Perplexity: 8.1908
Epoch [2/3], Step [726/3236], Loss: 2.1160, Perplexity: 8.2976
Epoch [2/3], Step [727/3236], Loss: 2.2012, Perplexity: 9.0360
Epoch [2/3], Step [728/3236], Loss: 2.4864, Perplexity: 12.0183
Epoch [2/3], Step [729/3236], Loss: 3.1466, Perplexity: 23.2561
Epoch [2/3], Step [730/3236], Loss: 2.0925, Perplexity: 8.1051
Epoch [2/3], Step [731/3236], Loss: 2.0598, Perplexity: 7.8441
Epoch [2/3], Step [732/3236], Loss: 2.0699, Perplexity: 7.9244
Epoch [2/3], Step [733/3236], Loss: 2.1148, Perplexity: 8.2877
Epoch [2/3], Step [734/3236], Loss: 2.1717, Perplexity: 8.7729
Epoch [2/3], Step [735/3236], Loss: 2.2778, Perplexity: 9.7549
Epoch [2/3], Step [736/3236], Loss: 2.1554, Perplexity: 8.6314
Epoch [2/3], Step [737/3236], Loss: 1.9974, Perplexity: 7.3700
Epoch [2/3], Step [738/3236], Loss: 2.2929, Perplexity: 9.9032
Epoch [2/3], Step [739/3236], Loss: 2.1464, Perplexity: 8.5536
Epoch [2/3], Step [740/3236], Loss: 2.1580, Perplexity: 8.6537
Epoch [2/3], Step [741/3236], Loss: 2.2418, Perplexity: 9.4102
Epoch [2/3], Step [742/3236], Loss: 2.3146, Perplexity: 10.1212
Epoch [2/3], Step [743/3236], Loss: 2.3923, Perplexity: 10.9387
Epoch [2/3], Step [744/3236], Loss: 2.3214, Perplexity: 10.1897
Epoch [2/3], Step [745/3236], Loss: 2.1358, Perplexity: 8.4642
Epoch [2/3], Step [746/3236], Loss: 2.1233, Perplexity: 8.3590
Epoch [2/3], Step [747/3236], Loss: 2.1881, Perplexity: 8.9182
Epoch [2/3], Step [748/3236], Loss: 2.2973, Perplexity: 9.9470
Epoch [2/3], Step [749/3236], Loss: 2.1237, Perplexity: 8.3618
Epoch [2/3], Step [750/3236], Loss: 2.2286, Perplexity: 9.2871
Epoch [2/3], Step [751/3236], Loss: 2.4706, Perplexity: 11.8296
Epoch [2/3], Step [752/3236], Loss: 2.1879, Perplexity: 8.9162
Epoch [2/3], Step [753/3236], Loss: 2.1726, Perplexity: 8.7811
Epoch [2/3], Step [754/3236], Loss: 2.6333, Perplexity: 13.9200
Epoch [2/3], Step [755/3236], Loss: 2.2405, Perplexity: 9.3982
Epoch [2/3], Step [756/3236], Loss: 2.0265, Perplexity: 7.5872
Epoch [2/3], Step [757/3236], Loss: 2.3310, Perplexity: 10.2887
Epoch [2/3], Step [758/3236], Loss: 2.0221, Perplexity: 7.5540
Epoch [2/3], Step [759/3236], Loss: 2.1055, Perplexity: 8.2113
Epoch [2/3], Step [760/3236], Loss: 2.0683, Perplexity: 7.9111
Epoch [2/3], Step [761/3236], Loss: 2.1324, Perplexity: 8.4353
Epoch [2/3], Step [762/3236], Loss: 2.2827, Perplexity: 9.8027
Epoch [2/3], Step [763/3236], Loss: 2.3217, Perplexity: 10.1930
Epoch [2/3], Step [764/3236], Loss: 2.2764, Perplexity: 9.7415
Epoch [2/3], Step [765/3236], Loss: 2.2266, Perplexity: 9.2681
Epoch [2/3], Step [766/3236], Loss: 2.4882, Perplexity: 12.0394
Epoch [2/3], Step [767/3236], Loss: 2.1318, Perplexity: 8.4297
Epoch [2/3], Step [768/3236], Loss: 2.1347, Perplexity: 8.4548
Epoch [2/3], Step [769/3236], Loss: 2.0566, Perplexity: 7.8190
Epoch [2/3], Step [770/3236], Loss: 2.1524, Perplexity: 8.6056
Epoch [2/3], Step [771/3236], Loss: 2.1320, Perplexity: 8.4320
Epoch [2/3], Step [772/3236], Loss: 4.0895, Perplexity: 59.7078
Epoch [2/3], Step [773/3236], Loss: 1.9854, Perplexity: 7.2819
Epoch [2/3], Step [774/3236], Loss: 2.2842, Perplexity: 9.8177
Epoch [2/3], Step [775/3236], Loss: 2.1339, Perplexity: 8.4473
Epoch [2/3], Step [776/3236], Loss: 2.2437, Perplexity: 9.4286
Epoch [2/3], Step [777/3236], Loss: 2.2264, Perplexity: 9.2661
Epoch [2/3], Step [778/3236], Loss: 2.1069, Perplexity: 8.2231
Epoch [2/3], Step [779/3236], Loss: 2.4769, Perplexity: 11.9044
Epoch [2/3], Step [780/3236], Loss: 1.9964, Perplexity: 7.3623
Epoch [2/3], Step [781/3236], Loss: 2.1772, Perplexity: 8.8212
Epoch [2/3], Step [782/3236], Loss: 2.3436, Perplexity: 10.4188
Epoch [2/3], Step [783/3236], Loss: 2.0990, Perplexity: 8.1580
Epoch [2/3], Step [784/3236], Loss: 2.1245, Perplexity: 8.3691
Epoch [2/3], Step [785/3236], Loss: 1.9436, Perplexity: 6.9840
Epoch [2/3], Step [786/3236], Loss: 2.1785, Perplexity: 8.8332
Epoch [2/3], Step [787/3236], Loss: 2.0508, Perplexity: 7.7740
Epoch [2/3], Step [788/3236], Loss: 2.0917, Perplexity: 8.0989
Epoch [2/3], Step [789/3236], Loss: 2.1640, Perplexity: 8.7056
Epoch [2/3], Step [790/3236], Loss: 2.1719, Perplexity: 8.7749
Epoch [2/3], Step [791/3236], Loss: 2.3029, Perplexity: 10.0031
Epoch [2/3], Step [792/3236], Loss: 2.0131, Perplexity: 7.4865
Epoch [2/3], Step [793/3236], Loss: 2.1414, Perplexity: 8.5117
Epoch [2/3], Step [794/3236], Loss: 1.9769, Perplexity: 7.2204
Epoch [2/3], Step [795/3236], Loss: 2.0123, Perplexity: 7.4807
Epoch [2/3], Step [796/3236], Loss: 2.2539, Perplexity: 9.5253
Epoch [2/3], Step [797/3236], Loss: 2.1863, Perplexity: 8.9020
Epoch [2/3], Step [798/3236], Loss: 2.0382, Perplexity: 7.6768
Epoch [2/3], Step [799/3236], Loss: 1.9412, Perplexity: 6.9668
Epoch [2/3], Step [800/3236], Loss: 2.1140, Perplexity: 8.2815
Epoch [2/3], Step [801/3236], Loss: 2.0303, Perplexity: 7.6163
Epoch [2/3], Step [802/3236], Loss: 2.3984, Perplexity: 11.0050
Epoch [2/3], Step [803/3236], Loss: 2.1429, Perplexity: 8.5241
Epoch [2/3], Step [804/3236], Loss: 2.0462, Perplexity: 7.7384
Epoch [2/3], Step [805/3236], Loss: 2.0498, Perplexity: 7.7661
Epoch [2/3], Step [806/3236], Loss: 2.0596, Perplexity: 7.8430
Epoch [2/3], Step [807/3236], Loss: 2.1600, Perplexity: 8.6713
Epoch [2/3], Step [808/3236], Loss: 2.0580, Perplexity: 7.8301
Epoch [2/3], Step [809/3236], Loss: 2.1970, Perplexity: 8.9981
Epoch [2/3], Step [810/3236], Loss: 2.1458, Perplexity: 8.5490
Epoch [2/3], Step [811/3236], Loss: 2.0588, Perplexity: 7.8365
Epoch [2/3], Step [812/3236], Loss: 2.0562, Perplexity: 7.8159
Epoch [2/3], Step [813/3236], Loss: 2.1198, Perplexity: 8.3293
Epoch [2/3], Step [814/3236], Loss: 2.0543, Perplexity: 7.8015
Epoch [2/3], Step [815/3236], Loss: 1.9293, Perplexity: 6.8844
Epoch [2/3], Step [816/3236], Loss: 2.3048, Perplexity: 10.0219
Epoch [2/3], Step [817/3236], Loss: 2.2393, Perplexity: 9.3864
Epoch [2/3], Step [818/3236], Loss: 2.9275, Perplexity: 18.6814
Epoch [2/3], Step [819/3236], Loss: 2.1250, Perplexity: 8.3725
Epoch [2/3], Step [820/3236], Loss: 2.0456, Perplexity: 7.7339
Epoch [2/3], Step [821/3236], Loss: 2.1102, Perplexity: 8.2499
Epoch [2/3], Step [822/3236], Loss: 2.0211, Perplexity: 7.5464
Epoch [2/3], Step [823/3236], Loss: 2.0343, Perplexity: 7.6469
Epoch [2/3], Step [824/3236], Loss: 1.9819, Perplexity: 7.2563
Epoch [2/3], Step [825/3236], Loss: 2.3747, Perplexity: 10.7479
Epoch [2/3], Step [826/3236], Loss: 2.0329, Perplexity: 7.6358
Epoch [2/3], Step [827/3236], Loss: 3.0681, Perplexity: 21.5012
Epoch [2/3], Step [828/3236], Loss: 2.2188, Perplexity: 9.1959
Epoch [2/3], Step [829/3236], Loss: 2.3083, Perplexity: 10.0573
Epoch [2/3], Step [830/3236], Loss: 2.0746, Perplexity: 7.9616
Epoch [2/3], Step [831/3236], Loss: 2.2104, Perplexity: 9.1190
Epoch [2/3], Step [832/3236], Loss: 2.0328, Perplexity: 7.6356
Epoch [2/3], Step [833/3236], Loss: 2.2036, Perplexity: 9.0579
Epoch [2/3], Step [834/3236], Loss: 2.1565, Perplexity: 8.6405
Epoch [2/3], Step [835/3236], Loss: 2.0015, Perplexity: 7.3999
Epoch [2/3], Step [836/3236], Loss: 2.1020, Perplexity: 8.1826
Epoch [2/3], Step [837/3236], Loss: 2.0961, Perplexity: 8.1346
Epoch [2/3], Step [838/3236], Loss: 2.6840, Perplexity: 14.6434
Epoch [2/3], Step [839/3236], Loss: 2.3570, Perplexity: 10.5589
Epoch [2/3], Step [840/3236], Loss: 2.0371, Perplexity: 7.6680
Epoch [2/3], Step [841/3236], Loss: 2.1882, Perplexity: 8.9187
Epoch [2/3], Step [842/3236], Loss: 2.0324, Perplexity: 7.6326
Epoch [2/3], Step [843/3236], Loss: 2.0612, Perplexity: 7.8556
Epoch [2/3], Step [844/3236], Loss: 2.2736, Perplexity: 9.7141
Epoch [2/3], Step [845/3236], Loss: 2.0392, Perplexity: 7.6844
Epoch [2/3], Step [846/3236], Loss: 2.7572, Perplexity: 15.7557
Epoch [2/3], Step [847/3236], Loss: 2.1413, Perplexity: 8.5103
Epoch [2/3], Step [848/3236], Loss: 2.0971, Perplexity: 8.1425
Epoch [2/3], Step [849/3236], Loss: 2.2002, Perplexity: 9.0272
Epoch [2/3], Step [850/3236], Loss: 1.9191, Perplexity: 6.8151
Epoch [2/3], Step [851/3236], Loss: 2.0561, Perplexity: 7.8155
Epoch [2/3], Step [852/3236], Loss: 2.4429, Perplexity: 11.5063
Epoch [2/3], Step [853/3236], Loss: 2.7074, Perplexity: 14.9899
Epoch [2/3], Step [854/3236], Loss: 2.1924, Perplexity: 8.9569
Epoch [2/3], Step [855/3236], Loss: 2.0165, Perplexity: 7.5119
Epoch [2/3], Step [856/3236], Loss: 2.2333, Perplexity: 9.3309
Epoch [2/3], Step [857/3236], Loss: 1.9881, Perplexity: 7.3014
Epoch [2/3], Step [858/3236], Loss: 2.2369, Perplexity: 9.3638
Epoch [2/3], Step [859/3236], Loss: 2.0743, Perplexity: 7.9591
Epoch [2/3], Step [860/3236], Loss: 2.1482, Perplexity: 8.5692
Epoch [2/3], Step [861/3236], Loss: 2.0090, Perplexity: 7.4557
Epoch [2/3], Step [862/3236], Loss: 2.2262, Perplexity: 9.2650
Epoch [2/3], Step [863/3236], Loss: 2.3978, Perplexity: 10.9985
Epoch [2/3], Step [864/3236], Loss: 2.1159, Perplexity: 8.2973
Epoch [2/3], Step [865/3236], Loss: 2.1689, Perplexity: 8.7482
Epoch [2/3], Step [866/3236], Loss: 2.4157, Perplexity: 11.1978
Epoch [2/3], Step [867/3236], Loss: 2.3616, Perplexity: 10.6084
Epoch [2/3], Step [868/3236], Loss: 2.2216, Perplexity: 9.2217
Epoch [2/3], Step [869/3236], Loss: 2.0741, Perplexity: 7.9574
Epoch [2/3], Step [870/3236], Loss: 2.3075, Perplexity: 10.0489
Epoch [2/3], Step [871/3236], Loss: 1.9803, Perplexity: 7.2452
Epoch [2/3], Step [872/3236], Loss: 2.1950, Perplexity: 8.9800
Epoch [2/3], Step [873/3236], Loss: 1.9872, Perplexity: 7.2950
Epoch [2/3], Step [874/3236], Loss: 2.1983, Perplexity: 9.0100
Epoch [2/3], Step [875/3236], Loss: 2.0905, Perplexity: 8.0887
Epoch [2/3], Step [876/3236], Loss: 2.0378, Perplexity: 7.6736
Epoch [2/3], Step [877/3236], Loss: 2.2515, Perplexity: 9.5022
Epoch [2/3], Step [878/3236], Loss: 2.3324, Perplexity: 10.3026
Epoch [2/3], Step [879/3236], Loss: 2.0929, Perplexity: 8.1082
Epoch [2/3], Step [880/3236], Loss: 2.0498, Perplexity: 7.7667
Epoch [2/3], Step [881/3236], Loss: 2.0163, Perplexity: 7.5102
Epoch [2/3], Step [882/3236], Loss: 1.9431, Perplexity: 6.9803
Epoch [2/3], Step [883/3236], Loss: 2.3029, Perplexity: 10.0033
Epoch [2/3], Step [884/3236], Loss: 2.2060, Perplexity: 9.0791
Epoch [2/3], Step [885/3236], Loss: 2.0294, Perplexity: 7.6092
Epoch [2/3], Step [886/3236], Loss: 2.0383, Perplexity: 7.6777
Epoch [2/3], Step [887/3236], Loss: 2.0881, Perplexity: 8.0696
Epoch [2/3], Step [888/3236], Loss: 2.1028, Perplexity: 8.1888
Epoch [2/3], Step [889/3236], Loss: 2.4117, Perplexity: 11.1533
Epoch [2/3], Step [890/3236], Loss: 2.1095, Perplexity: 8.2445
Epoch [2/3], Step [891/3236], Loss: 2.1619, Perplexity: 8.6879
Epoch [2/3], Step [892/3236], Loss: 2.1088, Perplexity: 8.2380
Epoch [2/3], Step [893/3236], Loss: 2.2083, Perplexity: 9.0998
Epoch [2/3], Step [894/3236], Loss: 2.0786, Perplexity: 7.9932
Epoch [2/3], Step [895/3236], Loss: 2.0709, Perplexity: 7.9321
Epoch [2/3], Step [896/3236], Loss: 2.2526, Perplexity: 9.5126
Epoch [2/3], Step [897/3236], Loss: 2.1702, Perplexity: 8.7600
Epoch [2/3], Step [898/3236], Loss: 2.2378, Perplexity: 9.3732
Epoch [2/3], Step [899/3236], Loss: 2.0875, Perplexity: 8.0649
Epoch [2/3], Step [900/3236], Loss: 2.2322, Perplexity: 9.3200
Epoch [2/3], Step [901/3236], Loss: 2.2596, Perplexity: 9.5793
Epoch [2/3], Step [902/3236], Loss: 1.9938, Perplexity: 7.3435
Epoch [2/3], Step [903/3236], Loss: 2.2954, Perplexity: 9.9284
Epoch [2/3], Step [904/3236], Loss: 2.1109, Perplexity: 8.2558
Epoch [2/3], Step [905/3236], Loss: 2.5475, Perplexity: 12.7748
Epoch [2/3], Step [906/3236], Loss: 2.1735, Perplexity: 8.7889
Epoch [2/3], Step [907/3236], Loss: 2.0178, Perplexity: 7.5214
Epoch [2/3], Step [908/3236], Loss: 1.9689, Perplexity: 7.1627
Epoch [2/3], Step [909/3236], Loss: 1.9803, Perplexity: 7.2447
Epoch [2/3], Step [910/3236], Loss: 2.3746, Perplexity: 10.7468
Epoch [2/3], Step [911/3236], Loss: 2.0408, Perplexity: 7.6966
Epoch [2/3], Step [912/3236], Loss: 2.2886, Perplexity: 9.8614
Epoch [2/3], Step [913/3236], Loss: 1.9733, Perplexity: 7.1942
Epoch [2/3], Step [914/3236], Loss: 2.2346, Perplexity: 9.3427
Epoch [2/3], Step [915/3236], Loss: 2.0314, Perplexity: 7.6251
Epoch [2/3], Step [916/3236], Loss: 2.2730, Perplexity: 9.7089
Epoch [2/3], Step [917/3236], Loss: 2.2058, Perplexity: 9.0777
Epoch [2/3], Step [918/3236], Loss: 2.1729, Perplexity: 8.7834
Epoch [2/3], Step [919/3236], Loss: 2.1190, Perplexity: 8.3224
Epoch [2/3], Step [920/3236], Loss: 2.1211, Perplexity: 8.3406
Epoch [2/3], Step [921/3236], Loss: 2.0515, Perplexity: 7.7798
Epoch [2/3], Step [922/3236], Loss: 2.0054, Perplexity: 7.4289
Epoch [2/3], Step [923/3236], Loss: 2.1308, Perplexity: 8.4214
Epoch [2/3], Step [924/3236], Loss: 2.1364, Perplexity: 8.4692
Epoch [2/3], Step [925/3236], Loss: 2.2073, Perplexity: 9.0915
Epoch [2/3], Step [926/3236], Loss: 2.2969, Perplexity: 9.9430
Epoch [2/3], Step [927/3236], Loss: 2.0300, Perplexity: 7.6142
Epoch [2/3], Step [928/3236], Loss: 1.9589, Perplexity: 7.0912
Epoch [2/3], Step [929/3236], Loss: 2.7005, Perplexity: 14.8876
Epoch [2/3], Step [930/3236], Loss: 2.5752, Perplexity: 13.1335
Epoch [2/3], Step [931/3236], Loss: 2.3132, Perplexity: 10.1063
Epoch [2/3], Step [932/3236], Loss: 1.9775, Perplexity: 7.2244
Epoch [2/3], Step [933/3236], Loss: 2.1610, Perplexity: 8.6794
Epoch [2/3], Step [934/3236], Loss: 2.1426, Perplexity: 8.5219
Epoch [2/3], Step [935/3236], Loss: 2.7150, Perplexity: 15.1050
Epoch [2/3], Step [936/3236], Loss: 1.9467, Perplexity: 7.0055
Epoch [2/3], Step [937/3236], Loss: 2.1465, Perplexity: 8.5547
Epoch [2/3], Step [938/3236], Loss: 2.3888, Perplexity: 10.9007
Epoch [2/3], Step [939/3236], Loss: 2.1276, Perplexity: 8.3945
Epoch [2/3], Step [940/3236], Loss: 2.0580, Perplexity: 7.8305
Epoch [2/3], Step [941/3236], Loss: 2.1525, Perplexity: 8.6064
Epoch [2/3], Step [942/3236], Loss: 2.6333, Perplexity: 13.9201
Epoch [2/3], Step [943/3236], Loss: 2.1623, Perplexity: 8.6915
Epoch [2/3], Step [944/3236], Loss: 2.7907, Perplexity: 16.2931
Epoch [2/3], Step [945/3236], Loss: 2.1072, Perplexity: 8.2250
Epoch [2/3], Step [946/3236], Loss: 1.9893, Perplexity: 7.3106
Epoch [2/3], Step [947/3236], Loss: 2.1421, Perplexity: 8.5173
Epoch [2/3], Step [948/3236], Loss: 2.0487, Perplexity: 7.7579
Epoch [2/3], Step [949/3236], Loss: 2.0924, Perplexity: 8.1043
Epoch [2/3], Step [950/3236], Loss: 2.0858, Perplexity: 8.0514
Epoch [2/3], Step [951/3236], Loss: 2.0918, Perplexity: 8.0994
Epoch [2/3], Step [952/3236], Loss: 2.2314, Perplexity: 9.3127
Epoch [2/3], Step [953/3236], Loss: 2.1327, Perplexity: 8.4373
Epoch [2/3], Step [954/3236], Loss: 1.9275, Perplexity: 6.8720
Epoch [2/3], Step [955/3236], Loss: 2.1458, Perplexity: 8.5489
Epoch [2/3], Step [956/3236], Loss: 2.1233, Perplexity: 8.3587
Epoch [2/3], Step [957/3236], Loss: 1.9487, Perplexity: 7.0194
Epoch [2/3], Step [958/3236], Loss: 2.0353, Perplexity: 7.6547
Epoch [2/3], Step [959/3236], Loss: 2.0443, Perplexity: 7.7234
Epoch [2/3], Step [960/3236], Loss: 2.4875, Perplexity: 12.0316
Epoch [2/3], Step [961/3236], Loss: 2.0636, Perplexity: 7.8743
Epoch [2/3], Step [962/3236], Loss: 2.0584, Perplexity: 7.8334
Epoch [2/3], Step [963/3236], Loss: 2.0782, Perplexity: 7.9899
Epoch [2/3], Step [964/3236], Loss: 2.2047, Perplexity: 9.0672
Epoch [2/3], Step [965/3236], Loss: 2.5473, Perplexity: 12.7732
Epoch [2/3], Step [966/3236], Loss: 2.2758, Perplexity: 9.7357
Epoch [2/3], Step [967/3236], Loss: 2.1565, Perplexity: 8.6406
Epoch [2/3], Step [968/3236], Loss: 2.1162, Perplexity: 8.2995
Epoch [2/3], Step [969/3236], Loss: 2.1054, Perplexity: 8.2107
Epoch [2/3], Step [970/3236], Loss: 2.2151, Perplexity: 9.1624
Epoch [2/3], Step [971/3236], Loss: 2.1239, Perplexity: 8.3638
Epoch [2/3], Step [972/3236], Loss: 2.1226, Perplexity: 8.3532
Epoch [2/3], Step [973/3236], Loss: 2.0917, Perplexity: 8.0989
Epoch [2/3], Step [974/3236], Loss: 2.1283, Perplexity: 8.4008
Epoch [2/3], Step [975/3236], Loss: 2.1393, Perplexity: 8.4938
Epoch [2/3], Step [976/3236], Loss: 2.0063, Perplexity: 7.4359
Epoch [2/3], Step [977/3236], Loss: 2.3373, Perplexity: 10.3529
Epoch [2/3], Step [978/3236], Loss: 2.0439, Perplexity: 7.7209
Epoch [2/3], Step [979/3236], Loss: 1.9775, Perplexity: 7.2250
Epoch [2/3], Step [980/3236], Loss: 2.0125, Perplexity: 7.4819
Epoch [2/3], Step [981/3236], Loss: 2.0459, Perplexity: 7.7361
Epoch [2/3], Step [982/3236], Loss: 2.2455, Perplexity: 9.4448
Epoch [2/3], Step [983/3236], Loss: 2.1165, Perplexity: 8.3022
Epoch [2/3], Step [984/3236], Loss: 2.0722, Perplexity: 7.9421
Epoch [2/3], Step [985/3236], Loss: 2.1532, Perplexity: 8.6121
Epoch [2/3], Step [986/3236], Loss: 2.0459, Perplexity: 7.7360
Epoch [2/3], Step [987/3236], Loss: 1.9560, Perplexity: 7.0707
Epoch [2/3], Step [988/3236], Loss: 2.0673, Perplexity: 7.9035
Epoch [2/3], Step [989/3236], Loss: 2.0170, Perplexity: 7.5160
Epoch [2/3], Step [990/3236], Loss: 2.1366, Perplexity: 8.4706
Epoch [2/3], Step [991/3236], Loss: 2.0132, Perplexity: 7.4870
Epoch [2/3], Step [992/3236], Loss: 2.1323, Perplexity: 8.4341
Epoch [2/3], Step [993/3236], Loss: 2.1288, Perplexity: 8.4048
Epoch [2/3], Step [994/3236], Loss: 1.9950, Perplexity: 7.3521
Epoch [2/3], Step [995/3236], Loss: 2.8681, Perplexity: 17.6043
Epoch [2/3], Step [996/3236], Loss: 2.0512, Perplexity: 7.7774
Epoch [2/3], Step [997/3236], Loss: 2.1436, Perplexity: 8.5303
Epoch [2/3], Step [998/3236], Loss: 2.0364, Perplexity: 7.6633
Epoch [2/3], Step [999/3236], Loss: 2.0422, Perplexity: 7.7077
Epoch [2/3], Step [1000/3236], Loss: 2.3558, Perplexity: 10.5471
Epoch [2/3], Step [1001/3236], Loss: 2.0289, Perplexity: 7.6058
Epoch [2/3], Step [1002/3236], Loss: 2.0597, Perplexity: 7.8434
Epoch [2/3], Step [1003/3236], Loss: 2.4489, Perplexity: 11.5759
Epoch [2/3], Step [1004/3236], Loss: 2.1635, Perplexity: 8.7014
Epoch [2/3], Step [1005/3236], Loss: 2.0677, Perplexity: 7.9066
Epoch [2/3], Step [1006/3236], Loss: 2.0592, Perplexity: 7.8395
Epoch [2/3], Step [1007/3236], Loss: 2.1805, Perplexity: 8.8505
Epoch [2/3], Step [1008/3236], Loss: 2.1466, Perplexity: 8.5558
Epoch [2/3], Step [1009/3236], Loss: 2.1838, Perplexity: 8.8802
Epoch [2/3], Step [1010/3236], Loss: 2.2184, Perplexity: 9.1927
Epoch [2/3], Step [1011/3236], Loss: 1.9263, Perplexity: 6.8644
Epoch [2/3], Step [1012/3236], Loss: 2.1698, Perplexity: 8.7568
Epoch [2/3], Step [1013/3236], Loss: 2.1737, Perplexity: 8.7907
Epoch [2/3], Step [1014/3236], Loss: 2.2230, Perplexity: 9.2353
Epoch [2/3], Step [1015/3236], Loss: 2.0138, Perplexity: 7.4920
Epoch [2/3], Step [1016/3236], Loss: 2.0890, Perplexity: 8.0769
Epoch [2/3], Step [1017/3236], Loss: 2.1670, Perplexity: 8.7318
Epoch [2/3], Step [1018/3236], Loss: 2.2439, Perplexity: 9.4300
Epoch [2/3], Step [1019/3236], Loss: 2.1577, Perplexity: 8.6508
Epoch [2/3], Step [1020/3236], Loss: 2.1143, Perplexity: 8.2835
Epoch [2/3], Step [1021/3236], Loss: 2.2436, Perplexity: 9.4274
Epoch [2/3], Step [1022/3236], Loss: 2.0041, Perplexity: 7.4196
Epoch [2/3], Step [1023/3236], Loss: 2.6870, Perplexity: 14.6868
Epoch [2/3], Step [1024/3236], Loss: 2.0467, Perplexity: 7.7425
Epoch [2/3], Step [1025/3236], Loss: 2.0568, Perplexity: 7.8212
Epoch [2/3], Step [1026/3236], Loss: 2.0386, Perplexity: 7.6797
Epoch [2/3], Step [1027/3236], Loss: 2.2482, Perplexity: 9.4703
Epoch [2/3], Step [1028/3236], Loss: 2.4758, Perplexity: 11.8912
Epoch [2/3], Step [1029/3236], Loss: 2.0493, Perplexity: 7.7623
Epoch [2/3], Step [1030/3236], Loss: 2.1336, Perplexity: 8.4453
Epoch [2/3], Step [1031/3236], Loss: 1.9052, Perplexity: 6.7210
Epoch [2/3], Step [1032/3236], Loss: 2.3831, Perplexity: 10.8387
Epoch [2/3], Step [1033/3236], Loss: 2.2000, Perplexity: 9.0255
Epoch [2/3], Step [1034/3236], Loss: 2.1444, Perplexity: 8.5366
Epoch [2/3], Step [1035/3236], Loss: 2.0151, Perplexity: 7.5016
Epoch [2/3], Step [1036/3236], Loss: 2.1391, Perplexity: 8.4915
Epoch [2/3], Step [1037/3236], Loss: 2.1502, Perplexity: 8.5866
Epoch [2/3], Step [1038/3236], Loss: 2.1922, Perplexity: 8.9547
Epoch [2/3], Step [1039/3236], Loss: 2.0662, Perplexity: 7.8945
Epoch [2/3], Step [1040/3236], Loss: 2.0252, Perplexity: 7.5774
Epoch [2/3], Step [1041/3236], Loss: 2.3067, Perplexity: 10.0410
Epoch [2/3], Step [1042/3236], Loss: 2.1458, Perplexity: 8.5492
Epoch [2/3], Step [1043/3236], Loss: 2.5236, Perplexity: 12.4739
Epoch [2/3], Step [1044/3236], Loss: 2.1749, Perplexity: 8.8009
Epoch [2/3], Step [1045/3236], Loss: 1.9856, Perplexity: 7.2835
Epoch [2/3], Step [1046/3236], Loss: 2.1738, Perplexity: 8.7916
Epoch [2/3], Step [1047/3236], Loss: 2.1425, Perplexity: 8.5209
Epoch [2/3], Step [1048/3236], Loss: 1.8624, Perplexity: 6.4390
Epoch [2/3], Step [1049/3236], Loss: 1.9858, Perplexity: 7.2846
Epoch [2/3], Step [1050/3236], Loss: 2.0113, Perplexity: 7.4731
Epoch [2/3], Step [1051/3236], Loss: 2.1022, Perplexity: 8.1841
Epoch [2/3], Step [1052/3236], Loss: 2.0794, Perplexity: 7.9994
Epoch [2/3], Step [1053/3236], Loss: 2.5295, Perplexity: 12.5470
Epoch [2/3], Step [1054/3236], Loss: 2.2855, Perplexity: 9.8310
Epoch [2/3], Step [1055/3236], Loss: 1.9686, Perplexity: 7.1609
Epoch [2/3], Step [1056/3236], Loss: 2.1723, Perplexity: 8.7788
Epoch [2/3], Step [1057/3236], Loss: 2.0768, Perplexity: 7.9789
Epoch [2/3], Step [1058/3236], Loss: 2.0628, Perplexity: 7.8682
Epoch [2/3], Step [1059/3236], Loss: 2.2360, Perplexity: 9.3562
Epoch [2/3], Step [1060/3236], Loss: 2.0700, Perplexity: 7.9249
Epoch [2/3], Step [1061/3236], Loss: 2.0654, Perplexity: 7.8887
Epoch [2/3], Step [1062/3236], Loss: 2.0396, Perplexity: 7.6873
Epoch [2/3], Step [1063/3236], Loss: 2.0205, Perplexity: 7.5425
Epoch [2/3], Step [1064/3236], Loss: 1.9827, Perplexity: 7.2626
Epoch [2/3], Step [1065/3236], Loss: 2.6680, Perplexity: 14.4105
Epoch [2/3], Step [1066/3236], Loss: 2.0003, Perplexity: 7.3910
Epoch [2/3], Step [1067/3236], Loss: 2.1124, Perplexity: 8.2680
Epoch [2/3], Step [1068/3236], Loss: 2.0205, Perplexity: 7.5419
Epoch [2/3], Step [1069/3236], Loss: 2.1233, Perplexity: 8.3586
Epoch [2/3], Step [1070/3236], Loss: 2.1028, Perplexity: 8.1890
Epoch [2/3], Step [1071/3236], Loss: 2.1400, Perplexity: 8.4993
Epoch [2/3], Step [1072/3236], Loss: 1.9303, Perplexity: 6.8918
Epoch [2/3], Step [1073/3236], Loss: 2.0672, Perplexity: 7.9029
Epoch [2/3], Step [1074/3236], Loss: 2.0103, Perplexity: 7.4652
Epoch [2/3], Step [1075/3236], Loss: 2.3378, Perplexity: 10.3584
Epoch [2/3], Step [1076/3236], Loss: 2.0755, Perplexity: 7.9684
Epoch [2/3], Step [1077/3236], Loss: 2.1738, Perplexity: 8.7917
Epoch [2/3], Step [1078/3236], Loss: 2.0075, Perplexity: 7.4445
Epoch [2/3], Step [1079/3236], Loss: 2.1541, Perplexity: 8.6197
Epoch [2/3], Step [1080/3236], Loss: 2.1344, Perplexity: 8.4519
Epoch [2/3], Step [1081/3236], Loss: 2.2639, Perplexity: 9.6207
Epoch [2/3], Step [1082/3236], Loss: 1.9216, Perplexity: 6.8318
Epoch [2/3], Step [1083/3236], Loss: 2.0668, Perplexity: 7.8993
Epoch [2/3], Step [1084/3236], Loss: 2.0791, Perplexity: 7.9974
Epoch [2/3], Step [1085/3236], Loss: 2.2045, Perplexity: 9.0655
Epoch [2/3], Step [1086/3236], Loss: 2.0316, Perplexity: 7.6265
Epoch [2/3], Step [1087/3236], Loss: 2.0549, Perplexity: 7.8060
Epoch [2/3], Step [1088/3236], Loss: 2.0139, Perplexity: 7.4926
Epoch [2/3], Step [1089/3236], Loss: 2.1321, Perplexity: 8.4326
Epoch [2/3], Step [1090/3236], Loss: 2.2040, Perplexity: 9.0614
Epoch [2/3], Step [1091/3236], Loss: 2.1266, Perplexity: 8.3862
Epoch [2/3], Step [1092/3236], Loss: 1.9510, Perplexity: 7.0355
Epoch [2/3], Step [1093/3236], Loss: 2.2492, Perplexity: 9.4801
Epoch [2/3], Step [1094/3236], Loss: 2.5057, Perplexity: 12.2519
Epoch [2/3], Step [1095/3236], Loss: 2.6309, Perplexity: 13.8866
Epoch [2/3], Step [1096/3236], Loss: 2.0951, Perplexity: 8.1266
Epoch [2/3], Step [1097/3236], Loss: 1.9898, Perplexity: 7.3140
Epoch [2/3], Step [1098/3236], Loss: 2.1701, Perplexity: 8.7591
Epoch [2/3], Step [1099/3236], Loss: 2.0372, Perplexity: 7.6689
Epoch [2/3], Step [1100/3236], Loss: 2.1663, Perplexity: 8.7261
Epoch [2/3], Step [1101/3236], Loss: 2.0162, Perplexity: 7.5095
Epoch [2/3], Step [1102/3236], Loss: 2.2316, Perplexity: 9.3148
Epoch [2/3], Step [1103/3236], Loss: 2.0837, Perplexity: 8.0340
Epoch [2/3], Step [1104/3236], Loss: 2.0797, Perplexity: 8.0021
Epoch [2/3], Step [1105/3236], Loss: 2.1475, Perplexity: 8.5633
Epoch [2/3], Step [1106/3236], Loss: 1.9774, Perplexity: 7.2236
Epoch [2/3], Step [1107/3236], Loss: 2.6035, Perplexity: 13.5108
Epoch [2/3], Step [1108/3236], Loss: 2.0788, Perplexity: 7.9952
Epoch [2/3], Step [1109/3236], Loss: 2.0494, Perplexity: 7.7636
Epoch [2/3], Step [1110/3236], Loss: 2.1559, Perplexity: 8.6356
Epoch [2/3], Step [1111/3236], Loss: 2.3566, Perplexity: 10.5552
Epoch [2/3], Step [1112/3236], Loss: 2.0757, Perplexity: 7.9699
Epoch [2/3], Step [1113/3236], Loss: 1.9641, Perplexity: 7.1286
Epoch [2/3], Step [1114/3236], Loss: 2.1701, Perplexity: 8.7589
Epoch [2/3], Step [1115/3236], Loss: 2.1289, Perplexity: 8.4052
Epoch [2/3], Step [1116/3236], Loss: 2.0072, Perplexity: 7.4424
Epoch [2/3], Step [1117/3236], Loss: 2.4456, Perplexity: 11.5380
Epoch [2/3], Step [1118/3236], Loss: 2.1177, Perplexity: 8.3122
Epoch [2/3], Step [1119/3236], Loss: 2.1113, Perplexity: 8.2592
Epoch [2/3], Step [1120/3236], Loss: 2.3899, Perplexity: 10.9122
Epoch [2/3], Step [1121/3236], Loss: 2.0613, Perplexity: 7.8560
Epoch [2/3], Step [1122/3236], Loss: 2.0175, Perplexity: 7.5196
Epoch [2/3], Step [1123/3236], Loss: 2.5564, Perplexity: 12.8888
Epoch [2/3], Step [1124/3236], Loss: 2.2775, Perplexity: 9.7525
Epoch [2/3], Step [1125/3236], Loss: 2.1230, Perplexity: 8.3564
Epoch [2/3], Step [1126/3236], Loss: 2.1333, Perplexity: 8.4430
Epoch [2/3], Step [1127/3236], Loss: 2.1626, Perplexity: 8.6933
Epoch [2/3], Step [1128/3236], Loss: 1.9845, Perplexity: 7.2754
Epoch [2/3], Step [1129/3236], Loss: 2.0391, Perplexity: 7.6834
Epoch [2/3], Step [1130/3236], Loss: 2.0723, Perplexity: 7.9430
Epoch [2/3], Step [1131/3236], Loss: 1.9672, Perplexity: 7.1503
Epoch [2/3], Step [1132/3236], Loss: 2.0815, Perplexity: 8.0163
Epoch [2/3], Step [1133/3236], Loss: 2.1283, Perplexity: 8.4008
Epoch [2/3], Step [1134/3236], Loss: 2.1049, Perplexity: 8.2064
Epoch [2/3], Step [1135/3236], Loss: 2.0287, Perplexity: 7.6038
Epoch [2/3], Step [1136/3236], Loss: 2.0748, Perplexity: 7.9632
Epoch [2/3], Step [1137/3236], Loss: 2.0468, Perplexity: 7.7431
Epoch [2/3], Step [1138/3236], Loss: 2.1480, Perplexity: 8.5677
Epoch [2/3], Step [1139/3236], Loss: 2.0857, Perplexity: 8.0499
Epoch [2/3], Step [1140/3236], Loss: 2.0774, Perplexity: 7.9839
Epoch [2/3], Step [1141/3236], Loss: 2.2152, Perplexity: 9.1630
Epoch [2/3], Step [1142/3236], Loss: 2.3639, Perplexity: 10.6326
Epoch [2/3], Step [1143/3236], Loss: 2.6040, Perplexity: 13.5183
Epoch [2/3], Step [1144/3236], Loss: 2.0966, Perplexity: 8.1387
Epoch [2/3], Step [1145/3236], Loss: 2.1687, Perplexity: 8.7473
Epoch [2/3], Step [1146/3236], Loss: 2.2027, Perplexity: 9.0491
Epoch [2/3], Step [1147/3236], Loss: 2.0165, Perplexity: 7.5122
Epoch [2/3], Step [1148/3236], Loss: 2.0556, Perplexity: 7.8113
Epoch [2/3], Step [1149/3236], Loss: 2.1327, Perplexity: 8.4375
Epoch [2/3], Step [1150/3236], Loss: 2.3810, Perplexity: 10.8160
Epoch [2/3], Step [1151/3236], Loss: 2.0297, Perplexity: 7.6117
Epoch [2/3], Step [1152/3236], Loss: 2.1817, Perplexity: 8.8609
Epoch [2/3], Step [1153/3236], Loss: 2.0332, Perplexity: 7.6381
Epoch [2/3], Step [1154/3236], Loss: 1.8149, Perplexity: 6.1407
Epoch [2/3], Step [1155/3236], Loss: 2.0150, Perplexity: 7.5006
Epoch [2/3], Step [1156/3236], Loss: 2.2023, Perplexity: 9.0455
Epoch [2/3], Step [1157/3236], Loss: 2.0181, Perplexity: 7.5238
Epoch [2/3], Step [1158/3236], Loss: 2.0750, Perplexity: 7.9648
Epoch [2/3], Step [1159/3236], Loss: 2.0740, Perplexity: 7.9570
Epoch [2/3], Step [1160/3236], Loss: 2.0456, Perplexity: 7.7341
Epoch [2/3], Step [1161/3236], Loss: 1.9889, Perplexity: 7.3077
Epoch [2/3], Step [1162/3236], Loss: 1.9469, Perplexity: 7.0068
Epoch [2/3], Step [1163/3236], Loss: 2.1507, Perplexity: 8.5907
Epoch [2/3], Step [1164/3236], Loss: 2.1738, Perplexity: 8.7921
Epoch [2/3], Step [1165/3236], Loss: 2.1662, Perplexity: 8.7253
Epoch [2/3], Step [1166/3236], Loss: 2.0733, Perplexity: 7.9511
Epoch [2/3], Step [1167/3236], Loss: 2.0697, Perplexity: 7.9222
Epoch [2/3], Step [1168/3236], Loss: 2.0913, Perplexity: 8.0952
Epoch [2/3], Step [1169/3236], Loss: 2.1577, Perplexity: 8.6513
Epoch [2/3], Step [1170/3236], Loss: 2.0198, Perplexity: 7.5365
Epoch [2/3], Step [1171/3236], Loss: 2.0128, Perplexity: 7.4841
Epoch [2/3], Step [1172/3236], Loss: 2.0049, Perplexity: 7.4251
Epoch [2/3], Step [1173/3236], Loss: 2.2272, Perplexity: 9.2738
Epoch [2/3], Step [1174/3236], Loss: 2.1461, Perplexity: 8.5517
Epoch [2/3], Step [1175/3236], Loss: 2.0439, Perplexity: 7.7204
Epoch [2/3], Step [1176/3236], Loss: 1.9182, Perplexity: 6.8089
Epoch [2/3], Step [1177/3236], Loss: 2.0315, Perplexity: 7.6252
Epoch [2/3], Step [1178/3236], Loss: 2.1300, Perplexity: 8.4146
Epoch [2/3], Step [1179/3236], Loss: 2.0132, Perplexity: 7.4872
Epoch [2/3], Step [1180/3236], Loss: 2.0102, Perplexity: 7.4649
Epoch [2/3], Step [1181/3236], Loss: 2.1583, Perplexity: 8.6567
Epoch [2/3], Step [1182/3236], Loss: 2.0007, Perplexity: 7.3941
Epoch [2/3], Step [1183/3236], Loss: 2.1862, Perplexity: 8.9016
Epoch [2/3], Step [1184/3236], Loss: 2.0349, Perplexity: 7.6511
Epoch [2/3], Step [1185/3236], Loss: 2.4437, Perplexity: 11.5160
Epoch [2/3], Step [1186/3236], Loss: 2.0439, Perplexity: 7.7210
Epoch [2/3], Step [1187/3236], Loss: 2.1909, Perplexity: 8.9432
Epoch [2/3], Step [1188/3236], Loss: 1.8599, Perplexity: 6.4232
Epoch [2/3], Step [1189/3236], Loss: 1.9732, Perplexity: 7.1939
Epoch [2/3], Step [1190/3236], Loss: 2.4061, Perplexity: 11.0911
Epoch [2/3], Step [1191/3236], Loss: 3.0461, Perplexity: 21.0325
Epoch [2/3], Step [1192/3236], Loss: 2.1309, Perplexity: 8.4221
Epoch [2/3], Step [1193/3236], Loss: 2.1773, Perplexity: 8.8229
Epoch [2/3], Step [1194/3236], Loss: 2.3674, Perplexity: 10.6694
Epoch [2/3], Step [1195/3236], Loss: 2.3702, Perplexity: 10.6999
Epoch [2/3], Step [1196/3236], Loss: 2.2593, Perplexity: 9.5763
Epoch [2/3], Step [1197/3236], Loss: 1.9795, Perplexity: 7.2389
Epoch [2/3], Step [1198/3236], Loss: 2.0348, Perplexity: 7.6508
Epoch [2/3], Step [1199/3236], Loss: 2.8618, Perplexity: 17.4922
Epoch [2/3], Step [1200/3236], Loss: 2.0722, Perplexity: 7.9419
Epoch [2/3], Step [1201/3236], Loss: 2.0898, Perplexity: 8.0835
Epoch [2/3], Step [1202/3236], Loss: 2.3738, Perplexity: 10.7376
Epoch [2/3], Step [1203/3236], Loss: 2.1347, Perplexity: 8.4543
Epoch [2/3], Step [1204/3236], Loss: 2.0730, Perplexity: 7.9489
Epoch [2/3], Step [1205/3236], Loss: 2.0996, Perplexity: 8.1625
Epoch [2/3], Step [1206/3236], Loss: 2.2575, Perplexity: 9.5593
Epoch [2/3], Step [1207/3236], Loss: 2.3744, Perplexity: 10.7446
Epoch [2/3], Step [1208/3236], Loss: 2.0639, Perplexity: 7.8767
Epoch [2/3], Step [1209/3236], Loss: 2.0056, Perplexity: 7.4309
Epoch [2/3], Step [1210/3236], Loss: 2.1328, Perplexity: 8.4381
Epoch [2/3], Step [1211/3236], Loss: 2.0910, Perplexity: 8.0933
Epoch [2/3], Step [1212/3236], Loss: 2.1861, Perplexity: 8.9004
Epoch [2/3], Step [1213/3236], Loss: 2.0294, Perplexity: 7.6094
Epoch [2/3], Step [1214/3236], Loss: 2.2605, Perplexity: 9.5876
Epoch [2/3], Step [1215/3236], Loss: 2.3116, Perplexity: 10.0910
Epoch [2/3], Step [1216/3236], Loss: 2.2354, Perplexity: 9.3501
Epoch [2/3], Step [1217/3236], Loss: 2.1200, Perplexity: 8.3309
Epoch [2/3], Step [1218/3236], Loss: 1.8983, Perplexity: 6.6748
Epoch [2/3], Step [1219/3236], Loss: 2.0310, Perplexity: 7.6216
Epoch [2/3], Step [1220/3236], Loss: 2.0945, Perplexity: 8.1210
Epoch [2/3], Step [1221/3236], Loss: 2.2589, Perplexity: 9.5727
Epoch [2/3], Step [1222/3236], Loss: 1.9526, Perplexity: 7.0471
Epoch [2/3], Step [1223/3236], Loss: 2.3290, Perplexity: 10.2680
Epoch [2/3], Step [1224/3236], Loss: 1.8805, Perplexity: 6.5568
Epoch [2/3], Step [1225/3236], Loss: 2.0594, Perplexity: 7.8414
Epoch [2/3], Step [1226/3236], Loss: 1.9439, Perplexity: 6.9862
Epoch [2/3], Step [1227/3236], Loss: 2.3554, Perplexity: 10.5419
Epoch [2/3], Step [1228/3236], Loss: 2.0922, Perplexity: 8.1025
Epoch [2/3], Step [1229/3236], Loss: 1.9326, Perplexity: 6.9073
Epoch [2/3], Step [1230/3236], Loss: 2.0099, Perplexity: 7.4624
Epoch [2/3], Step [1231/3236], Loss: 2.1010, Perplexity: 8.1747
Epoch [2/3], Step [1232/3236], Loss: 2.4939, Perplexity: 12.1088
Epoch [2/3], Step [1233/3236], Loss: 2.0633, Perplexity: 7.8721
Epoch [2/3], Step [1234/3236], Loss: 2.2799, Perplexity: 9.7758
Epoch [2/3], Step [1235/3236], Loss: 2.0891, Perplexity: 8.0777
Epoch [2/3], Step [1236/3236], Loss: 2.2846, Perplexity: 9.8222
Epoch [2/3], Step [1237/3236], Loss: 2.4484, Perplexity: 11.5704
Epoch [2/3], Step [1238/3236], Loss: 2.2382, Perplexity: 9.3767
Epoch [2/3], Step [1239/3236], Loss: 2.2139, Perplexity: 9.1509
Epoch [2/3], Step [1240/3236], Loss: 2.2663, Perplexity: 9.6436
Epoch [2/3], Step [1241/3236], Loss: 2.0064, Perplexity: 7.4366
Epoch [2/3], Step [1242/3236], Loss: 2.0652, Perplexity: 7.8868
Epoch [2/3], Step [1243/3236], Loss: 2.3438, Perplexity: 10.4211
Epoch [2/3], Step [1244/3236], Loss: 2.0917, Perplexity: 8.0987
Epoch [2/3], Step [1245/3236], Loss: 2.3703, Perplexity: 10.7005
Epoch [2/3], Step [1246/3236], Loss: 2.1143, Perplexity: 8.2834
Epoch [2/3], Step [1247/3236], Loss: 2.0200, Perplexity: 7.5387
Epoch [2/3], Step [1248/3236], Loss: 2.0260, Perplexity: 7.5838
Epoch [2/3], Step [1249/3236], Loss: 1.9899, Perplexity: 7.3151
Epoch [2/3], Step [1250/3236], Loss: 2.0079, Perplexity: 7.4480
Epoch [2/3], Step [1251/3236], Loss: 2.2193, Perplexity: 9.2007
Epoch [2/3], Step [1252/3236], Loss: 2.0204, Perplexity: 7.5413
Epoch [2/3], Step [1253/3236], Loss: 2.0739, Perplexity: 7.9558
Epoch [2/3], Step [1254/3236], Loss: 2.3667, Perplexity: 10.6617
Epoch [2/3], Step [1255/3236], Loss: 3.0502, Perplexity: 21.1203
Epoch [2/3], Step [1256/3236], Loss: 2.0202, Perplexity: 7.5399
Epoch [2/3], Step [1257/3236], Loss: 2.1013, Perplexity: 8.1771
Epoch [2/3], Step [1258/3236], Loss: 2.3193, Perplexity: 10.1687
Epoch [2/3], Step [1259/3236], Loss: 2.1976, Perplexity: 9.0038
Epoch [2/3], Step [1260/3236], Loss: 2.0447, Perplexity: 7.7267
Epoch [2/3], Step [1261/3236], Loss: 2.1726, Perplexity: 8.7810
Epoch [2/3], Step [1262/3236], Loss: 1.9771, Perplexity: 7.2218
Epoch [2/3], Step [1263/3236], Loss: 2.2820, Perplexity: 9.7966
Epoch [2/3], Step [1264/3236], Loss: 2.0490, Perplexity: 7.7599
Epoch [2/3], Step [1265/3236], Loss: 2.5887, Perplexity: 13.3122
Epoch [2/3], Step [1266/3236], Loss: 2.1934, Perplexity: 8.9655
Epoch [2/3], Step [1267/3236], Loss: 2.1404, Perplexity: 8.5030
Epoch [2/3], Step [1268/3236], Loss: 2.0487, Perplexity: 7.7577
Epoch [2/3], Step [1269/3236], Loss: 2.0837, Perplexity: 8.0341
Epoch [2/3], Step [1270/3236], Loss: 2.0279, Perplexity: 7.5980
Epoch [2/3], Step [1271/3236], Loss: 2.0585, Perplexity: 7.8340
Epoch [2/3], Step [1272/3236], Loss: 2.0521, Perplexity: 7.7839
Epoch [2/3], Step [1273/3236], Loss: 2.0379, Perplexity: 7.6748
Epoch [2/3], Step [1274/3236], Loss: 1.9901, Perplexity: 7.3166
Epoch [2/3], Step [1275/3236], Loss: 2.1114, Perplexity: 8.2599
Epoch [2/3], Step [1276/3236], Loss: 2.4600, Perplexity: 11.7043
Epoch [2/3], Step [1277/3236], Loss: 2.0331, Perplexity: 7.6380
Epoch [2/3], Step [1278/3236], Loss: 2.1355, Perplexity: 8.4613
Epoch [2/3], Step [1279/3236], Loss: 2.0614, Perplexity: 7.8566
Epoch [2/3], Step [1280/3236], Loss: 2.1080, Perplexity: 8.2315
Epoch [2/3], Step [1281/3236], Loss: 2.0892, Perplexity: 8.0782
Epoch [2/3], Step [1282/3236], Loss: 2.4144, Perplexity: 11.1831
Epoch [2/3], Step [1283/3236], Loss: 2.3288, Perplexity: 10.2653
Epoch [2/3], Step [1284/3236], Loss: 2.3085, Perplexity: 10.0591
Epoch [2/3], Step [1285/3236], Loss: 1.9499, Perplexity: 7.0281
Epoch [2/3], Step [1286/3236], Loss: 2.0224, Perplexity: 7.5562
Epoch [2/3], Step [1287/3236], Loss: 2.2159, Perplexity: 9.1693
Epoch [2/3], Step [1288/3236], Loss: 1.9037, Perplexity: 6.7104
Epoch [2/3], Step [1289/3236], Loss: 2.1026, Perplexity: 8.1874
Epoch [2/3], Step [1290/3236], Loss: 2.5465, Perplexity: 12.7621
Epoch [2/3], Step [1291/3236], Loss: 2.5752, Perplexity: 13.1346
Epoch [2/3], Step [1292/3236], Loss: 2.2018, Perplexity: 9.0415
Epoch [2/3], Step [1293/3236], Loss: 2.0293, Perplexity: 7.6090
Epoch [2/3], Step [1294/3236], Loss: 2.2263, Perplexity: 9.2660
Epoch [2/3], Step [1295/3236], Loss: 2.2030, Perplexity: 9.0524
Epoch [2/3], Step [1296/3236], Loss: 2.2831, Perplexity: 9.8068
Epoch [2/3], Step [1297/3236], Loss: 2.6617, Perplexity: 14.3199
Epoch [2/3], Step [1298/3236], Loss: 2.4560, Perplexity: 11.6579
Epoch [2/3], Step [1299/3236], Loss: 2.1878, Perplexity: 8.9159
Epoch [2/3], Step [1300/3236], Loss: 2.0992, Perplexity: 8.1599
Epoch [2/3], Step [1301/3236], Loss: 2.1211, Perplexity: 8.3406
Epoch [2/3], Step [1302/3236], Loss: 2.2236, Perplexity: 9.2402
Epoch [2/3], Step [1303/3236], Loss: 2.2391, Perplexity: 9.3846
Epoch [2/3], Step [1304/3236], Loss: 2.3679, Perplexity: 10.6753
Epoch [2/3], Step [1305/3236], Loss: 1.9436, Perplexity: 6.9840
Epoch [2/3], Step [1306/3236], Loss: 2.0284, Perplexity: 7.6022
Epoch [2/3], Step [1307/3236], Loss: 2.1583, Perplexity: 8.6567
Epoch [2/3], Step [1308/3236], Loss: 2.0286, Perplexity: 7.6034
Epoch [2/3], Step [1309/3236], Loss: 2.0967, Perplexity: 8.1397
Epoch [2/3], Step [1310/3236], Loss: 2.0156, Perplexity: 7.5055
Epoch [2/3], Step [1311/3236], Loss: 2.1065, Perplexity: 8.2198
Epoch [2/3], Step [1312/3236], Loss: 2.0499, Perplexity: 7.7669
Epoch [2/3], Step [1313/3236], Loss: 2.1719, Perplexity: 8.7747
Epoch [2/3], Step [1314/3236], Loss: 2.0748, Perplexity: 7.9633
Epoch [2/3], Step [1315/3236], Loss: 2.1717, Perplexity: 8.7731
Epoch [2/3], Step [1316/3236], Loss: 1.9893, Perplexity: 7.3101
Epoch [2/3], Step [1317/3236], Loss: 2.0569, Perplexity: 7.8217
Epoch [2/3], Step [1318/3236], Loss: 2.2185, Perplexity: 9.1936
Epoch [2/3], Step [1319/3236], Loss: 2.0458, Perplexity: 7.7355
Epoch [2/3], Step [1320/3236], Loss: 2.4188, Perplexity: 11.2329
Epoch [2/3], Step [1321/3236], Loss: 2.0084, Perplexity: 7.4511
Epoch [2/3], Step [1322/3236], Loss: 2.2444, Perplexity: 9.4343
Epoch [2/3], Step [1323/3236], Loss: 2.1134, Perplexity: 8.2764
Epoch [2/3], Step [1324/3236], Loss: 2.4927, Perplexity: 12.0942
Epoch [2/3], Step [1325/3236], Loss: 2.1300, Perplexity: 8.4148
Epoch [2/3], Step [1326/3236], Loss: 1.9613, Perplexity: 7.1083
Epoch [2/3], Step [1327/3236], Loss: 2.8015, Perplexity: 16.4687
Epoch [2/3], Step [1328/3236], Loss: 1.9892, Perplexity: 7.3093
Epoch [2/3], Step [1329/3236], Loss: 2.0119, Perplexity: 7.4777
Epoch [2/3], Step [1330/3236], Loss: 2.1104, Perplexity: 8.2515
Epoch [2/3], Step [1331/3236], Loss: 2.0430, Perplexity: 7.7138
Epoch [2/3], Step [1332/3236], Loss: 2.0597, Perplexity: 7.8439
Epoch [2/3], Step [1333/3236], Loss: 2.0008, Perplexity: 7.3952
Epoch [2/3], Step [1334/3236], Loss: 1.9810, Perplexity: 7.2500
Epoch [2/3], Step [1335/3236], Loss: 2.5356, Perplexity: 12.6244
Epoch [2/3], Step [1336/3236], Loss: 2.1018, Perplexity: 8.1810
Epoch [2/3], Step [1337/3236], Loss: 2.3347, Perplexity: 10.3263
Epoch [2/3], Step [1338/3236], Loss: 2.2505, Perplexity: 9.4923
Epoch [2/3], Step [1339/3236], Loss: 2.5006, Perplexity: 12.1902
Epoch [2/3], Step [1340/3236], Loss: 2.3223, Perplexity: 10.1993
Epoch [2/3], Step [1341/3236], Loss: 1.9926, Perplexity: 7.3345
Epoch [2/3], Step [1342/3236], Loss: 1.9784, Perplexity: 7.2309
Epoch [2/3], Step [1343/3236], Loss: 2.2190, Perplexity: 9.1983
Epoch [2/3], Step [1344/3236], Loss: 2.1673, Perplexity: 8.7346
Epoch [2/3], Step [1345/3236], Loss: 2.2346, Perplexity: 9.3428
Epoch [2/3], Step [1346/3236], Loss: 2.1494, Perplexity: 8.5801
Epoch [2/3], Step [1347/3236], Loss: 1.9184, Perplexity: 6.8100
Epoch [2/3], Step [1348/3236], Loss: 2.0180, Perplexity: 7.5231
Epoch [2/3], Step [1349/3236], Loss: 2.1032, Perplexity: 8.1927
Epoch [2/3], Step [1350/3236], Loss: 1.9882, Perplexity: 7.3024
Epoch [2/3], Step [1351/3236], Loss: 2.1468, Perplexity: 8.5578
Epoch [2/3], Step [1352/3236], Loss: 1.9129, Perplexity: 6.7729
Epoch [2/3], Step [1353/3236], Loss: 2.0545, Perplexity: 7.8033
Epoch [2/3], Step [1354/3236], Loss: 2.1149, Perplexity: 8.2890
Epoch [2/3], Step [1355/3236], Loss: 2.0495, Perplexity: 7.7637
Epoch [2/3], Step [1356/3236], Loss: 2.0766, Perplexity: 7.9774
Epoch [2/3], Step [1357/3236], Loss: 2.0074, Perplexity: 7.4437
Epoch [2/3], Step [1358/3236], Loss: 1.9630, Perplexity: 7.1207
Epoch [2/3], Step [1359/3236], Loss: 2.1120, Perplexity: 8.2651
Epoch [2/3], Step [1360/3236], Loss: 1.9693, Perplexity: 7.1656
Epoch [2/3], Step [1361/3236], Loss: 2.1615, Perplexity: 8.6844
Epoch [2/3], Step [1362/3236], Loss: 2.3015, Perplexity: 9.9889
Epoch [2/3], Step [1363/3236], Loss: 2.3948, Perplexity: 10.9655
Epoch [2/3], Step [1364/3236], Loss: 2.0181, Perplexity: 7.5237
Epoch [2/3], Step [1365/3236], Loss: 2.0682, Perplexity: 7.9108
Epoch [2/3], Step [1366/3236], Loss: 2.1065, Perplexity: 8.2197
Epoch [2/3], Step [1367/3236], Loss: 2.1401, Perplexity: 8.5007
Epoch [2/3], Step [1368/3236], Loss: 1.9257, Perplexity: 6.8599
Epoch [2/3], Step [1369/3236], Loss: 2.1919, Perplexity: 8.9518
Epoch [2/3], Step [1370/3236], Loss: 2.8347, Perplexity: 17.0246
Epoch [2/3], Step [1371/3236], Loss: 2.0818, Perplexity: 8.0190
Epoch [2/3], Step [1372/3236], Loss: 2.4073, Perplexity: 11.1036
Epoch [2/3], Step [1373/3236], Loss: 2.3573, Perplexity: 10.5621
Epoch [2/3], Step [1374/3236], Loss: 2.0842, Perplexity: 8.0381
Epoch [2/3], Step [1375/3236], Loss: 2.0793, Perplexity: 7.9988
Epoch [2/3], Step [1376/3236], Loss: 2.0409, Perplexity: 7.6978
Epoch [2/3], Step [1377/3236], Loss: 2.0076, Perplexity: 7.4452
Epoch [2/3], Step [1378/3236], Loss: 2.0802, Perplexity: 8.0058
Epoch [2/3], Step [1379/3236], Loss: 2.0406, Perplexity: 7.6954
Epoch [2/3], Step [1380/3236], Loss: 1.9904, Perplexity: 7.3183
Epoch [2/3], Step [1381/3236], Loss: 2.0263, Perplexity: 7.5859
Epoch [2/3], Step [1382/3236], Loss: 2.0851, Perplexity: 8.0455
Epoch [2/3], Step [1383/3236], Loss: 3.4284, Perplexity: 30.8275
Epoch [2/3], Step [1384/3236], Loss: 2.0567, Perplexity: 7.8200
Epoch [2/3], Step [1385/3236], Loss: 2.0016, Perplexity: 7.4012
Epoch [2/3], Step [1386/3236], Loss: 2.1724, Perplexity: 8.7797
Epoch [2/3], Step [1387/3236], Loss: 1.9481, Perplexity: 7.0155
Epoch [2/3], Step [1388/3236], Loss: 2.0320, Perplexity: 7.6296
Epoch [2/3], Step [1389/3236], Loss: 2.1230, Perplexity: 8.3564
Epoch [2/3], Step [1390/3236], Loss: 2.1838, Perplexity: 8.8804
Epoch [2/3], Step [1391/3236], Loss: 2.1971, Perplexity: 8.9988
Epoch [2/3], Step [1392/3236], Loss: 2.3673, Perplexity: 10.6685
Epoch [2/3], Step [1393/3236], Loss: 2.0390, Perplexity: 7.6833
Epoch [2/3], Step [1394/3236], Loss: 1.9852, Perplexity: 7.2807
Epoch [2/3], Step [1395/3236], Loss: 2.0595, Perplexity: 7.8419
Epoch [2/3], Step [1396/3236], Loss: 1.9477, Perplexity: 7.0128
Epoch [2/3], Step [1397/3236], Loss: 2.3783, Perplexity: 10.7862
Epoch [2/3], Step [1398/3236], Loss: 2.0461, Perplexity: 7.7379
Epoch [2/3], Step [1399/3236], Loss: 2.1205, Perplexity: 8.3353
Epoch [2/3], Step [1400/3236], Loss: 2.2872, Perplexity: 9.8477
Epoch [2/3], Step [1401/3236], Loss: 2.1460, Perplexity: 8.5506
Epoch [2/3], Step [1402/3236], Loss: 2.0193, Perplexity: 7.5329
Epoch [2/3], Step [1403/3236], Loss: 2.2506, Perplexity: 9.4936
Epoch [2/3], Step [1404/3236], Loss: 2.0818, Perplexity: 8.0189
Epoch [2/3], Step [1405/3236], Loss: 2.7292, Perplexity: 15.3202
Epoch [2/3], Step [1406/3236], Loss: 2.6779, Perplexity: 14.5544
Epoch [2/3], Step [1407/3236], Loss: 1.9303, Perplexity: 6.8913
Epoch [2/3], Step [1408/3236], Loss: 2.0542, Perplexity: 7.8005
Epoch [2/3], Step [1409/3236], Loss: 1.9777, Perplexity: 7.2259
Epoch [2/3], Step [1410/3236], Loss: 2.2663, Perplexity: 9.6435
Epoch [2/3], Step [1411/3236], Loss: 2.2108, Perplexity: 9.1234
Epoch [2/3], Step [1412/3236], Loss: 1.9288, Perplexity: 6.8815
Epoch [2/3], Step [1413/3236], Loss: 2.1251, Perplexity: 8.3740
Epoch [2/3], Step [1414/3236], Loss: 1.9313, Perplexity: 6.8982
Epoch [2/3], Step [1415/3236], Loss: 2.2214, Perplexity: 9.2202
Epoch [2/3], Step [1416/3236], Loss: 2.0442, Perplexity: 7.7230
Epoch [2/3], Step [1417/3236], Loss: 2.7435, Perplexity: 15.5414
Epoch [2/3], Step [1418/3236], Loss: 2.3650, Perplexity: 10.6441
Epoch [2/3], Step [1419/3236], Loss: 1.9730, Perplexity: 7.1923
Epoch [2/3], Step [1420/3236], Loss: 2.0736, Perplexity: 7.9536
Epoch [2/3], Step [1421/3236], Loss: 2.1308, Perplexity: 8.4217
Epoch [2/3], Step [1422/3236], Loss: 2.2604, Perplexity: 9.5868
Epoch [2/3], Step [1423/3236], Loss: 2.1592, Perplexity: 8.6644
Epoch [2/3], Step [1424/3236], Loss: 2.0723, Perplexity: 7.9427
Epoch [2/3], Step [1425/3236], Loss: 2.3036, Perplexity: 10.0105
Epoch [2/3], Step [1426/3236], Loss: 2.0426, Perplexity: 7.7103
Epoch [2/3], Step [1427/3236], Loss: 1.9779, Perplexity: 7.2275
Epoch [2/3], Step [1428/3236], Loss: 2.0858, Perplexity: 8.0507
Epoch [2/3], Step [1429/3236], Loss: 2.0223, Perplexity: 7.5559
Epoch [2/3], Step [1430/3236], Loss: 2.0861, Perplexity: 8.0534
Epoch [2/3], Step [1431/3236], Loss: 2.2634, Perplexity: 9.6155
Epoch [2/3], Step [1432/3236], Loss: 2.1660, Perplexity: 8.7234
Epoch [2/3], Step [1433/3236], Loss: 2.0011, Perplexity: 7.3968
Epoch [2/3], Step [1434/3236], Loss: 2.4280, Perplexity: 11.3358
Epoch [2/3], Step [1435/3236], Loss: 2.0091, Perplexity: 7.4565
Epoch [2/3], Step [1436/3236], Loss: 2.0599, Perplexity: 7.8456
Epoch [2/3], Step [1437/3236], Loss: 1.9221, Perplexity: 6.8355
Epoch [2/3], Step [1438/3236], Loss: 2.2703, Perplexity: 9.6825
Epoch [2/3], Step [1439/3236], Loss: 2.1550, Perplexity: 8.6276
Epoch [2/3], Step [1440/3236], Loss: 2.1495, Perplexity: 8.5808
Epoch [2/3], Step [1441/3236], Loss: 2.1483, Perplexity: 8.5704
Epoch [2/3], Step [1442/3236], Loss: 2.0361, Perplexity: 7.6603
Epoch [2/3], Step [1443/3236], Loss: 2.0773, Perplexity: 7.9830
Epoch [2/3], Step [1444/3236], Loss: 2.0558, Perplexity: 7.8128
Epoch [2/3], Step [1445/3236], Loss: 2.1699, Perplexity: 8.7570
Epoch [2/3], Step [1446/3236], Loss: 2.0670, Perplexity: 7.9011
Epoch [2/3], Step [1447/3236], Loss: 2.2049, Perplexity: 9.0695
Epoch [2/3], Step [1448/3236], Loss: 2.1697, Perplexity: 8.7556
Epoch [2/3], Step [1449/3236], Loss: 2.1638, Perplexity: 8.7046
Epoch [2/3], Step [1450/3236], Loss: 2.8059, Perplexity: 16.5419
Epoch [2/3], Step [1451/3236], Loss: 2.1086, Perplexity: 8.2370
Epoch [2/3], Step [1452/3236], Loss: 2.1113, Perplexity: 8.2593
Epoch [2/3], Step [1453/3236], Loss: 1.9618, Perplexity: 7.1124
Epoch [2/3], Step [1454/3236], Loss: 2.1771, Perplexity: 8.8203
Epoch [2/3], Step [1455/3236], Loss: 2.3974, Perplexity: 10.9945
Epoch [2/3], Step [1456/3236], Loss: 2.0631, Perplexity: 7.8702
Epoch [2/3], Step [1457/3236], Loss: 2.0838, Perplexity: 8.0346
Epoch [2/3], Step [1458/3236], Loss: 2.1036, Perplexity: 8.1956
Epoch [2/3], Step [1459/3236], Loss: 2.0405, Perplexity: 7.6941
Epoch [2/3], Step [1460/3236], Loss: 2.2658, Perplexity: 9.6389
Epoch [2/3], Step [1461/3236], Loss: 2.1704, Perplexity: 8.7619
Epoch [2/3], Step [1462/3236], Loss: 1.9558, Perplexity: 7.0699
Epoch [2/3], Step [1463/3236], Loss: 2.0346, Perplexity: 7.6492
Epoch [2/3], Step [1464/3236], Loss: 2.2643, Perplexity: 9.6247
Epoch [2/3], Step [1465/3236], Loss: 1.9956, Perplexity: 7.3565
Epoch [2/3], Step [1466/3236], Loss: 2.4053, Perplexity: 11.0816
Epoch [2/3], Step [1467/3236], Loss: 2.0005, Perplexity: 7.3927
Epoch [2/3], Step [1468/3236], Loss: 2.0713, Perplexity: 7.9349
Epoch [2/3], Step [1469/3236], Loss: 2.0816, Perplexity: 8.0172
Epoch [2/3], Step [1470/3236], Loss: 2.2610, Perplexity: 9.5923
Epoch [2/3], Step [1471/3236], Loss: 2.1498, Perplexity: 8.5832
Epoch [2/3], Step [1472/3236], Loss: 2.1075, Perplexity: 8.2275
Epoch [2/3], Step [1473/3236], Loss: 2.0894, Perplexity: 8.0797
Epoch [2/3], Step [1474/3236], Loss: 2.1321, Perplexity: 8.4328
Epoch [2/3], Step [1475/3236], Loss: 1.9992, Perplexity: 7.3832
Epoch [2/3], Step [1476/3236], Loss: 2.1333, Perplexity: 8.4429
Epoch [2/3], Step [1477/3236], Loss: 2.2294, Perplexity: 9.2945
Epoch [2/3], Step [1478/3236], Loss: 2.2991, Perplexity: 9.9650
Epoch [2/3], Step [1479/3236], Loss: 2.0941, Perplexity: 8.1184
Epoch [2/3], Step [1480/3236], Loss: 2.4889, Perplexity: 12.0479
Epoch [2/3], Step [1481/3236], Loss: 2.2091, Perplexity: 9.1073
Epoch [2/3], Step [1482/3236], Loss: 1.8952, Perplexity: 6.6540
Epoch [2/3], Step [1483/3236], Loss: 2.0433, Perplexity: 7.7159
Epoch [2/3], Step [1484/3236], Loss: 2.2131, Perplexity: 9.1437
Epoch [2/3], Step [1485/3236], Loss: 2.0882, Perplexity: 8.0703
Epoch [2/3], Step [1486/3236], Loss: 2.0667, Perplexity: 7.8988
Epoch [2/3], Step [1487/3236], Loss: 2.6444, Perplexity: 14.0750
Epoch [2/3], Step [1488/3236], Loss: 2.0332, Perplexity: 7.6389
Epoch [2/3], Step [1489/3236], Loss: 2.2699, Perplexity: 9.6788
Epoch [2/3], Step [1490/3236], Loss: 2.0803, Perplexity: 8.0067
Epoch [2/3], Step [1491/3236], Loss: 2.0105, Perplexity: 7.4672
Epoch [2/3], Step [1492/3236], Loss: 2.2012, Perplexity: 9.0360
Epoch [2/3], Step [1493/3236], Loss: 2.2816, Perplexity: 9.7925
Epoch [2/3], Step [1494/3236], Loss: 2.0389, Perplexity: 7.6825
Epoch [2/3], Step [1495/3236], Loss: 2.3881, Perplexity: 10.8929
Epoch [2/3], Step [1496/3236], Loss: 2.2381, Perplexity: 9.3753
Epoch [2/3], Step [1497/3236], Loss: 2.0432, Perplexity: 7.7150
Epoch [2/3], Step [1498/3236], Loss: 2.0822, Perplexity: 8.0220
Epoch [2/3], Step [1499/3236], Loss: 1.9628, Perplexity: 7.1196
Epoch [2/3], Step [1500/3236], Loss: 2.0423, Perplexity: 7.7082
Epoch [2/3], Step [1501/3236], Loss: 2.0307, Perplexity: 7.6198
Epoch [2/3], Step [1502/3236], Loss: 2.0140, Perplexity: 7.4933
Epoch [2/3], Step [1503/3236], Loss: 2.0266, Perplexity: 7.5884
Epoch [2/3], Step [1504/3236], Loss: 2.0034, Perplexity: 7.4139
Epoch [2/3], Step [1505/3236], Loss: 2.0188, Perplexity: 7.5294
Epoch [2/3], Step [1506/3236], Loss: 2.0505, Perplexity: 7.7720
Epoch [2/3], Step [1507/3236], Loss: 2.0982, Perplexity: 8.1518
Epoch [2/3], Step [1508/3236], Loss: 2.0086, Perplexity: 7.4529
Epoch [2/3], Step [1509/3236], Loss: 2.0121, Perplexity: 7.4793
Epoch [2/3], Step [1510/3236], Loss: 2.2603, Perplexity: 9.5856
Epoch [2/3], Step [1511/3236], Loss: 2.0735, Perplexity: 7.9523
Epoch [2/3], Step [1512/3236], Loss: 2.1143, Perplexity: 8.2841
Epoch [2/3], Step [1513/3236], Loss: 2.0200, Perplexity: 7.5383
Epoch [2/3], Step [1514/3236], Loss: 2.0922, Perplexity: 8.1031
Epoch [2/3], Step [1515/3236], Loss: 2.0684, Perplexity: 7.9123
Epoch [2/3], Step [1516/3236], Loss: 2.1007, Perplexity: 8.1718
Epoch [2/3], Step [1517/3236], Loss: 2.4028, Perplexity: 11.0538
Epoch [2/3], Step [1518/3236], Loss: 2.0856, Perplexity: 8.0490
Epoch [2/3], Step [1519/3236], Loss: 2.0727, Perplexity: 7.9466
Epoch [2/3], Step [1520/3236], Loss: 1.9295, Perplexity: 6.8863
Epoch [2/3], Step [1521/3236], Loss: 2.1111, Perplexity: 8.2577
Epoch [2/3], Step [1522/3236], Loss: 2.3370, Perplexity: 10.3503
Epoch [2/3], Step [1523/3236], Loss: 1.9960, Perplexity: 7.3593
Epoch [2/3], Step [1524/3236], Loss: 2.0847, Perplexity: 8.0424
Epoch [2/3], Step [1525/3236], Loss: 2.2053, Perplexity: 9.0730
Epoch [2/3], Step [1526/3236], Loss: 2.0175, Perplexity: 7.5193
Epoch [2/3], Step [1527/3236], Loss: 1.9662, Perplexity: 7.1432
Epoch [2/3], Step [1528/3236], Loss: 2.3231, Perplexity: 10.2070
Epoch [2/3], Step [1529/3236], Loss: 2.0012, Perplexity: 7.3982
Epoch [2/3], Step [1530/3236], Loss: 2.1543, Perplexity: 8.6216
Epoch [2/3], Step [1531/3236], Loss: 2.7997, Perplexity: 16.4402
Epoch [2/3], Step [1532/3236], Loss: 2.0779, Perplexity: 7.9877
Epoch [2/3], Step [1533/3236], Loss: 2.0614, Perplexity: 7.8568
Epoch [2/3], Step [1534/3236], Loss: 2.0288, Perplexity: 7.6050
Epoch [2/3], Step [1535/3236], Loss: 1.9828, Perplexity: 7.2627
Epoch [2/3], Step [1536/3236], Loss: 2.1348, Perplexity: 8.4550
Epoch [2/3], Step [1537/3236], Loss: 3.5438, Perplexity: 34.5970
Epoch [2/3], Step [1538/3236], Loss: 2.0819, Perplexity: 8.0199
Epoch [2/3], Step [1539/3236], Loss: 2.1300, Perplexity: 8.4151
Epoch [2/3], Step [1540/3236], Loss: 2.2287, Perplexity: 9.2881
Epoch [2/3], Step [1541/3236], Loss: 2.0338, Perplexity: 7.6429
Epoch [2/3], Step [1542/3236], Loss: 2.1587, Perplexity: 8.6596
Epoch [2/3], Step [1543/3236], Loss: 1.8872, Perplexity: 6.6011
Epoch [2/3], Step [1544/3236], Loss: 2.2170, Perplexity: 9.1799
Epoch [2/3], Step [1545/3236], Loss: 2.1058, Perplexity: 8.2140
Epoch [2/3], Step [1546/3236], Loss: 2.9938, Perplexity: 19.9623
Epoch [2/3], Step [1547/3236], Loss: 1.9720, Perplexity: 7.1847
Epoch [2/3], Step [1548/3236], Loss: 2.1442, Perplexity: 8.5351
Epoch [2/3], Step [1549/3236], Loss: 2.1500, Perplexity: 8.5850
Epoch [2/3], Step [1550/3236], Loss: 2.5137, Perplexity: 12.3508
Epoch [2/3], Step [1551/3236], Loss: 2.9654, Perplexity: 19.4034
Epoch [2/3], Step [1552/3236], Loss: 2.1159, Perplexity: 8.2973
Epoch [2/3], Step [1553/3236], Loss: 2.0826, Perplexity: 8.0255
Epoch [2/3], Step [1554/3236], Loss: 2.1704, Perplexity: 8.7614
Epoch [2/3], Step [1555/3236], Loss: 2.0089, Perplexity: 7.4555
Epoch [2/3], Step [1556/3236], Loss: 2.1018, Perplexity: 8.1809
Epoch [2/3], Step [1557/3236], Loss: 2.0465, Perplexity: 7.7411
Epoch [2/3], Step [1558/3236], Loss: 2.0672, Perplexity: 7.9024
Epoch [2/3], Step [1559/3236], Loss: 2.1252, Perplexity: 8.3749
Epoch [2/3], Step [1560/3236], Loss: 2.1573, Perplexity: 8.6479
Epoch [2/3], Step [1561/3236], Loss: 2.0175, Perplexity: 7.5194
Epoch [2/3], Step [1562/3236], Loss: 1.9804, Perplexity: 7.2457
Epoch [2/3], Step [1563/3236], Loss: 2.2191, Perplexity: 9.1987
Epoch [2/3], Step [1564/3236], Loss: 1.9801, Perplexity: 7.2433
Epoch [2/3], Step [1565/3236], Loss: 2.0986, Perplexity: 8.1546
Epoch [2/3], Step [1566/3236], Loss: 1.9658, Perplexity: 7.1404
Epoch [2/3], Step [1567/3236], Loss: 2.1203, Perplexity: 8.3334
Epoch [2/3], Step [1568/3236], Loss: 2.1058, Perplexity: 8.2141
Epoch [2/3], Step [1569/3236], Loss: 2.0650, Perplexity: 7.8855
Epoch [2/3], Step [1570/3236], Loss: 2.3884, Perplexity: 10.8956
Epoch [2/3], Step [1571/3236], Loss: 2.3558, Perplexity: 10.5462
Epoch [2/3], Step [1572/3236], Loss: 1.9104, Perplexity: 6.7560
Epoch [2/3], Step [1573/3236], Loss: 2.1964, Perplexity: 8.9929
Epoch [2/3], Step [1574/3236], Loss: 2.5399, Perplexity: 12.6789
Epoch [2/3], Step [1575/3236], Loss: 2.0542, Perplexity: 7.8009
Epoch [2/3], Step [1576/3236], Loss: 2.1262, Perplexity: 8.3833
Epoch [2/3], Step [1577/3236], Loss: 1.9907, Perplexity: 7.3204
Epoch [2/3], Step [1578/3236], Loss: 2.0288, Perplexity: 7.6052
Epoch [2/3], Step [1579/3236], Loss: 2.1781, Perplexity: 8.8292
Epoch [2/3], Step [1580/3236], Loss: 2.2036, Perplexity: 9.0579
Epoch [2/3], Step [1581/3236], Loss: 2.0405, Perplexity: 7.6944
Epoch [2/3], Step [1582/3236], Loss: 2.0786, Perplexity: 7.9937
Epoch [2/3], Step [1583/3236], Loss: 2.4051, Perplexity: 11.0797
Epoch [2/3], Step [1584/3236], Loss: 2.0464, Perplexity: 7.7401
Epoch [2/3], Step [1585/3236], Loss: 2.0392, Perplexity: 7.6846
Epoch [2/3], Step [1586/3236], Loss: 2.1726, Perplexity: 8.7807
Epoch [2/3], Step [1587/3236], Loss: 2.5773, Perplexity: 13.1622
Epoch [2/3], Step [1588/3236], Loss: 2.2000, Perplexity: 9.0253
Epoch [2/3], Step [1589/3236], Loss: 2.0103, Perplexity: 7.4652
Epoch [2/3], Step [1590/3236], Loss: 2.0659, Perplexity: 7.8926
Epoch [2/3], Step [1591/3236], Loss: 1.9302, Perplexity: 6.8910
Epoch [2/3], Step [1592/3236], Loss: 2.1010, Perplexity: 8.1743
Epoch [2/3], Step [1593/3236], Loss: 2.0530, Perplexity: 7.7911
Epoch [2/3], Step [1594/3236], Loss: 2.2023, Perplexity: 9.0458
Epoch [2/3], Step [1595/3236], Loss: 2.0326, Perplexity: 7.6342
Epoch [2/3], Step [1596/3236], Loss: 2.1073, Perplexity: 8.2263
Epoch [2/3], Step [1597/3236], Loss: 1.9402, Perplexity: 6.9602
Epoch [2/3], Step [1598/3236], Loss: 2.0069, Perplexity: 7.4401
Epoch [2/3], Step [1599/3236], Loss: 2.3718, Perplexity: 10.7167
Epoch [2/3], Step [1600/3236], Loss: 2.1831, Perplexity: 8.8739
Epoch [2/3], Step [1601/3236], Loss: 1.9873, Perplexity: 7.2959
Epoch [2/3], Step [1602/3236], Loss: 1.9724, Perplexity: 7.1878
Epoch [2/3], Step [1603/3236], Loss: 1.9386, Perplexity: 6.9489
Epoch [2/3], Step [1604/3236], Loss: 1.8757, Perplexity: 6.5255
Epoch [2/3], Step [1605/3236], Loss: 1.9387, Perplexity: 6.9499
Epoch [2/3], Step [1606/3236], Loss: 2.2532, Perplexity: 9.5180
Epoch [2/3], Step [1607/3236], Loss: 1.9812, Perplexity: 7.2512
Epoch [2/3], Step [1608/3236], Loss: 2.1408, Perplexity: 8.5066
Epoch [2/3], Step [1609/3236], Loss: 1.9988, Perplexity: 7.3800
Epoch [2/3], Step [1610/3236], Loss: 2.2516, Perplexity: 9.5025
Epoch [2/3], Step [1611/3236], Loss: 2.0282, Perplexity: 7.6004
Epoch [2/3], Step [1612/3236], Loss: 2.1355, Perplexity: 8.4616
Epoch [2/3], Step [1613/3236], Loss: 2.1766, Perplexity: 8.8163
Epoch [2/3], Step [1614/3236], Loss: 1.9235, Perplexity: 6.8447
Epoch [2/3], Step [1615/3236], Loss: 2.1807, Perplexity: 8.8521
Epoch [2/3], Step [1616/3236], Loss: 2.0896, Perplexity: 8.0819
Epoch [2/3], Step [1617/3236], Loss: 2.0346, Perplexity: 7.6489
Epoch [2/3], Step [1618/3236], Loss: 2.0348, Perplexity: 7.6505
Epoch [2/3], Step [1619/3236], Loss: 1.9888, Perplexity: 7.3070
Epoch [2/3], Step [1620/3236], Loss: 2.2249, Perplexity: 9.2522
Epoch [2/3], Step [1621/3236], Loss: 1.9808, Perplexity: 7.2483
Epoch [2/3], Step [1622/3236], Loss: 2.1376, Perplexity: 8.4793
Epoch [2/3], Step [1623/3236], Loss: 2.1251, Perplexity: 8.3736
Epoch [2/3], Step [1624/3236], Loss: 2.0957, Perplexity: 8.1308
Epoch [2/3], Step [1625/3236], Loss: 2.0779, Perplexity: 7.9877
Epoch [2/3], Step [1626/3236], Loss: 2.5308, Perplexity: 12.5634
Epoch [2/3], Step [1627/3236], Loss: 2.2399, Perplexity: 9.3928
Epoch [2/3], Step [1628/3236], Loss: 2.1006, Perplexity: 8.1712
Epoch [2/3], Step [1629/3236], Loss: 2.2702, Perplexity: 9.6814
Epoch [2/3], Step [1630/3236], Loss: 2.0824, Perplexity: 8.0241
Epoch [2/3], Step [1631/3236], Loss: 1.9524, Perplexity: 7.0459
Epoch [2/3], Step [1632/3236], Loss: 2.7341, Perplexity: 15.3953
Epoch [2/3], Step [1633/3236], Loss: 2.1619, Perplexity: 8.6879
Epoch [2/3], Step [1634/3236], Loss: 2.3250, Perplexity: 10.2269
Epoch [2/3], Step [1635/3236], Loss: 2.0480, Perplexity: 7.7527
Epoch [2/3], Step [1636/3236], Loss: 2.8843, Perplexity: 17.8903
Epoch [2/3], Step [1637/3236], Loss: 2.0302, Perplexity: 7.6158
Epoch [2/3], Step [1638/3236], Loss: 2.1288, Perplexity: 8.4050
Epoch [2/3], Step [1639/3236], Loss: 2.0447, Perplexity: 7.7271
Epoch [2/3], Step [1640/3236], Loss: 2.6144, Perplexity: 13.6588
Epoch [2/3], Step [1641/3236], Loss: 2.0076, Perplexity: 7.4454
Epoch [2/3], Step [1642/3236], Loss: 2.1940, Perplexity: 8.9711
Epoch [2/3], Step [1643/3236], Loss: 1.9241, Perplexity: 6.8492
Epoch [2/3], Step [1644/3236], Loss: 2.3133, Perplexity: 10.1078
Epoch [2/3], Step [1645/3236], Loss: 2.0726, Perplexity: 7.9458
Epoch [2/3], Step [1646/3236], Loss: 2.2034, Perplexity: 9.0555
Epoch [2/3], Step [1647/3236], Loss: 2.0937, Perplexity: 8.1150
Epoch [2/3], Step [1648/3236], Loss: 2.1198, Perplexity: 8.3295
Epoch [2/3], Step [1649/3236], Loss: 2.0664, Perplexity: 7.8965
Epoch [2/3], Step [1650/3236], Loss: 2.1920, Perplexity: 8.9527
Epoch [2/3], Step [1651/3236], Loss: 2.1329, Perplexity: 8.4393
Epoch [2/3], Step [1652/3236], Loss: 2.1840, Perplexity: 8.8818
Epoch [2/3], Step [1653/3236], Loss: 2.0781, Perplexity: 7.9890
Epoch [2/3], Step [1654/3236], Loss: 1.9645, Perplexity: 7.1314
Epoch [2/3], Step [1655/3236], Loss: 2.2893, Perplexity: 9.8684
Epoch [2/3], Step [1656/3236], Loss: 2.2435, Perplexity: 9.4260
Epoch [2/3], Step [1657/3236], Loss: 2.0193, Perplexity: 7.5333
Epoch [2/3], Step [1658/3236], Loss: 2.0714, Perplexity: 7.9363
Epoch [2/3], Step [1659/3236], Loss: 2.1458, Perplexity: 8.5492
Epoch [2/3], Step [1660/3236], Loss: 2.0544, Perplexity: 7.8018
Epoch [2/3], Step [1661/3236], Loss: 2.4593, Perplexity: 11.6961
Epoch [2/3], Step [1662/3236], Loss: 2.0394, Perplexity: 7.6860
Epoch [2/3], Step [1663/3236], Loss: 1.9284, Perplexity: 6.8783
Epoch [2/3], Step [1664/3236], Loss: 2.0725, Perplexity: 7.9450
Epoch [2/3], Step [1665/3236], Loss: 2.0501, Perplexity: 7.7684
Epoch [2/3], Step [1666/3236], Loss: 2.1345, Perplexity: 8.4532
Epoch [2/3], Step [1667/3236], Loss: 1.9387, Perplexity: 6.9500
Epoch [2/3], Step [1668/3236], Loss: 2.0367, Perplexity: 7.6656
Epoch [2/3], Step [1669/3236], Loss: 2.2514, Perplexity: 9.5010
Epoch [2/3], Step [1670/3236], Loss: 2.1609, Perplexity: 8.6787
Epoch [2/3], Step [1671/3236], Loss: 2.0958, Perplexity: 8.1315
Epoch [2/3], Step [1672/3236], Loss: 2.5620, Perplexity: 12.9619
Epoch [2/3], Step [1673/3236], Loss: 2.1725, Perplexity: 8.7803
Epoch [2/3], Step [1674/3236], Loss: 1.9961, Perplexity: 7.3599
Epoch [2/3], Step [1675/3236], Loss: 2.0876, Perplexity: 8.0657
Epoch [2/3], Step [1676/3236], Loss: 2.3031, Perplexity: 10.0052
Epoch [2/3], Step [1677/3236], Loss: 2.0994, Perplexity: 8.1612
Epoch [2/3], Step [1678/3236], Loss: 2.0136, Perplexity: 7.4903
Epoch [2/3], Step [1679/3236], Loss: 2.0429, Perplexity: 7.7128
Epoch [2/3], Step [1680/3236], Loss: 2.1988, Perplexity: 9.0143
Epoch [2/3], Step [1681/3236], Loss: 1.9868, Perplexity: 7.2918
Epoch [2/3], Step [1682/3236], Loss: 1.9298, Perplexity: 6.8884
Epoch [2/3], Step [1683/3236], Loss: 2.0431, Perplexity: 7.7145
Epoch [2/3], Step [1684/3236], Loss: 2.0683, Perplexity: 7.9116
Epoch [2/3], Step [1685/3236], Loss: 2.0711, Perplexity: 7.9336
Epoch [2/3], Step [1686/3236], Loss: 2.1692, Perplexity: 8.7515
Epoch [2/3], Step [1687/3236], Loss: 1.9525, Perplexity: 7.0463
Epoch [2/3], Step [1688/3236], Loss: 2.0939, Perplexity: 8.1165
Epoch [2/3], Step [1689/3236], Loss: 2.1593, Perplexity: 8.6650
Epoch [2/3], Step [1690/3236], Loss: 2.3489, Perplexity: 10.4737
Epoch [2/3], Step [1691/3236], Loss: 2.0230, Perplexity: 7.5608
Epoch [2/3], Step [1692/3236], Loss: 1.9562, Perplexity: 7.0726
Epoch [2/3], Step [1693/3236], Loss: 1.9913, Perplexity: 7.3254
Epoch [2/3], Step [1694/3236], Loss: 2.1237, Perplexity: 8.3622
Epoch [2/3], Step [1695/3236], Loss: 1.8054, Perplexity: 6.0822
Epoch [2/3], Step [1696/3236], Loss: 2.5252, Perplexity: 12.4939
Epoch [2/3], Step [1697/3236], Loss: 1.9356, Perplexity: 6.9283
Epoch [2/3], Step [1698/3236], Loss: 2.0591, Perplexity: 7.8389
Epoch [2/3], Step [1699/3236], Loss: 2.4227, Perplexity: 11.2763
Epoch [2/3], Step [1700/3236], Loss: 2.0211, Perplexity: 7.5463
Epoch [2/3], Step [1701/3236], Loss: 2.0924, Perplexity: 8.1045
Epoch [2/3], Step [1702/3236], Loss: 1.9275, Perplexity: 6.8726
Epoch [2/3], Step [1703/3236], Loss: 2.0868, Perplexity: 8.0589
Epoch [2/3], Step [1704/3236], Loss: 1.9511, Perplexity: 7.0366
Epoch [2/3], Step [1705/3236], Loss: 2.7449, Perplexity: 15.5629
Epoch [2/3], Step [1706/3236], Loss: 2.0202, Perplexity: 7.5397
Epoch [2/3], Step [1707/3236], Loss: 1.9526, Perplexity: 7.0467
Epoch [2/3], Step [1708/3236], Loss: 2.0524, Perplexity: 7.7866
Epoch [2/3], Step [1709/3236], Loss: 2.3159, Perplexity: 10.1343
Epoch [2/3], Step [1710/3236], Loss: 1.9634, Perplexity: 7.1234
Epoch [2/3], Step [1711/3236], Loss: 2.0715, Perplexity: 7.9371
Epoch [2/3], Step [1712/3236], Loss: 2.0212, Perplexity: 7.5471
Epoch [2/3], Step [1713/3236], Loss: 2.1338, Perplexity: 8.4473
Epoch [2/3], Step [1714/3236], Loss: 1.8596, Perplexity: 6.4209
Epoch [2/3], Step [1715/3236], Loss: 2.0339, Perplexity: 7.6440
Epoch [2/3], Step [1716/3236], Loss: 1.9248, Perplexity: 6.8539
Epoch [2/3], Step [1717/3236], Loss: 2.2841, Perplexity: 9.8167
Epoch [2/3], Step [1718/3236], Loss: 2.3362, Perplexity: 10.3417
Epoch [2/3], Step [1719/3236], Loss: 2.6621, Perplexity: 14.3263
Epoch [2/3], Step [1720/3236], Loss: 2.5494, Perplexity: 12.7997
Epoch [2/3], Step [1721/3236], Loss: 1.8821, Perplexity: 6.5674
Epoch [2/3], Step [1722/3236], Loss: 2.0014, Perplexity: 7.3997
Epoch [2/3], Step [1723/3236], Loss: 1.9965, Perplexity: 7.3632
Epoch [2/3], Step [1724/3236], Loss: 2.0516, Perplexity: 7.7802
Epoch [2/3], Step [1725/3236], Loss: 2.0616, Perplexity: 7.8587
Epoch [2/3], Step [1726/3236], Loss: 2.7072, Perplexity: 14.9869
Epoch [2/3], Step [1727/3236], Loss: 2.1108, Perplexity: 8.2551
Epoch [2/3], Step [1728/3236], Loss: 2.2058, Perplexity: 9.0776
Epoch [2/3], Step [1729/3236], Loss: 2.0825, Perplexity: 8.0247
Epoch [2/3], Step [1730/3236], Loss: 2.0126, Perplexity: 7.4828
Epoch [2/3], Step [1731/3236], Loss: 2.0680, Perplexity: 7.9088
Epoch [2/3], Step [1732/3236], Loss: 2.1162, Perplexity: 8.2998
Epoch [2/3], Step [1733/3236], Loss: 2.0465, Perplexity: 7.7409
Epoch [2/3], Step [1734/3236], Loss: 1.9969, Perplexity: 7.3661
Epoch [2/3], Step [1735/3236], Loss: 2.0444, Perplexity: 7.7244
Epoch [2/3], Step [1736/3236], Loss: 2.0077, Perplexity: 7.4461
Epoch [2/3], Step [1737/3236], Loss: 2.0417, Perplexity: 7.7033
Epoch [2/3], Step [1738/3236], Loss: 2.3990, Perplexity: 11.0123
Epoch [2/3], Step [1739/3236], Loss: 2.1012, Perplexity: 8.1760
Epoch [2/3], Step [1740/3236], Loss: 2.3641, Perplexity: 10.6345
Epoch [2/3], Step [1741/3236], Loss: 2.0744, Perplexity: 7.9601
Epoch [2/3], Step [1742/3236], Loss: 2.0322, Perplexity: 7.6308
Epoch [2/3], Step [1743/3236], Loss: 2.0821, Perplexity: 8.0209
Epoch [2/3], Step [1744/3236], Loss: 1.9701, Perplexity: 7.1717
Epoch [2/3], Step [1745/3236], Loss: 2.0963, Perplexity: 8.1358
Epoch [2/3], Step [1746/3236], Loss: 2.1205, Perplexity: 8.3351
Epoch [2/3], Step [1747/3236], Loss: 2.1744, Perplexity: 8.7968
Epoch [2/3], Step [1748/3236], Loss: 2.0481, Perplexity: 7.7534
Epoch [2/3], Step [1749/3236], Loss: 2.0314, Perplexity: 7.6245
Epoch [2/3], Step [1750/3236], Loss: 2.0164, Perplexity: 7.5110
Epoch [2/3], Step [1751/3236], Loss: 2.0819, Perplexity: 8.0195
Epoch [2/3], Step [1752/3236], Loss: 2.0132, Perplexity: 7.4875
Epoch [2/3], Step [1753/3236], Loss: 2.3268, Perplexity: 10.2448
Epoch [2/3], Step [1754/3236], Loss: 1.9580, Perplexity: 7.0851
Epoch [2/3], Step [1755/3236], Loss: 2.1528, Perplexity: 8.6091
Epoch [2/3], Step [1756/3236], Loss: 2.1089, Perplexity: 8.2391
Epoch [2/3], Step [1757/3236], Loss: 1.8974, Perplexity: 6.6684
Epoch [2/3], Step [1758/3236], Loss: 2.0084, Perplexity: 7.4517
Epoch [2/3], Step [1759/3236], Loss: 2.2186, Perplexity: 9.1942
Epoch [2/3], Step [1760/3236], Loss: 2.0899, Perplexity: 8.0841
Epoch [2/3], Step [1761/3236], Loss: 1.9822, Perplexity: 7.2588
Epoch [2/3], Step [1762/3236], Loss: 1.9664, Perplexity: 7.1448
Epoch [2/3], Step [1763/3236], Loss: 2.0525, Perplexity: 7.7872
Epoch [2/3], Step [1764/3236], Loss: 2.0696, Perplexity: 7.9213
Epoch [2/3], Step [1765/3236], Loss: 1.9243, Perplexity: 6.8501
Epoch [2/3], Step [1766/3236], Loss: 2.2757, Perplexity: 9.7348
Epoch [2/3], Step [1767/3236], Loss: 1.9265, Perplexity: 6.8655
Epoch [2/3], Step [1768/3236], Loss: 2.3167, Perplexity: 10.1419
Epoch [2/3], Step [1769/3236], Loss: 1.9121, Perplexity: 6.7673
Epoch [2/3], Step [1770/3236], Loss: 2.1154, Perplexity: 8.2931
Epoch [2/3], Step [1771/3236], Loss: 2.1181, Perplexity: 8.3152
Epoch [2/3], Step [1772/3236], Loss: 1.9169, Perplexity: 6.7999
Epoch [2/3], Step [1773/3236], Loss: 1.9624, Perplexity: 7.1161
Epoch [2/3], Step [1774/3236], Loss: 2.0989, Perplexity: 8.1572
Epoch [2/3], Step [1775/3236], Loss: 2.0156, Perplexity: 7.5054
Epoch [2/3], Step [1776/3236], Loss: 2.0886, Perplexity: 8.0736
Epoch [2/3], Step [1777/3236], Loss: 2.3274, Perplexity: 10.2508
Epoch [2/3], Step [1778/3236], Loss: 2.0655, Perplexity: 7.8891
Epoch [2/3], Step [1779/3236], Loss: 1.8692, Perplexity: 6.4829
Epoch [2/3], Step [1780/3236], Loss: 2.1385, Perplexity: 8.4866
Epoch [2/3], Step [1781/3236], Loss: 2.0255, Perplexity: 7.5795
Epoch [2/3], Step [1782/3236], Loss: 2.7095, Perplexity: 15.0225
Epoch [2/3], Step [1783/3236], Loss: 1.9859, Perplexity: 7.2854
Epoch [2/3], Step [1784/3236], Loss: 1.9857, Perplexity: 7.2840
Epoch [2/3], Step [1785/3236], Loss: 2.3210, Perplexity: 10.1855
Epoch [2/3], Step [1786/3236], Loss: 1.9877, Perplexity: 7.2990
Epoch [2/3], Step [1787/3236], Loss: 1.9344, Perplexity: 6.9197
Epoch [2/3], Step [1788/3236], Loss: 2.1168, Perplexity: 8.3044
Epoch [2/3], Step [1789/3236], Loss: 2.4049, Perplexity: 11.0777
Epoch [2/3], Step [1790/3236], Loss: 1.9979, Perplexity: 7.3734
Epoch [2/3], Step [1791/3236], Loss: 2.2074, Perplexity: 9.0920
Epoch [2/3], Step [1792/3236], Loss: 2.0026, Perplexity: 7.4080
Epoch [2/3], Step [1793/3236], Loss: 1.8482, Perplexity: 6.3483
Epoch [2/3], Step [1794/3236], Loss: 1.9703, Perplexity: 7.1728
Epoch [2/3], Step [1795/3236], Loss: 1.9809, Perplexity: 7.2490
Epoch [2/3], Step [1796/3236], Loss: 2.0548, Perplexity: 7.8052
Epoch [2/3], Step [1797/3236], Loss: 2.1956, Perplexity: 8.9852
Epoch [2/3], Step [1798/3236], Loss: 2.4164, Perplexity: 11.2053
Epoch [2/3], Step [1799/3236], Loss: 1.9599, Perplexity: 7.0987
Epoch [2/3], Step [1800/3236], Loss: 2.2471, Perplexity: 9.4603
Epoch [2/3], Step [1801/3236], Loss: 2.1074, Perplexity: 8.2268
Epoch [2/3], Step [1802/3236], Loss: 1.9704, Perplexity: 7.1736
Epoch [2/3], Step [1803/3236], Loss: 2.3416, Perplexity: 10.3975
Epoch [2/3], Step [1804/3236], Loss: 2.0243, Perplexity: 7.5710
Epoch [2/3], Step [1805/3236], Loss: 2.0167, Perplexity: 7.5138
Epoch [2/3], Step [1806/3236], Loss: 1.9824, Perplexity: 7.2600
Epoch [2/3], Step [1807/3236], Loss: 2.0300, Perplexity: 7.6138
Epoch [2/3], Step [1808/3236], Loss: 2.0902, Perplexity: 8.0862
Epoch [2/3], Step [1809/3236], Loss: 2.0465, Perplexity: 7.7407
Epoch [2/3], Step [1810/3236], Loss: 2.0505, Perplexity: 7.7715
Epoch [2/3], Step [1811/3236], Loss: 2.0753, Perplexity: 7.9666
Epoch [2/3], Step [1812/3236], Loss: 1.9986, Perplexity: 7.3788
Epoch [2/3], Step [1813/3236], Loss: 2.2994, Perplexity: 9.9678
Epoch [2/3], Step [1814/3236], Loss: 1.9640, Perplexity: 7.1281
Epoch [2/3], Step [1815/3236], Loss: 2.3167, Perplexity: 10.1425
Epoch [2/3], Step [1816/3236], Loss: 2.1404, Perplexity: 8.5024
Epoch [2/3], Step [1817/3236], Loss: 1.9180, Perplexity: 6.8076
Epoch [2/3], Step [1818/3236], Loss: 1.9774, Perplexity: 7.2236
Epoch [2/3], Step [1819/3236], Loss: 1.9868, Perplexity: 7.2918
Epoch [2/3], Step [1820/3236], Loss: 2.1375, Perplexity: 8.4785
Epoch [2/3], Step [1821/3236], Loss: 1.9617, Perplexity: 7.1117
Epoch [2/3], Step [1822/3236], Loss: 2.2456, Perplexity: 9.4463
Epoch [2/3], Step [1823/3236], Loss: 2.1237, Perplexity: 8.3618
Epoch [2/3], Step [1824/3236], Loss: 2.0390, Perplexity: 7.6831
Epoch [2/3], Step [1825/3236], Loss: 2.1905, Perplexity: 8.9395
Epoch [2/3], Step [1826/3236], Loss: 2.1171, Perplexity: 8.3066
Epoch [2/3], Step [1827/3236], Loss: 2.1116, Perplexity: 8.2613
Epoch [2/3], Step [1828/3236], Loss: 3.1689, Perplexity: 23.7807
Epoch [2/3], Step [1829/3236], Loss: 2.2659, Perplexity: 9.6399
Epoch [2/3], Step [1830/3236], Loss: 2.0234, Perplexity: 7.5641
Epoch [2/3], Step [1831/3236], Loss: 2.0347, Perplexity: 7.6503
Epoch [2/3], Step [1832/3236], Loss: 2.0831, Perplexity: 8.0295
Epoch [2/3], Step [1833/3236], Loss: 2.3703, Perplexity: 10.7003
Epoch [2/3], Step [1834/3236], Loss: 2.0335, Perplexity: 7.6405
Epoch [2/3], Step [1835/3236], Loss: 2.0567, Perplexity: 7.8203
Epoch [2/3], Step [1836/3236], Loss: 2.4244, Perplexity: 11.2953
Epoch [2/3], Step [1837/3236], Loss: 2.3117, Perplexity: 10.0913
Epoch [2/3], Step [1838/3236], Loss: 2.0686, Perplexity: 7.9137
Epoch [2/3], Step [1839/3236], Loss: 1.9811, Perplexity: 7.2508
Epoch [2/3], Step [1840/3236], Loss: 2.0897, Perplexity: 8.0823
Epoch [2/3], Step [1841/3236], Loss: 2.1966, Perplexity: 8.9939
Epoch [2/3], Step [1842/3236], Loss: 2.2437, Perplexity: 9.4283
Epoch [2/3], Step [1843/3236], Loss: 2.0137, Perplexity: 7.4906
Epoch [2/3], Step [1844/3236], Loss: 1.9456, Perplexity: 6.9982
Epoch [2/3], Step [1845/3236], Loss: 2.2403, Perplexity: 9.3963
Epoch [2/3], Step [1846/3236], Loss: 1.9587, Perplexity: 7.0902
Epoch [2/3], Step [1847/3236], Loss: 2.2680, Perplexity: 9.6596
Epoch [2/3], Step [1848/3236], Loss: 2.0516, Perplexity: 7.7805
Epoch [2/3], Step [1849/3236], Loss: 1.9655, Perplexity: 7.1385
Epoch [2/3], Step [1850/3236], Loss: 2.1116, Perplexity: 8.2613
Epoch [2/3], Step [1851/3236], Loss: 1.9151, Perplexity: 6.7877
Epoch [2/3], Step [1852/3236], Loss: 2.0009, Perplexity: 7.3957
Epoch [2/3], Step [1853/3236], Loss: 2.1279, Perplexity: 8.3975
Epoch [2/3], Step [1854/3236], Loss: 1.9977, Perplexity: 7.3721
Epoch [2/3], Step [1855/3236], Loss: 1.9919, Perplexity: 7.3293
Epoch [2/3], Step [1856/3236], Loss: 2.1051, Perplexity: 8.2081
Epoch [2/3], Step [1857/3236], Loss: 2.0751, Perplexity: 7.9657
Epoch [2/3], Step [1858/3236], Loss: 2.1415, Perplexity: 8.5124
Epoch [2/3], Step [1859/3236], Loss: 2.2134, Perplexity: 9.1471
Epoch [2/3], Step [1860/3236], Loss: 2.0718, Perplexity: 7.9388
Epoch [2/3], Step [1861/3236], Loss: 2.0947, Perplexity: 8.1233
Epoch [2/3], Step [1862/3236], Loss: 2.0277, Perplexity: 7.5967
Epoch [2/3], Step [1863/3236], Loss: 2.2263, Perplexity: 9.2655
Epoch [2/3], Step [1864/3236], Loss: 2.0875, Perplexity: 8.0648
Epoch [2/3], Step [1865/3236], Loss: 2.1706, Perplexity: 8.7632
Epoch [2/3], Step [1866/3236], Loss: 1.9730, Perplexity: 7.1922
Epoch [2/3], Step [1867/3236], Loss: 2.0513, Perplexity: 7.7776
Epoch [2/3], Step [1868/3236], Loss: 2.1676, Perplexity: 8.7371
Epoch [2/3], Step [1869/3236], Loss: 2.4507, Perplexity: 11.5962
Epoch [2/3], Step [1870/3236], Loss: 1.9854, Perplexity: 7.2818
Epoch [2/3], Step [1871/3236], Loss: 2.1356, Perplexity: 8.4621
Epoch [2/3], Step [1872/3236], Loss: 2.2376, Perplexity: 9.3706
Epoch [2/3], Step [1873/3236], Loss: 2.0803, Perplexity: 8.0070
Epoch [2/3], Step [1874/3236], Loss: 2.0579, Perplexity: 7.8297
Epoch [2/3], Step [1875/3236], Loss: 2.5212, Perplexity: 12.4437
Epoch [2/3], Step [1876/3236], Loss: 1.9539, Perplexity: 7.0562
Epoch [2/3], Step [1877/3236], Loss: 2.0551, Perplexity: 7.8074
Epoch [2/3], Step [1878/3236], Loss: 1.9805, Perplexity: 7.2460
Epoch [2/3], Step [1879/3236], Loss: 2.0014, Perplexity: 7.3993
Epoch [2/3], Step [1880/3236], Loss: 1.9300, Perplexity: 6.8896
Epoch [2/3], Step [1881/3236], Loss: 2.2202, Perplexity: 9.2093
Epoch [2/3], Step [1882/3236], Loss: 1.9334, Perplexity: 6.9133
Epoch [2/3], Step [1883/3236], Loss: 3.1409, Perplexity: 23.1257
Epoch [2/3], Step [1884/3236], Loss: 1.9320, Perplexity: 6.9032
Epoch [2/3], Step [1885/3236], Loss: 2.0757, Perplexity: 7.9700
Epoch [2/3], Step [1886/3236], Loss: 2.0002, Perplexity: 7.3906
Epoch [2/3], Step [1887/3236], Loss: 1.9658, Perplexity: 7.1409
Epoch [2/3], Step [1888/3236], Loss: 2.1955, Perplexity: 8.9849
Epoch [2/3], Step [1889/3236], Loss: 2.0563, Perplexity: 7.8170
Epoch [2/3], Step [1890/3236], Loss: 2.0583, Perplexity: 7.8323
Epoch [2/3], Step [1891/3236], Loss: 2.1318, Perplexity: 8.4303
Epoch [2/3], Step [1892/3236], Loss: 2.3722, Perplexity: 10.7209
Epoch [2/3], Step [1893/3236], Loss: 1.9475, Perplexity: 7.0114
Epoch [2/3], Step [1894/3236], Loss: 2.0619, Perplexity: 7.8610
Epoch [2/3], Step [1895/3236], Loss: 1.9656, Perplexity: 7.1389
Epoch [2/3], Step [1896/3236], Loss: 2.0315, Perplexity: 7.6256
Epoch [2/3], Step [1897/3236], Loss: 2.0868, Perplexity: 8.0595
Epoch [2/3], Step [1898/3236], Loss: 2.1896, Perplexity: 8.9316
Epoch [2/3], Step [1899/3236], Loss: 2.0688, Perplexity: 7.9157
Epoch [2/3], Step [1900/3236], Loss: 2.1104, Perplexity: 8.2512
Epoch [2/3], Step [1901/3236], Loss: 2.3304, Perplexity: 10.2824
Epoch [2/3], Step [1902/3236], Loss: 2.0772, Perplexity: 7.9824
Epoch [2/3], Step [1903/3236], Loss: 2.0614, Perplexity: 7.8566
Epoch [2/3], Step [1904/3236], Loss: 2.1082, Perplexity: 8.2337
Epoch [2/3], Step [1905/3236], Loss: 2.1114, Perplexity: 8.2599
Epoch [2/3], Step [1906/3236], Loss: 1.9914, Perplexity: 7.3259
Epoch [2/3], Step [1907/3236], Loss: 2.0925, Perplexity: 8.1051
Epoch [2/3], Step [1908/3236], Loss: 2.8735, Perplexity: 17.6989
Epoch [2/3], Step [1909/3236], Loss: 1.9836, Perplexity: 7.2692
Epoch [2/3], Step [1910/3236], Loss: 2.0982, Perplexity: 8.1519
Epoch [2/3], Step [1911/3236], Loss: 2.1086, Perplexity: 8.2365
Epoch [2/3], Step [1912/3236], Loss: 2.2097, Perplexity: 9.1133
Epoch [2/3], Step [1913/3236], Loss: 2.1047, Perplexity: 8.2047
Epoch [2/3], Step [1914/3236], Loss: 2.0948, Perplexity: 8.1235
Epoch [2/3], Step [1915/3236], Loss: 2.1067, Perplexity: 8.2208
Epoch [2/3], Step [1916/3236], Loss: 2.0004, Perplexity: 7.3918
Epoch [2/3], Step [1917/3236], Loss: 2.0617, Perplexity: 7.8592
Epoch [2/3], Step [1918/3236], Loss: 1.9665, Perplexity: 7.1458
Epoch [2/3], Step [1919/3236], Loss: 2.1270, Perplexity: 8.3900
Epoch [2/3], Step [1920/3236], Loss: 2.0217, Perplexity: 7.5509
Epoch [2/3], Step [1921/3236], Loss: 2.2948, Perplexity: 9.9226
Epoch [2/3], Step [1922/3236], Loss: 2.1580, Perplexity: 8.6534
Epoch [2/3], Step [1923/3236], Loss: 2.1334, Perplexity: 8.4435
Epoch [2/3], Step [1924/3236], Loss: 1.9888, Perplexity: 7.3065
Epoch [2/3], Step [1925/3236], Loss: 2.1214, Perplexity: 8.3432
Epoch [2/3], Step [1926/3236], Loss: 2.0367, Perplexity: 7.6655
Epoch [2/3], Step [1927/3236], Loss: 1.9057, Perplexity: 6.7244
Epoch [2/3], Step [1928/3236], Loss: 1.9409, Perplexity: 6.9653
Epoch [2/3], Step [1929/3236], Loss: 1.9669, Perplexity: 7.1482
Epoch [2/3], Step [1930/3236], Loss: 1.9889, Perplexity: 7.3073
Epoch [2/3], Step [1931/3236], Loss: 2.0117, Perplexity: 7.4762
Epoch [2/3], Step [1932/3236], Loss: 2.1154, Perplexity: 8.2929
Epoch [2/3], Step [1933/3236], Loss: 2.0523, Perplexity: 7.7854
Epoch [2/3], Step [1934/3236], Loss: 2.0704, Perplexity: 7.9281
Epoch [2/3], Step [1935/3236], Loss: 2.1081, Perplexity: 8.2325
Epoch [2/3], Step [1936/3236], Loss: 2.3418, Perplexity: 10.3996
Epoch [2/3], Step [1937/3236], Loss: 2.0495, Perplexity: 7.7639
Epoch [2/3], Step [1938/3236], Loss: 1.9963, Perplexity: 7.3615
Epoch [2/3], Step [1939/3236], Loss: 2.0887, Perplexity: 8.0747
Epoch [2/3], Step [1940/3236], Loss: 1.9410, Perplexity: 6.9658
Epoch [2/3], Step [1941/3236], Loss: 1.9937, Perplexity: 7.3428
Epoch [2/3], Step [1942/3236], Loss: 2.6837, Perplexity: 14.6386
Epoch [2/3], Step [1943/3236], Loss: 1.9104, Perplexity: 6.7560
Epoch [2/3], Step [1944/3236], Loss: 2.0281, Perplexity: 7.5996
Epoch [2/3], Step [1945/3236], Loss: 2.0451, Perplexity: 7.7299
Epoch [2/3], Step [1946/3236], Loss: 2.0570, Perplexity: 7.8228
Epoch [2/3], Step [1947/3236], Loss: 2.3260, Perplexity: 10.2365
Epoch [2/3], Step [1948/3236], Loss: 2.1026, Perplexity: 8.1873
Epoch [2/3], Step [1949/3236], Loss: 2.2937, Perplexity: 9.9112
Epoch [2/3], Step [1950/3236], Loss: 2.0342, Perplexity: 7.6463
Epoch [2/3], Step [1951/3236], Loss: 2.0982, Perplexity: 8.1517
Epoch [2/3], Step [1952/3236], Loss: 2.0129, Perplexity: 7.4849
Epoch [2/3], Step [1953/3236], Loss: 2.0714, Perplexity: 7.9357
Epoch [2/3], Step [1954/3236], Loss: 1.9587, Perplexity: 7.0898
Epoch [2/3], Step [1955/3236], Loss: 2.1456, Perplexity: 8.5469
Epoch [2/3], Step [1956/3236], Loss: 2.0846, Perplexity: 8.0414
Epoch [2/3], Step [1957/3236], Loss: 1.9249, Perplexity: 6.8547
Epoch [2/3], Step [1958/3236], Loss: 2.1704, Perplexity: 8.7616
Epoch [2/3], Step [1959/3236], Loss: 2.0769, Perplexity: 7.9794
Epoch [2/3], Step [1960/3236], Loss: 2.0161, Perplexity: 7.5090
Epoch [2/3], Step [1961/3236], Loss: 2.0675, Perplexity: 7.9052
Epoch [2/3], Step [1962/3236], Loss: 2.3910, Perplexity: 10.9243
Epoch [2/3], Step [1963/3236], Loss: 2.0959, Perplexity: 8.1330
Epoch [2/3], Step [1964/3236], Loss: 2.0350, Perplexity: 7.6523
Epoch [2/3], Step [1965/3236], Loss: 2.0075, Perplexity: 7.4449
Epoch [2/3], Step [1966/3236], Loss: 2.1931, Perplexity: 8.9632
Epoch [2/3], Step [1967/3236], Loss: 2.1037, Perplexity: 8.1964
Epoch [2/3], Step [1968/3236], Loss: 2.2648, Perplexity: 9.6289
Epoch [2/3], Step [1969/3236], Loss: 1.9391, Perplexity: 6.9522
Epoch [2/3], Step [1970/3236], Loss: 2.0421, Perplexity: 7.7072
Epoch [2/3], Step [1971/3236], Loss: 2.2449, Perplexity: 9.4399
Epoch [2/3], Step [1972/3236], Loss: 2.0628, Perplexity: 7.8683
Epoch [2/3], Step [1973/3236], Loss: 2.1696, Perplexity: 8.7545
Epoch [2/3], Step [1974/3236], Loss: 2.2277, Perplexity: 9.2784
Epoch [2/3], Step [1975/3236], Loss: 2.0939, Perplexity: 8.1167
Epoch [2/3], Step [1976/3236], Loss: 1.9859, Perplexity: 7.2858
Epoch [2/3], Step [1977/3236], Loss: 2.0825, Perplexity: 8.0246
Epoch [2/3], Step [1978/3236], Loss: 2.0698, Perplexity: 7.9236
Epoch [2/3], Step [1979/3236], Loss: 1.9230, Perplexity: 6.8417
Epoch [2/3], Step [1980/3236], Loss: 2.0125, Perplexity: 7.4821
Epoch [2/3], Step [1981/3236], Loss: 2.0023, Perplexity: 7.4061
Epoch [2/3], Step [1982/3236], Loss: 2.6788, Perplexity: 14.5675
Epoch [2/3], Step [1983/3236], Loss: 1.9564, Perplexity: 7.0739
Epoch [2/3], Step [1984/3236], Loss: 2.0112, Perplexity: 7.4722
Epoch [2/3], Step [1985/3236], Loss: 2.0431, Perplexity: 7.7142
Epoch [2/3], Step [1986/3236], Loss: 2.0173, Perplexity: 7.5182
Epoch [2/3], Step [1987/3236], Loss: 1.9243, Perplexity: 6.8502
Epoch [2/3], Step [1988/3236], Loss: 2.1018, Perplexity: 8.1805
Epoch [2/3], Step [1989/3236], Loss: 2.1497, Perplexity: 8.5822
Epoch [2/3], Step [1990/3236], Loss: 2.0965, Perplexity: 8.1374
Epoch [2/3], Step [1991/3236], Loss: 2.4383, Perplexity: 11.4530
Epoch [2/3], Step [1992/3236], Loss: 2.1165, Perplexity: 8.3024
Epoch [2/3], Step [1993/3236], Loss: 2.1879, Perplexity: 8.9169
Epoch [2/3], Step [1994/3236], Loss: 2.0780, Perplexity: 7.9882
Epoch [2/3], Step [1995/3236], Loss: 2.0595, Perplexity: 7.8420
Epoch [2/3], Step [1996/3236], Loss: 1.9139, Perplexity: 6.7795
Epoch [2/3], Step [1997/3236], Loss: 2.1728, Perplexity: 8.7826
Epoch [2/3], Step [1998/3236], Loss: 2.6752, Perplexity: 14.5149
Epoch [2/3], Step [1999/3236], Loss: 2.0069, Perplexity: 7.4399
Epoch [2/3], Step [2000/3236], Loss: 2.1711, Perplexity: 8.7680
Epoch [2/3], Step [2001/3236], Loss: 1.9610, Perplexity: 7.1063
Epoch [2/3], Step [2002/3236], Loss: 2.0823, Perplexity: 8.0229
Epoch [2/3], Step [2003/3236], Loss: 2.4239, Perplexity: 11.2903
Epoch [2/3], Step [2004/3236], Loss: 2.4771, Perplexity: 11.9065
Epoch [2/3], Step [2005/3236], Loss: 1.9679, Perplexity: 7.1559
Epoch [2/3], Step [2006/3236], Loss: 2.1437, Perplexity: 8.5309
Epoch [2/3], Step [2007/3236], Loss: 2.0378, Perplexity: 7.6738
Epoch [2/3], Step [2008/3236], Loss: 2.0602, Perplexity: 7.8475
Epoch [2/3], Step [2009/3236], Loss: 2.0665, Perplexity: 7.8973
Epoch [2/3], Step [2010/3236], Loss: 2.0362, Perplexity: 7.6615
Epoch [2/3], Step [2011/3236], Loss: 1.9872, Perplexity: 7.2951
Epoch [2/3], Step [2012/3236], Loss: 2.0952, Perplexity: 8.1267
Epoch [2/3], Step [2013/3236], Loss: 2.1379, Perplexity: 8.4817
Epoch [2/3], Step [2014/3236], Loss: 2.0857, Perplexity: 8.0500
Epoch [2/3], Step [2015/3236], Loss: 2.2371, Perplexity: 9.3660
Epoch [2/3], Step [2016/3236], Loss: 2.1229, Perplexity: 8.3549
Epoch [2/3], Step [2017/3236], Loss: 2.0905, Perplexity: 8.0889
Epoch [2/3], Step [2018/3236], Loss: 2.1398, Perplexity: 8.4975
Epoch [2/3], Step [2019/3236], Loss: 2.1790, Perplexity: 8.8378
Epoch [2/3], Step [2020/3236], Loss: 1.9360, Perplexity: 6.9310
Epoch [2/3], Step [2021/3236], Loss: 1.9017, Perplexity: 6.6975
Epoch [2/3], Step [2022/3236], Loss: 1.8573, Perplexity: 6.4063
Epoch [2/3], Step [2023/3236], Loss: 1.9839, Perplexity: 7.2714
Epoch [2/3], Step [2024/3236], Loss: 1.9254, Perplexity: 6.8576
Epoch [2/3], Step [2025/3236], Loss: 2.3862, Perplexity: 10.8722
Epoch [2/3], Step [2026/3236], Loss: 2.1062, Perplexity: 8.2172
Epoch [2/3], Step [2027/3236], Loss: 1.9607, Perplexity: 7.1046
Epoch [2/3], Step [2028/3236], Loss: 2.0536, Perplexity: 7.7961
Epoch [2/3], Step [2029/3236], Loss: 1.9889, Perplexity: 7.3075
Epoch [2/3], Step [2030/3236], Loss: 1.9564, Perplexity: 7.0737
Epoch [2/3], Step [2031/3236], Loss: 1.9863, Perplexity: 7.2882
Epoch [2/3], Step [2032/3236], Loss: 2.0948, Perplexity: 8.1235
Epoch [2/3], Step [2033/3236], Loss: 1.9248, Perplexity: 6.8536
Epoch [2/3], Step [2034/3236], Loss: 1.9905, Perplexity: 7.3195
Epoch [2/3], Step [2035/3236], Loss: 2.2298, Perplexity: 9.2977
Epoch [2/3], Step [2036/3236], Loss: 2.0965, Perplexity: 8.1377
Epoch [2/3], Step [2037/3236], Loss: 2.3411, Perplexity: 10.3929
Epoch [2/3], Step [2038/3236], Loss: 1.9813, Perplexity: 7.2519
Epoch [2/3], Step [2039/3236], Loss: 1.7628, Perplexity: 5.8285
Epoch [2/3], Step [2040/3236], Loss: 1.8714, Perplexity: 6.4971
Epoch [2/3], Step [2041/3236], Loss: 2.0185, Perplexity: 7.5272
Epoch [2/3], Step [2042/3236], Loss: 1.9843, Perplexity: 7.2739
Epoch [2/3], Step [2043/3236], Loss: 2.1194, Perplexity: 8.3261
Epoch [2/3], Step [2044/3236], Loss: 1.9314, Perplexity: 6.8994
Epoch [2/3], Step [2045/3236], Loss: 2.3213, Perplexity: 10.1891
Epoch [2/3], Step [2046/3236], Loss: 2.5172, Perplexity: 12.3940
Epoch [2/3], Step [2047/3236], Loss: 2.0514, Perplexity: 7.7788
Epoch [2/3], Step [2048/3236], Loss: 2.0648, Perplexity: 7.8838
Epoch [2/3], Step [2049/3236], Loss: 2.6219, Perplexity: 13.7616
Epoch [2/3], Step [2050/3236], Loss: 2.1366, Perplexity: 8.4709
Epoch [2/3], Step [2051/3236], Loss: 2.0576, Perplexity: 7.8269
Epoch [2/3], Step [2052/3236], Loss: 2.0549, Perplexity: 7.8063
Epoch [2/3], Step [2053/3236], Loss: 2.2288, Perplexity: 9.2889
Epoch [2/3], Step [2054/3236], Loss: 1.9541, Perplexity: 7.0578
Epoch [2/3], Step [2055/3236], Loss: 2.0041, Perplexity: 7.4193
Epoch [2/3], Step [2056/3236], Loss: 2.0562, Perplexity: 7.8159
Epoch [2/3], Step [2057/3236], Loss: 2.1806, Perplexity: 8.8516
Epoch [2/3], Step [2058/3236], Loss: 2.4838, Perplexity: 11.9869
Epoch [2/3], Step [2059/3236], Loss: 2.0679, Perplexity: 7.9085
Epoch [2/3], Step [2060/3236], Loss: 1.8851, Perplexity: 6.5873
Epoch [2/3], Step [2061/3236], Loss: 2.0237, Perplexity: 7.5663
Epoch [2/3], Step [2062/3236], Loss: 2.1252, Perplexity: 8.3745
Epoch [2/3], Step [2063/3236], Loss: 2.0796, Perplexity: 8.0012
Epoch [2/3], Step [2064/3236], Loss: 2.1202, Perplexity: 8.3324
Epoch [2/3], Step [2065/3236], Loss: 2.0411, Perplexity: 7.6993
Epoch [2/3], Step [2066/3236], Loss: 1.8584, Perplexity: 6.4137
Epoch [2/3], Step [2067/3236], Loss: 2.0875, Perplexity: 8.0650
Epoch [2/3], Step [2068/3236], Loss: 2.1440, Perplexity: 8.5336
Epoch [2/3], Step [2069/3236], Loss: 2.0352, Perplexity: 7.6541
Epoch [2/3], Step [2070/3236], Loss: 2.0890, Perplexity: 8.0767
Epoch [2/3], Step [2071/3236], Loss: 2.2106, Perplexity: 9.1212
Epoch [2/3], Step [2072/3236], Loss: 2.0848, Perplexity: 8.0427
Epoch [2/3], Step [2073/3236], Loss: 2.2035, Perplexity: 9.0567
Epoch [2/3], Step [2074/3236], Loss: 2.0495, Perplexity: 7.7642
Epoch [2/3], Step [2075/3236], Loss: 2.1760, Perplexity: 8.8112
Epoch [2/3], Step [2076/3236], Loss: 1.9305, Perplexity: 6.8931
Epoch [2/3], Step [2077/3236], Loss: 2.2751, Perplexity: 9.7287
Epoch [2/3], Step [2078/3236], Loss: 2.0201, Perplexity: 7.5388
Epoch [2/3], Step [2079/3236], Loss: 2.2620, Perplexity: 9.6026
Epoch [2/3], Step [2080/3236], Loss: 1.9865, Perplexity: 7.2902
Epoch [2/3], Step [2081/3236], Loss: 1.9974, Perplexity: 7.3701
Epoch [2/3], Step [2082/3236], Loss: 2.3510, Perplexity: 10.4961
Epoch [2/3], Step [2083/3236], Loss: 1.9622, Perplexity: 7.1148
Epoch [2/3], Step [2084/3236], Loss: 2.0814, Perplexity: 8.0159
Epoch [2/3], Step [2085/3236], Loss: 2.1518, Perplexity: 8.6007
Epoch [2/3], Step [2086/3236], Loss: 2.2072, Perplexity: 9.0901
Epoch [2/3], Step [2087/3236], Loss: 2.3256, Perplexity: 10.2332
Epoch [2/3], Step [2088/3236], Loss: 2.0052, Perplexity: 7.4279
Epoch [2/3], Step [2089/3236], Loss: 1.9395, Perplexity: 6.9554
Epoch [2/3], Step [2090/3236], Loss: 2.0294, Perplexity: 7.6095
Epoch [2/3], Step [2091/3236], Loss: 2.1441, Perplexity: 8.5344
Epoch [2/3], Step [2092/3236], Loss: 1.9639, Perplexity: 7.1269
Epoch [2/3], Step [2093/3236], Loss: 2.1545, Perplexity: 8.6234
Epoch [2/3], Step [2094/3236], Loss: 1.9310, Perplexity: 6.8963
Epoch [2/3], Step [2095/3236], Loss: 2.0593, Perplexity: 7.8401
Epoch [2/3], Step [2096/3236], Loss: 1.9869, Perplexity: 7.2927
Epoch [2/3], Step [2097/3236], Loss: 2.1585, Perplexity: 8.6583
Epoch [2/3], Step [2098/3236], Loss: 2.0537, Perplexity: 7.7968
Epoch [2/3], Step [2099/3236], Loss: 2.1159, Perplexity: 8.2967
Epoch [2/3], Step [2100/3236], Loss: 1.8910, Perplexity: 6.6260
Epoch [2/3], Step [2101/3236], Loss: 2.4901, Perplexity: 12.0630
Epoch [2/3], Step [2102/3236], Loss: 2.1114, Perplexity: 8.2595
Epoch [2/3], Step [2103/3236], Loss: 2.9053, Perplexity: 18.2713
Epoch [2/3], Step [2104/3236], Loss: 2.0878, Perplexity: 8.0668
Epoch [2/3], Step [2105/3236], Loss: 1.9117, Perplexity: 6.7645
Epoch [2/3], Step [2106/3236], Loss: 2.0604, Perplexity: 7.8490
Epoch [2/3], Step [2107/3236], Loss: 2.1232, Perplexity: 8.3580
Epoch [2/3], Step [2108/3236], Loss: 2.0576, Perplexity: 7.8275
Epoch [2/3], Step [2109/3236], Loss: 1.9657, Perplexity: 7.1402
Epoch [2/3], Step [2110/3236], Loss: 2.0180, Perplexity: 7.5232
Epoch [2/3], Step [2111/3236], Loss: 3.6649, Perplexity: 39.0516
Epoch [2/3], Step [2112/3236], Loss: 2.3760, Perplexity: 10.7616
Epoch [2/3], Step [2113/3236], Loss: 2.7727, Perplexity: 16.0021
Epoch [2/3], Step [2114/3236], Loss: 2.1724, Perplexity: 8.7795
Epoch [2/3], Step [2115/3236], Loss: 2.0449, Perplexity: 7.7286
Epoch [2/3], Step [2116/3236], Loss: 1.9998, Perplexity: 7.3876
Epoch [2/3], Step [2117/3236], Loss: 2.1522, Perplexity: 8.6039
Epoch [2/3], Step [2118/3236], Loss: 2.1193, Perplexity: 8.3249
Epoch [2/3], Step [2119/3236], Loss: 2.0271, Perplexity: 7.5917
Epoch [2/3], Step [2120/3236], Loss: 2.2945, Perplexity: 9.9196
Epoch [2/3], Step [2121/3236], Loss: 2.0702, Perplexity: 7.9263
Epoch [2/3], Step [2122/3236], Loss: 2.0464, Perplexity: 7.7398
Epoch [2/3], Step [2123/3236], Loss: 2.0239, Perplexity: 7.5675
Epoch [2/3], Step [2124/3236], Loss: 2.0157, Perplexity: 7.5062
Epoch [2/3], Step [2125/3236], Loss: 1.8813, Perplexity: 6.5618
Epoch [2/3], Step [2126/3236], Loss: 1.9777, Perplexity: 7.2261
Epoch [2/3], Step [2127/3236], Loss: 2.1650, Perplexity: 8.7148
Epoch [2/3], Step [2128/3236], Loss: 1.9006, Perplexity: 6.6897
Epoch [2/3], Step [2129/3236], Loss: 2.4925, Perplexity: 12.0917
Epoch [2/3], Step [2130/3236], Loss: 1.9954, Perplexity: 7.3549
Epoch [2/3], Step [2131/3236], Loss: 2.1135, Perplexity: 8.2771
Epoch [2/3], Step [2132/3236], Loss: 2.0282, Perplexity: 7.6008
Epoch [2/3], Step [2133/3236], Loss: 2.0323, Perplexity: 7.6316
Epoch [2/3], Step [2134/3236], Loss: 2.3060, Perplexity: 10.0341
Epoch [2/3], Step [2135/3236], Loss: 2.2382, Perplexity: 9.3761
Epoch [2/3], Step [2136/3236], Loss: 2.0844, Perplexity: 8.0399
Epoch [2/3], Step [2137/3236], Loss: 2.0118, Perplexity: 7.4766
Epoch [2/3], Step [2138/3236], Loss: 2.2736, Perplexity: 9.7145
Epoch [2/3], Step [2139/3236], Loss: 1.9244, Perplexity: 6.8511
Epoch [2/3], Step [2140/3236], Loss: 2.0086, Perplexity: 7.4528
Epoch [2/3], Step [2141/3236], Loss: 2.3247, Perplexity: 10.2231
Epoch [2/3], Step [2142/3236], Loss: 1.9957, Perplexity: 7.3571
Epoch [2/3], Step [2143/3236], Loss: 1.8929, Perplexity: 6.6386
Epoch [2/3], Step [2144/3236], Loss: 1.9816, Perplexity: 7.2546
Epoch [2/3], Step [2145/3236], Loss: 2.1667, Perplexity: 8.7297
Epoch [2/3], Step [2146/3236], Loss: 1.9920, Perplexity: 7.3301
Epoch [2/3], Step [2147/3236], Loss: 1.9428, Perplexity: 6.9779
Epoch [2/3], Step [2148/3236], Loss: 2.0895, Perplexity: 8.0806
Epoch [2/3], Step [2149/3236], Loss: 2.0547, Perplexity: 7.8043
Epoch [2/3], Step [2150/3236], Loss: 2.1094, Perplexity: 8.2430
Epoch [2/3], Step [2151/3236], Loss: 2.0648, Perplexity: 7.8836
Epoch [2/3], Step [2152/3236], Loss: 2.2752, Perplexity: 9.7303
Epoch [2/3], Step [2153/3236], Loss: 2.0572, Perplexity: 7.8238
Epoch [2/3], Step [2154/3236], Loss: 2.4445, Perplexity: 11.5253
Epoch [2/3], Step [2155/3236], Loss: 2.2864, Perplexity: 9.8395
Epoch [2/3], Step [2156/3236], Loss: 2.1507, Perplexity: 8.5912
Epoch [2/3], Step [2157/3236], Loss: 2.0843, Perplexity: 8.0389
Epoch [2/3], Step [2158/3236], Loss: 2.1612, Perplexity: 8.6815
Epoch [2/3], Step [2159/3236], Loss: 1.8983, Perplexity: 6.6745
Epoch [2/3], Step [2160/3236], Loss: 2.0650, Perplexity: 7.8852
Epoch [2/3], Step [2161/3236], Loss: 2.0708, Perplexity: 7.9315
Epoch [2/3], Step [2162/3236], Loss: 2.0966, Perplexity: 8.1388
Epoch [2/3], Step [2163/3236], Loss: 1.9635, Perplexity: 7.1245
Epoch [2/3], Step [2164/3236], Loss: 2.3362, Perplexity: 10.3415
Epoch [2/3], Step [2165/3236], Loss: 2.4374, Perplexity: 11.4428
Epoch [2/3], Step [2166/3236], Loss: 2.0950, Perplexity: 8.1258
Epoch [2/3], Step [2167/3236], Loss: 2.0770, Perplexity: 7.9809
Epoch [2/3], Step [2168/3236], Loss: 2.0467, Perplexity: 7.7423
Epoch [2/3], Step [2169/3236], Loss: 2.1456, Perplexity: 8.5469
Epoch [2/3], Step [2170/3236], Loss: 2.0060, Perplexity: 7.4334
Epoch [2/3], Step [2171/3236], Loss: 2.0622, Perplexity: 7.8634
Epoch [2/3], Step [2172/3236], Loss: 2.1795, Perplexity: 8.8417
Epoch [2/3], Step [2173/3236], Loss: 1.9040, Perplexity: 6.7129
Epoch [2/3], Step [2174/3236], Loss: 1.9645, Perplexity: 7.1317
Epoch [2/3], Step [2175/3236], Loss: 1.9712, Perplexity: 7.1794
Epoch [2/3], Step [2176/3236], Loss: 2.1565, Perplexity: 8.6404
Epoch [2/3], Step [2177/3236], Loss: 1.9389, Perplexity: 6.9509
Epoch [2/3], Step [2178/3236], Loss: 2.1921, Perplexity: 8.9539
Epoch [2/3], Step [2179/3236], Loss: 2.0867, Perplexity: 8.0587
Epoch [2/3], Step [2180/3236], Loss: 2.3420, Perplexity: 10.4019
Epoch [2/3], Step [2181/3236], Loss: 2.1683, Perplexity: 8.7431
Epoch [2/3], Step [2182/3236], Loss: 1.9697, Perplexity: 7.1684
Epoch [2/3], Step [2183/3236], Loss: 2.0498, Perplexity: 7.7660
Epoch [2/3], Step [2184/3236], Loss: 2.0052, Perplexity: 7.4279
Epoch [2/3], Step [2185/3236], Loss: 2.0227, Perplexity: 7.5588
Epoch [2/3], Step [2186/3236], Loss: 2.2731, Perplexity: 9.7090
Epoch [2/3], Step [2187/3236], Loss: 2.1125, Perplexity: 8.2692
Epoch [2/3], Step [2188/3236], Loss: 2.3091, Perplexity: 10.0651
Epoch [2/3], Step [2189/3236], Loss: 1.8550, Perplexity: 6.3919
Epoch [2/3], Step [2190/3236], Loss: 2.0508, Perplexity: 7.7744
Epoch [2/3], Step [2191/3236], Loss: 1.8050, Perplexity: 6.0798
Epoch [2/3], Step [2192/3236], Loss: 2.0540, Perplexity: 7.7988
Epoch [2/3], Step [2193/3236], Loss: 1.8788, Perplexity: 6.5458
Epoch [2/3], Step [2194/3236], Loss: 2.2655, Perplexity: 9.6359
Epoch [2/3], Step [2195/3236], Loss: 2.0030, Perplexity: 7.4109
Epoch [2/3], Step [2196/3236], Loss: 2.1349, Perplexity: 8.4566
Epoch [2/3], Step [2197/3236], Loss: 2.1625, Perplexity: 8.6926
Epoch [2/3], Step [2198/3236], Loss: 1.8820, Perplexity: 6.5664
Epoch [2/3], Step [2199/3236], Loss: 2.4450, Perplexity: 11.5302
Epoch [2/3], Step [2200/3236], Loss: 1.8095, Perplexity: 6.1072
Epoch [2/3], Step [2201/3236], Loss: 2.0734, Perplexity: 7.9516
Epoch [2/3], Step [2202/3236], Loss: 2.0347, Perplexity: 7.6503
Epoch [2/3], Step [2203/3236], Loss: 2.2161, Perplexity: 9.1718
Epoch [2/3], Step [2204/3236], Loss: 2.3998, Perplexity: 11.0214
Epoch [2/3], Step [2205/3236], Loss: 2.0295, Perplexity: 7.6101
Epoch [2/3], Step [2206/3236], Loss: 2.2108, Perplexity: 9.1233
Epoch [2/3], Step [2207/3236], Loss: 2.0555, Perplexity: 7.8111
Epoch [2/3], Step [2208/3236], Loss: 2.0973, Perplexity: 8.1440
Epoch [2/3], Step [2209/3236], Loss: 2.0362, Perplexity: 7.6617
Epoch [2/3], Step [2210/3236], Loss: 1.9809, Perplexity: 7.2492
Epoch [2/3], Step [2211/3236], Loss: 1.9890, Perplexity: 7.3084
Epoch [2/3], Step [2212/3236], Loss: 2.1822, Perplexity: 8.8659
Epoch [2/3], Step [2213/3236], Loss: 2.0087, Perplexity: 7.4535
Epoch [2/3], Step [2214/3236], Loss: 2.0461, Perplexity: 7.7379
Epoch [2/3], Step [2215/3236], Loss: 1.9840, Perplexity: 7.2721
Epoch [2/3], Step [2216/3236], Loss: 1.9936, Perplexity: 7.3422
Epoch [2/3], Step [2217/3236], Loss: 2.0366, Perplexity: 7.6644
Epoch [2/3], Step [2218/3236], Loss: 2.2822, Perplexity: 9.7985
Epoch [2/3], Step [2219/3236], Loss: 1.9564, Perplexity: 7.0739
Epoch [2/3], Step [2220/3236], Loss: 2.0151, Perplexity: 7.5016
Epoch [2/3], Step [2221/3236], Loss: 2.2116, Perplexity: 9.1308
Epoch [2/3], Step [2222/3236], Loss: 1.9832, Perplexity: 7.2663
Epoch [2/3], Step [2223/3236], Loss: 2.1570, Perplexity: 8.6451
Epoch [2/3], Step [2224/3236], Loss: 2.0629, Perplexity: 7.8685
Epoch [2/3], Step [2225/3236], Loss: 2.4204, Perplexity: 11.2499
Epoch [2/3], Step [2226/3236], Loss: 2.1359, Perplexity: 8.4644
Epoch [2/3], Step [2227/3236], Loss: 2.7656, Perplexity: 15.8881
Epoch [2/3], Step [2228/3236], Loss: 2.0824, Perplexity: 8.0236
Epoch [2/3], Step [2229/3236], Loss: 1.9298, Perplexity: 6.8878
Epoch [2/3], Step [2230/3236], Loss: 2.4422, Perplexity: 11.4987
Epoch [2/3], Step [2231/3236], Loss: 2.0933, Perplexity: 8.1116
Epoch [2/3], Step [2232/3236], Loss: 1.8071, Perplexity: 6.0930
Epoch [2/3], Step [2233/3236], Loss: 1.8382, Perplexity: 6.2855
Epoch [2/3], Step [2234/3236], Loss: 2.1212, Perplexity: 8.3410
Epoch [2/3], Step [2235/3236], Loss: 2.3160, Perplexity: 10.1352
Epoch [2/3], Step [2236/3236], Loss: 2.0637, Perplexity: 7.8748
Epoch [2/3], Step [2237/3236], Loss: 1.9955, Perplexity: 7.3560
Epoch [2/3], Step [2238/3236], Loss: 2.0750, Perplexity: 7.9649
Epoch [2/3], Step [2239/3236], Loss: 2.1010, Perplexity: 8.1746
Epoch [2/3], Step [2240/3236], Loss: 2.0592, Perplexity: 7.8398
Epoch [2/3], Step [2241/3236], Loss: 1.8080, Perplexity: 6.0982
Epoch [2/3], Step [2242/3236], Loss: 2.1636, Perplexity: 8.7023
Epoch [2/3], Step [2243/3236], Loss: 1.9563, Perplexity: 7.0729
Epoch [2/3], Step [2244/3236], Loss: 1.9727, Perplexity: 7.1903
Epoch [2/3], Step [2245/3236], Loss: 2.1374, Perplexity: 8.4771
Epoch [2/3], Step [2246/3236], Loss: 2.3767, Perplexity: 10.7698
Epoch [2/3], Step [2247/3236], Loss: 2.3077, Perplexity: 10.0509
Epoch [2/3], Step [2248/3236], Loss: 2.3271, Perplexity: 10.2479
Epoch [2/3], Step [2249/3236], Loss: 2.0692, Perplexity: 7.9184
Epoch [2/3], Step [2250/3236], Loss: 2.1065, Perplexity: 8.2192
Epoch [2/3], Step [2251/3236], Loss: 1.9523, Perplexity: 7.0445
Epoch [2/3], Step [2252/3236], Loss: 2.3334, Perplexity: 10.3130
Epoch [2/3], Step [2253/3236], Loss: 2.0527, Perplexity: 7.7891
Epoch [2/3], Step [2254/3236], Loss: 2.1287, Perplexity: 8.4041
Epoch [2/3], Step [2255/3236], Loss: 1.8825, Perplexity: 6.5696
Epoch [2/3], Step [2256/3236], Loss: 2.3109, Perplexity: 10.0832
Epoch [2/3], Step [2257/3236], Loss: 2.5277, Perplexity: 12.5244
Epoch [2/3], Step [2258/3236], Loss: 1.9078, Perplexity: 6.7384
Epoch [2/3], Step [2259/3236], Loss: 2.1487, Perplexity: 8.5734
Epoch [2/3], Step [2260/3236], Loss: 1.9181, Perplexity: 6.8079
Epoch [2/3], Step [2261/3236], Loss: 2.0050, Perplexity: 7.4258
Epoch [2/3], Step [2262/3236], Loss: 2.1080, Perplexity: 8.2320
Epoch [2/3], Step [2263/3236], Loss: 1.9964, Perplexity: 7.3624
Epoch [2/3], Step [2264/3236], Loss: 1.9238, Perplexity: 6.8469
Epoch [2/3], Step [2265/3236], Loss: 1.9405, Perplexity: 6.9625
Epoch [2/3], Step [2266/3236], Loss: 2.3144, Perplexity: 10.1187
Epoch [2/3], Step [2267/3236], Loss: 2.0709, Perplexity: 7.9318
Epoch [2/3], Step [2268/3236], Loss: 2.0256, Perplexity: 7.5806
Epoch [2/3], Step [2269/3236], Loss: 2.5071, Perplexity: 12.2696
Epoch [2/3], Step [2270/3236], Loss: 1.9658, Perplexity: 7.1405
Epoch [2/3], Step [2271/3236], Loss: 1.9852, Perplexity: 7.2803
Epoch [2/3], Step [2272/3236], Loss: 2.0306, Perplexity: 7.6188
Epoch [2/3], Step [2273/3236], Loss: 1.8866, Perplexity: 6.5967
Epoch [2/3], Step [2274/3236], Loss: 1.8532, Perplexity: 6.3801
Epoch [2/3], Step [2275/3236], Loss: 2.7432, Perplexity: 15.5368
Epoch [2/3], Step [2276/3236], Loss: 1.9059, Perplexity: 6.7251
Epoch [2/3], Step [2277/3236], Loss: 2.0683, Perplexity: 7.9115
Epoch [2/3], Step [2278/3236], Loss: 2.0119, Perplexity: 7.4779
Epoch [2/3], Step [2279/3236], Loss: 1.9620, Perplexity: 7.1136
Epoch [2/3], Step [2280/3236], Loss: 2.0071, Perplexity: 7.4417
Epoch [2/3], Step [2281/3236], Loss: 2.3780, Perplexity: 10.7838
Epoch [2/3], Step [2282/3236], Loss: 2.3668, Perplexity: 10.6634
Epoch [2/3], Step [2283/3236], Loss: 2.0106, Perplexity: 7.4679
Epoch [2/3], Step [2284/3236], Loss: 2.1665, Perplexity: 8.7278
Epoch [2/3], Step [2285/3236], Loss: 2.2947, Perplexity: 9.9212
Epoch [2/3], Step [2286/3236], Loss: 2.0637, Perplexity: 7.8748
Epoch [2/3], Step [2287/3236], Loss: 2.0139, Perplexity: 7.4924
Epoch [2/3], Step [2288/3236], Loss: 2.1584, Perplexity: 8.6570
Epoch [2/3], Step [2289/3236], Loss: 2.1438, Perplexity: 8.5316
Epoch [2/3], Step [2290/3236], Loss: 2.0117, Perplexity: 7.4757
Epoch [2/3], Step [2291/3236], Loss: 2.0266, Perplexity: 7.5879
Epoch [2/3], Step [2292/3236], Loss: 2.0506, Perplexity: 7.7726
Epoch [2/3], Step [2293/3236], Loss: 2.0431, Perplexity: 7.7146
Epoch [2/3], Step [2294/3236], Loss: 2.0776, Perplexity: 7.9854
Epoch [2/3], Step [2295/3236], Loss: 2.0445, Perplexity: 7.7253
Epoch [2/3], Step [2296/3236], Loss: 2.0550, Perplexity: 7.8065
Epoch [2/3], Step [2297/3236], Loss: 2.0687, Perplexity: 7.9147
Epoch [2/3], Step [2298/3236], Loss: 2.0783, Perplexity: 7.9909
Epoch [2/3], Step [2299/3236], Loss: 2.0089, Perplexity: 7.4553
Epoch [2/3], Step [2300/3236], Loss: 2.0994, Perplexity: 8.1614
Epoch [2/3], Step [2301/3236], Loss: 1.9806, Perplexity: 7.2473
Epoch [2/3], Step [2302/3236], Loss: 2.2690, Perplexity: 9.6701
Epoch [2/3], Step [2303/3236], Loss: 2.0828, Perplexity: 8.0270
Epoch [2/3], Step [2304/3236], Loss: 1.9750, Perplexity: 7.2066
Epoch [2/3], Step [2305/3236], Loss: 2.0756, Perplexity: 7.9696
Epoch [2/3], Step [2306/3236], Loss: 2.0825, Perplexity: 8.0243
Epoch [2/3], Step [2307/3236], Loss: 2.0388, Perplexity: 7.6813
Epoch [2/3], Step [2308/3236], Loss: 2.0586, Perplexity: 7.8350
Epoch [2/3], Step [2309/3236], Loss: 1.9414, Perplexity: 6.9684
Epoch [2/3], Step [2310/3236], Loss: 1.9615, Perplexity: 7.1100
Epoch [2/3], Step [2311/3236], Loss: 3.0482, Perplexity: 21.0764
Epoch [2/3], Step [2312/3236], Loss: 1.9633, Perplexity: 7.1225
Epoch [2/3], Step [2313/3236], Loss: 2.1312, Perplexity: 8.4251
Epoch [2/3], Step [2314/3236], Loss: 1.9247, Perplexity: 6.8529
Epoch [2/3], Step [2315/3236], Loss: 1.9892, Perplexity: 7.3098
Epoch [2/3], Step [2316/3236], Loss: 1.9863, Perplexity: 7.2887
Epoch [2/3], Step [2317/3236], Loss: 1.8917, Perplexity: 6.6305
Epoch [2/3], Step [2318/3236], Loss: 1.9770, Perplexity: 7.2210
Epoch [2/3], Step [2319/3236], Loss: 2.2267, Perplexity: 9.2689
Epoch [2/3], Step [2320/3236], Loss: 2.0038, Perplexity: 7.4173
Epoch [2/3], Step [2321/3236], Loss: 1.9669, Perplexity: 7.1486
Epoch [2/3], Step [2322/3236], Loss: 1.9938, Perplexity: 7.3431
Epoch [2/3], Step [2323/3236], Loss: 1.8865, Perplexity: 6.5965
Epoch [2/3], Step [2324/3236], Loss: 2.1598, Perplexity: 8.6693
Epoch [2/3], Step [2325/3236], Loss: 2.0815, Perplexity: 8.0166
Epoch [2/3], Step [2326/3236], Loss: 2.0229, Perplexity: 7.5601
Epoch [2/3], Step [2327/3236], Loss: 2.0568, Perplexity: 7.8208
Epoch [2/3], Step [2328/3236], Loss: 2.1076, Perplexity: 8.2289
Epoch [2/3], Step [2329/3236], Loss: 1.9344, Perplexity: 6.9200
Epoch [2/3], Step [2330/3236], Loss: 2.0432, Perplexity: 7.7151
Epoch [2/3], Step [2331/3236], Loss: 1.8186, Perplexity: 6.1633
Epoch [2/3], Step [2332/3236], Loss: 2.0203, Perplexity: 7.5404
Epoch [2/3], Step [2333/3236], Loss: 2.6509, Perplexity: 14.1667
Epoch [2/3], Step [2334/3236], Loss: 2.0145, Perplexity: 7.4968
Epoch [2/3], Step [2335/3236], Loss: 1.9169, Perplexity: 6.7995
Epoch [2/3], Step [2336/3236], Loss: 1.9162, Perplexity: 6.7948
Epoch [2/3], Step [2337/3236], Loss: 2.1595, Perplexity: 8.6664
Epoch [2/3], Step [2338/3236], Loss: 2.0414, Perplexity: 7.7012
Epoch [2/3], Step [2339/3236], Loss: 2.1198, Perplexity: 8.3292
Epoch [2/3], Step [2340/3236], Loss: 2.2192, Perplexity: 9.2002
Epoch [2/3], Step [2341/3236], Loss: 1.9731, Perplexity: 7.1931
Epoch [2/3], Step [2342/3236], Loss: 2.0691, Perplexity: 7.9177
Epoch [2/3], Step [2343/3236], Loss: 2.0472, Perplexity: 7.7459
Epoch [2/3], Step [2344/3236], Loss: 1.9916, Perplexity: 7.3274
Epoch [2/3], Step [2345/3236], Loss: 1.9384, Perplexity: 6.9475
Epoch [2/3], Step [2346/3236], Loss: 2.0273, Perplexity: 7.5935
Epoch [2/3], Step [2347/3236], Loss: 2.1801, Perplexity: 8.8474
Epoch [2/3], Step [2348/3236], Loss: 1.9183, Perplexity: 6.8096
Epoch [2/3], Step [2349/3236], Loss: 2.1306, Perplexity: 8.4197
Epoch [2/3], Step [2350/3236], Loss: 2.0103, Perplexity: 7.4653
Epoch [2/3], Step [2351/3236], Loss: 2.0310, Perplexity: 7.6217
Epoch [2/3], Step [2352/3236], Loss: 2.1420, Perplexity: 8.5163
Epoch [2/3], Step [2353/3236], Loss: 1.9195, Perplexity: 6.8174
Epoch [2/3], Step [2354/3236], Loss: 2.6487, Perplexity: 14.1359
Epoch [2/3], Step [2355/3236], Loss: 1.8633, Perplexity: 6.4447
Epoch [2/3], Step [2356/3236], Loss: 1.9337, Perplexity: 6.9151
Epoch [2/3], Step [2357/3236], Loss: 2.0512, Perplexity: 7.7774
Epoch [2/3], Step [2358/3236], Loss: 1.9846, Perplexity: 7.2763
Epoch [2/3], Step [2359/3236], Loss: 1.9049, Perplexity: 6.7190
Epoch [2/3], Step [2360/3236], Loss: 2.0259, Perplexity: 7.5832
Epoch [2/3], Step [2361/3236], Loss: 2.5691, Perplexity: 13.0542
Epoch [2/3], Step [2362/3236], Loss: 2.0530, Perplexity: 7.7915
Epoch [2/3], Step [2363/3236], Loss: 1.9245, Perplexity: 6.8516
Epoch [2/3], Step [2364/3236], Loss: 1.9923, Perplexity: 7.3320
Epoch [2/3], Step [2365/3236], Loss: 2.0320, Perplexity: 7.6296
Epoch [2/3], Step [2366/3236], Loss: 2.2985, Perplexity: 9.9595
Epoch [2/3], Step [2367/3236], Loss: 2.1220, Perplexity: 8.3481
Epoch [2/3], Step [2368/3236], Loss: 2.7081, Perplexity: 15.0013
Epoch [2/3], Step [2369/3236], Loss: 1.9522, Perplexity: 7.0443
Epoch [2/3], Step [2370/3236], Loss: 2.5841, Perplexity: 13.2514
Epoch [2/3], Step [2371/3236], Loss: 2.0905, Perplexity: 8.0888
Epoch [2/3], Step [2372/3236], Loss: 1.8662, Perplexity: 6.4634
Epoch [2/3], Step [2373/3236], Loss: 2.0245, Perplexity: 7.5727
Epoch [2/3], Step [2374/3236], Loss: 2.3424, Perplexity: 10.4067
Epoch [2/3], Step [2375/3236], Loss: 1.9993, Perplexity: 7.3836
Epoch [2/3], Step [2376/3236], Loss: 2.4395, Perplexity: 11.4676
Epoch [2/3], Step [2377/3236], Loss: 2.1094, Perplexity: 8.2434
Epoch [2/3], Step [2378/3236], Loss: 2.2906, Perplexity: 9.8806
Epoch [2/3], Step [2379/3236], Loss: 1.9739, Perplexity: 7.1987
Epoch [2/3], Step [2380/3236], Loss: 1.9515, Perplexity: 7.0394
Epoch [2/3], Step [2381/3236], Loss: 2.1510, Perplexity: 8.5936
Epoch [2/3], Step [2382/3236], Loss: 1.9571, Perplexity: 7.0789
Epoch [2/3], Step [2383/3236], Loss: 2.0163, Perplexity: 7.5105
Epoch [2/3], Step [2384/3236], Loss: 2.0284, Perplexity: 7.6020
Epoch [2/3], Step [2385/3236], Loss: 2.1841, Perplexity: 8.8825
Epoch [2/3], Step [2386/3236], Loss: 2.0811, Perplexity: 8.0136
Epoch [2/3], Step [2387/3236], Loss: 2.0976, Perplexity: 8.1462
Epoch [2/3], Step [2388/3236], Loss: 1.9994, Perplexity: 7.3849
Epoch [2/3], Step [2389/3236], Loss: 1.8874, Perplexity: 6.6024
Epoch [2/3], Step [2390/3236], Loss: 2.0378, Perplexity: 7.6741
Epoch [2/3], Step [2391/3236], Loss: 2.0048, Perplexity: 7.4249
Epoch [2/3], Step [2392/3236], Loss: 2.0712, Perplexity: 7.9341
Epoch [2/3], Step [2393/3236], Loss: 2.5810, Perplexity: 13.2104
Epoch [2/3], Step [2394/3236], Loss: 2.1679, Perplexity: 8.7401
Epoch [2/3], Step [2395/3236], Loss: 2.3596, Perplexity: 10.5862
Epoch [2/3], Step [2396/3236], Loss: 1.8135, Perplexity: 6.1317
Epoch [2/3], Step [2397/3236], Loss: 2.0425, Perplexity: 7.7099
Epoch [2/3], Step [2398/3236], Loss: 1.9269, Perplexity: 6.8679
Epoch [2/3], Step [2399/3236], Loss: 2.2403, Perplexity: 9.3965
Epoch [2/3], Step [2400/3236], Loss: 2.0225, Perplexity: 7.5574
Epoch [2/3], Step [2401/3236], Loss: 1.9988, Perplexity: 7.3803
Epoch [2/3], Step [2402/3236], Loss: 2.0864, Perplexity: 8.0557
Epoch [2/3], Step [2403/3236], Loss: 2.1364, Perplexity: 8.4689
Epoch [2/3], Step [2404/3236], Loss: 2.0860, Perplexity: 8.0526
Epoch [2/3], Step [2405/3236], Loss: 2.7180, Perplexity: 15.1499
Epoch [2/3], Step [2406/3236], Loss: 2.0635, Perplexity: 7.8734
Epoch [2/3], Step [2407/3236], Loss: 1.8965, Perplexity: 6.6625
Epoch [2/3], Step [2408/3236], Loss: 1.9448, Perplexity: 6.9921
Epoch [2/3], Step [2409/3236], Loss: 2.1865, Perplexity: 8.9044
Epoch [2/3], Step [2410/3236], Loss: 1.9931, Perplexity: 7.3381
Epoch [2/3], Step [2411/3236], Loss: 2.1290, Perplexity: 8.4064
Epoch [2/3], Step [2412/3236], Loss: 2.0402, Perplexity: 7.6921
Epoch [2/3], Step [2413/3236], Loss: 2.0228, Perplexity: 7.5592
Epoch [2/3], Step [2414/3236], Loss: 2.1175, Perplexity: 8.3104
Epoch [2/3], Step [2415/3236], Loss: 2.0651, Perplexity: 7.8858
Epoch [2/3], Step [2416/3236], Loss: 1.9660, Perplexity: 7.1417
Epoch [2/3], Step [2417/3236], Loss: 2.1596, Perplexity: 8.6679
Epoch [2/3], Step [2418/3236], Loss: 2.1449, Perplexity: 8.5412
Epoch [2/3], Step [2419/3236], Loss: 2.2998, Perplexity: 9.9727
Epoch [2/3], Step [2420/3236], Loss: 2.0211, Perplexity: 7.5470
Epoch [2/3], Step [2421/3236], Loss: 1.9806, Perplexity: 7.2474
Epoch [2/3], Step [2422/3236], Loss: 1.9319, Perplexity: 6.9029
Epoch [2/3], Step [2423/3236], Loss: 1.9815, Perplexity: 7.2539
Epoch [2/3], Step [2424/3236], Loss: 2.0262, Perplexity: 7.5852
Epoch [2/3], Step [2425/3236], Loss: 1.9626, Perplexity: 7.1175
Epoch [2/3], Step [2426/3236], Loss: 2.0419, Perplexity: 7.7053
Epoch [2/3], Step [2427/3236], Loss: 2.1667, Perplexity: 8.7292
Epoch [2/3], Step [2428/3236], Loss: 2.0454, Perplexity: 7.7324
Epoch [2/3], Step [2429/3236], Loss: 1.9886, Perplexity: 7.3051
Epoch [2/3], Step [2430/3236], Loss: 2.7722, Perplexity: 15.9930
Epoch [2/3], Step [2431/3236], Loss: 2.3645, Perplexity: 10.6383
Epoch [2/3], Step [2432/3236], Loss: 2.0972, Perplexity: 8.1434
Epoch [2/3], Step [2433/3236], Loss: 2.0024, Perplexity: 7.4071
Epoch [2/3], Step [2434/3236], Loss: 2.2578, Perplexity: 9.5620
Epoch [2/3], Step [2435/3236], Loss: 2.6372, Perplexity: 13.9740
Epoch [2/3], Step [2436/3236], Loss: 1.9748, Perplexity: 7.2049
Epoch [2/3], Step [2437/3236], Loss: 2.0216, Perplexity: 7.5507
Epoch [2/3], Step [2438/3236], Loss: 1.9421, Perplexity: 6.9734
Epoch [2/3], Step [2439/3236], Loss: 2.3151, Perplexity: 10.1264
Epoch [2/3], Step [2440/3236], Loss: 1.9737, Perplexity: 7.1969
Epoch [2/3], Step [2441/3236], Loss: 1.9540, Perplexity: 7.0567
Epoch [2/3], Step [2442/3236], Loss: 1.9415, Perplexity: 6.9695
Epoch [2/3], Step [2443/3236], Loss: 2.1292, Perplexity: 8.4083
Epoch [2/3], Step [2444/3236], Loss: 2.0978, Perplexity: 8.1483
Epoch [2/3], Step [2445/3236], Loss: 2.0225, Perplexity: 7.5575
Epoch [2/3], Step [2446/3236], Loss: 2.0783, Perplexity: 7.9908
Epoch [2/3], Step [2447/3236], Loss: 2.1250, Perplexity: 8.3729
Epoch [2/3], Step [2448/3236], Loss: 2.2729, Perplexity: 9.7072
Epoch [2/3], Step [2449/3236], Loss: 1.9240, Perplexity: 6.8482
Epoch [2/3], Step [2450/3236], Loss: 2.1102, Perplexity: 8.2496
Epoch [2/3], Step [2451/3236], Loss: 2.2327, Perplexity: 9.3252
Epoch [2/3], Step [2452/3236], Loss: 2.0442, Perplexity: 7.7227
Epoch [2/3], Step [2453/3236], Loss: 2.0356, Perplexity: 7.6567
Epoch [2/3], Step [2454/3236], Loss: 1.8784, Perplexity: 6.5428
Epoch [2/3], Step [2455/3236], Loss: 1.8974, Perplexity: 6.6686
Epoch [2/3], Step [2456/3236], Loss: 2.0489, Perplexity: 7.7596
Epoch [2/3], Step [2457/3236], Loss: 1.9802, Perplexity: 7.2445
Epoch [2/3], Step [2458/3236], Loss: 1.9927, Perplexity: 7.3355
Epoch [2/3], Step [2459/3236], Loss: 2.1128, Perplexity: 8.2717
Epoch [2/3], Step [2460/3236], Loss: 1.9993, Perplexity: 7.3840
Epoch [2/3], Step [2461/3236], Loss: 2.3463, Perplexity: 10.4468
Epoch [2/3], Step [2462/3236], Loss: 1.9849, Perplexity: 7.2781
Epoch [2/3], Step [2463/3236], Loss: 1.9189, Perplexity: 6.8131
Epoch [2/3], Step [2464/3236], Loss: 1.9298, Perplexity: 6.8880
Epoch [2/3], Step [2465/3236], Loss: 1.9558, Perplexity: 7.0698
Epoch [2/3], Step [2466/3236], Loss: 2.1154, Perplexity: 8.2932
Epoch [2/3], Step [2467/3236], Loss: 2.2529, Perplexity: 9.5157
Epoch [2/3], Step [2468/3236], Loss: 2.0445, Perplexity: 7.7252
Epoch [2/3], Step [2469/3236], Loss: 2.0407, Perplexity: 7.6963
Epoch [2/3], Step [2470/3236], Loss: 2.0663, Perplexity: 7.8957
Epoch [2/3], Step [2471/3236], Loss: 2.1704, Perplexity: 8.7616
Epoch [2/3], Step [2472/3236], Loss: 2.0227, Perplexity: 7.5583
Epoch [2/3], Step [2473/3236], Loss: 2.0040, Perplexity: 7.4189
Epoch [2/3], Step [2474/3236], Loss: 2.1188, Perplexity: 8.3214
Epoch [2/3], Step [2475/3236], Loss: 2.0713, Perplexity: 7.9351
Epoch [2/3], Step [2476/3236], Loss: 2.0472, Perplexity: 7.7458
Epoch [2/3], Step [2477/3236], Loss: 2.1487, Perplexity: 8.5740
Epoch [2/3], Step [2478/3236], Loss: 2.0374, Perplexity: 7.6709
Epoch [2/3], Step [2479/3236], Loss: 2.0316, Perplexity: 7.6263
Epoch [2/3], Step [2480/3236], Loss: 1.9107, Perplexity: 6.7580
Epoch [2/3], Step [2481/3236], Loss: 1.9878, Perplexity: 7.2992
Epoch [2/3], Step [2482/3236], Loss: 2.0756, Perplexity: 7.9695
Epoch [2/3], Step [2483/3236], Loss: 2.0484, Perplexity: 7.7558
Epoch [2/3], Step [2484/3236], Loss: 2.0313, Perplexity: 7.6240
Epoch [2/3], Step [2485/3236], Loss: 3.0342, Perplexity: 20.7837
Epoch [2/3], Step [2486/3236], Loss: 1.9503, Perplexity: 7.0309
Epoch [2/3], Step [2487/3236], Loss: 1.9542, Perplexity: 7.0580
Epoch [2/3], Step [2488/3236], Loss: 1.9901, Perplexity: 7.3159
Epoch [2/3], Step [2489/3236], Loss: 1.9723, Perplexity: 7.1869
Epoch [2/3], Step [2490/3236], Loss: 2.0383, Perplexity: 7.6776
Epoch [2/3], Step [2491/3236], Loss: 1.8984, Perplexity: 6.6751
Epoch [2/3], Step [2492/3236], Loss: 1.9632, Perplexity: 7.1219
Epoch [2/3], Step [2493/3236], Loss: 2.0110, Perplexity: 7.4706
Epoch [2/3], Step [2494/3236], Loss: 1.9887, Perplexity: 7.3059
Epoch [2/3], Step [2495/3236], Loss: 2.0181, Perplexity: 7.5238
Epoch [2/3], Step [2496/3236], Loss: 1.9449, Perplexity: 6.9927
Epoch [2/3], Step [2497/3236], Loss: 2.0851, Perplexity: 8.0454
Epoch [2/3], Step [2498/3236], Loss: 1.9209, Perplexity: 6.8274
Epoch [2/3], Step [2499/3236], Loss: 2.0263, Perplexity: 7.5862
Epoch [2/3], Step [2500/3236], Loss: 1.9985, Perplexity: 7.3783
Epoch [2/3], Step [2501/3236], Loss: 2.1412, Perplexity: 8.5095
Epoch [2/3], Step [2502/3236], Loss: 1.9638, Perplexity: 7.1263
Epoch [2/3], Step [2503/3236], Loss: 2.0498, Perplexity: 7.7663
Epoch [2/3], Step [2504/3236], Loss: 2.0181, Perplexity: 7.5237
Epoch [2/3], Step [2505/3236], Loss: 2.0757, Perplexity: 7.9700
Epoch [2/3], Step [2506/3236], Loss: 1.9888, Perplexity: 7.3071
Epoch [2/3], Step [2507/3236], Loss: 2.1187, Perplexity: 8.3206
Epoch [2/3], Step [2508/3236], Loss: 2.5125, Perplexity: 12.3352
Epoch [2/3], Step [2509/3236], Loss: 2.0442, Perplexity: 7.7232
Epoch [2/3], Step [2510/3236], Loss: 1.8873, Perplexity: 6.6014
Epoch [2/3], Step [2511/3236], Loss: 1.9798, Perplexity: 7.2416
Epoch [2/3], Step [2512/3236], Loss: 1.9524, Perplexity: 7.0457
Epoch [2/3], Step [2513/3236], Loss: 1.8725, Perplexity: 6.5046
Epoch [2/3], Step [2514/3236], Loss: 1.9958, Perplexity: 7.3584
Epoch [2/3], Step [2515/3236], Loss: 2.0755, Perplexity: 7.9687
Epoch [2/3], Step [2516/3236], Loss: 1.9736, Perplexity: 7.1963
Epoch [2/3], Step [2517/3236], Loss: 2.1081, Perplexity: 8.2322
Epoch [2/3], Step [2518/3236], Loss: 2.0720, Perplexity: 7.9407
Epoch [2/3], Step [2519/3236], Loss: 2.0441, Perplexity: 7.7226
Epoch [2/3], Step [2520/3236], Loss: 2.1614, Perplexity: 8.6829
Epoch [2/3], Step [2521/3236], Loss: 2.0156, Perplexity: 7.5052
Epoch [2/3], Step [2522/3236], Loss: 2.7868, Perplexity: 16.2287
Epoch [2/3], Step [2523/3236], Loss: 2.6178, Perplexity: 13.7061
Epoch [2/3], Step [2524/3236], Loss: 2.1568, Perplexity: 8.6435
Epoch [2/3], Step [2525/3236], Loss: 2.3725, Perplexity: 10.7244
Epoch [2/3], Step [2526/3236], Loss: 2.2478, Perplexity: 9.4669
Epoch [2/3], Step [2527/3236], Loss: 1.9604, Perplexity: 7.1023
Epoch [2/3], Step [2528/3236], Loss: 1.9050, Perplexity: 6.7193
Epoch [2/3], Step [2529/3236], Loss: 1.9238, Perplexity: 6.8470
Epoch [2/3], Step [2530/3236], Loss: 2.3129, Perplexity: 10.1039
Epoch [2/3], Step [2531/3236], Loss: 1.9023, Perplexity: 6.7012
Epoch [2/3], Step [2532/3236], Loss: 1.8087, Perplexity: 6.1024
Epoch [2/3], Step [2533/3236], Loss: 2.1123, Perplexity: 8.2674
Epoch [2/3], Step [2534/3236], Loss: 2.2009, Perplexity: 9.0334
Epoch [2/3], Step [2535/3236], Loss: 1.9703, Perplexity: 7.1732
Epoch [2/3], Step [2536/3236], Loss: 1.8508, Perplexity: 6.3651
Epoch [2/3], Step [2537/3236], Loss: 2.7942, Perplexity: 16.3503
Epoch [2/3], Step [2538/3236], Loss: 2.0874, Perplexity: 8.0637
Epoch [2/3], Step [2539/3236], Loss: 2.0239, Perplexity: 7.5678
Epoch [2/3], Step [2540/3236], Loss: 2.0865, Perplexity: 8.0566
Epoch [2/3], Step [2541/3236], Loss: 1.9043, Perplexity: 6.7145
Epoch [2/3], Step [2542/3236], Loss: 2.1567, Perplexity: 8.6424
Epoch [2/3], Step [2543/3236], Loss: 1.9919, Perplexity: 7.3296
Epoch [2/3], Step [2544/3236], Loss: 1.9242, Perplexity: 6.8494
Epoch [2/3], Step [2545/3236], Loss: 2.0681, Perplexity: 7.9097
Epoch [2/3], Step [2546/3236], Loss: 2.6610, Perplexity: 14.3110
Epoch [2/3], Step [2547/3236], Loss: 2.0026, Perplexity: 7.4084
Epoch [2/3], Step [2548/3236], Loss: 1.9460, Perplexity: 7.0004
Epoch [2/3], Step [2549/3236], Loss: 2.1062, Perplexity: 8.2171
Epoch [2/3], Step [2550/3236], Loss: 1.9672, Perplexity: 7.1503
Epoch [2/3], Step [2551/3236], Loss: 2.0824, Perplexity: 8.0238
Epoch [2/3], Step [2552/3236], Loss: 2.2892, Perplexity: 9.8674
Epoch [2/3], Step [2553/3236], Loss: 1.8399, Perplexity: 6.2960
Epoch [2/3], Step [2554/3236], Loss: 1.9837, Perplexity: 7.2695
Epoch [2/3], Step [2555/3236], Loss: 2.1675, Perplexity: 8.7365
Epoch [2/3], Step [2556/3236], Loss: 1.9775, Perplexity: 7.2244
Epoch [2/3], Step [2557/3236], Loss: 2.0712, Perplexity: 7.9341
Epoch [2/3], Step [2558/3236], Loss: 2.3513, Perplexity: 10.4991
Epoch [2/3], Step [2559/3236], Loss: 1.9904, Perplexity: 7.3187
Epoch [2/3], Step [2560/3236], Loss: 2.3828, Perplexity: 10.8349
Epoch [2/3], Step [2561/3236], Loss: 2.1846, Perplexity: 8.8873
Epoch [2/3], Step [2562/3236], Loss: 1.9576, Perplexity: 7.0822
Epoch [2/3], Step [2563/3236], Loss: 2.0504, Perplexity: 7.7709
Epoch [2/3], Step [2564/3236], Loss: 1.8842, Perplexity: 6.5808
Epoch [2/3], Step [2565/3236], Loss: 1.9147, Perplexity: 6.7847
Epoch [2/3], Step [2566/3236], Loss: 1.9665, Perplexity: 7.1455
Epoch [2/3], Step [2567/3236], Loss: 2.0759, Perplexity: 7.9714
Epoch [2/3], Step [2568/3236], Loss: 2.0408, Perplexity: 7.6965
Epoch [2/3], Step [2569/3236], Loss: 2.2120, Perplexity: 9.1336
Epoch [2/3], Step [2570/3236], Loss: 1.8827, Perplexity: 6.5709
Epoch [2/3], Step [2571/3236], Loss: 2.0425, Perplexity: 7.7101
Epoch [2/3], Step [2572/3236], Loss: 1.9629, Perplexity: 7.1197
Epoch [2/3], Step [2573/3236], Loss: 2.1568, Perplexity: 8.6437
Epoch [2/3], Step [2574/3236], Loss: 1.8526, Perplexity: 6.3762
Epoch [2/3], Step [2575/3236], Loss: 2.3416, Perplexity: 10.3976
Epoch [2/3], Step [2576/3236], Loss: 2.2285, Perplexity: 9.2856
Epoch [2/3], Step [2577/3236], Loss: 1.9466, Perplexity: 7.0047
Epoch [2/3], Step [2578/3236], Loss: 1.9312, Perplexity: 6.8976
Epoch [2/3], Step [2579/3236], Loss: 1.8342, Perplexity: 6.2603
Epoch [2/3], Step [2580/3236], Loss: 2.1172, Perplexity: 8.3079
Epoch [2/3], Step [2581/3236], Loss: 2.2005, Perplexity: 9.0296
Epoch [2/3], Step [2582/3236], Loss: 2.0638, Perplexity: 7.8756
Epoch [2/3], Step [2583/3236], Loss: 2.0152, Perplexity: 7.5020
Epoch [2/3], Step [2584/3236], Loss: 2.2234, Perplexity: 9.2385
Epoch [2/3], Step [2585/3236], Loss: 2.1111, Perplexity: 8.2571
Epoch [2/3], Step [2586/3236], Loss: 2.4367, Perplexity: 11.4352
Epoch [2/3], Step [2587/3236], Loss: 1.9732, Perplexity: 7.1936
Epoch [2/3], Step [2588/3236], Loss: 2.5673, Perplexity: 13.0305
Epoch [2/3], Step [2589/3236], Loss: 2.3487, Perplexity: 10.4722
Epoch [2/3], Step [2590/3236], Loss: 1.9802, Perplexity: 7.2440
Epoch [2/3], Step [2591/3236], Loss: 1.9661, Perplexity: 7.1430
Epoch [2/3], Step [2592/3236], Loss: 1.8966, Perplexity: 6.6633
Epoch [2/3], Step [2593/3236], Loss: 2.1542, Perplexity: 8.6213
Epoch [2/3], Step [2594/3236], Loss: 2.1058, Perplexity: 8.2133
Epoch [2/3], Step [2595/3236], Loss: 2.3528, Perplexity: 10.5150
Epoch [2/3], Step [2596/3236], Loss: 2.0451, Perplexity: 7.7301
Epoch [2/3], Step [2597/3236], Loss: 1.9705, Perplexity: 7.1745
Epoch [2/3], Step [2598/3236], Loss: 2.4694, Perplexity: 11.8153
Epoch [2/3], Step [2599/3236], Loss: 2.0122, Perplexity: 7.4797
Epoch [2/3], Step [2600/3236], Loss: 1.9969, Perplexity: 7.3662
Epoch [2/3], Step [2601/3236], Loss: 1.9685, Perplexity: 7.1603
Epoch [2/3], Step [2602/3236], Loss: 2.0374, Perplexity: 7.6709
Epoch [2/3], Step [2603/3236], Loss: 2.0202, Perplexity: 7.5400
Epoch [2/3], Step [2604/3236], Loss: 2.1255, Perplexity: 8.3773
Epoch [2/3], Step [2605/3236], Loss: 2.1909, Perplexity: 8.9434
Epoch [2/3], Step [2606/3236], Loss: 1.9467, Perplexity: 7.0056
Epoch [2/3], Step [2607/3236], Loss: 1.9359, Perplexity: 6.9302
Epoch [2/3], Step [2608/3236], Loss: 2.2588, Perplexity: 9.5713
Epoch [2/3], Step [2609/3236], Loss: 2.1719, Perplexity: 8.7752
Epoch [2/3], Step [2610/3236], Loss: 2.0368, Perplexity: 7.6662
Epoch [2/3], Step [2611/3236], Loss: 2.1945, Perplexity: 8.9755
Epoch [2/3], Step [2612/3236], Loss: 2.0143, Perplexity: 7.4951
Epoch [2/3], Step [2613/3236], Loss: 1.9319, Perplexity: 6.9024
Epoch [2/3], Step [2614/3236], Loss: 1.9642, Perplexity: 7.1290
Epoch [2/3], Step [2615/3236], Loss: 1.9743, Perplexity: 7.2017
Epoch [2/3], Step [2616/3236], Loss: 1.9633, Perplexity: 7.1228
Epoch [2/3], Step [2617/3236], Loss: 1.9185, Perplexity: 6.8109
Epoch [2/3], Step [2618/3236], Loss: 2.0843, Perplexity: 8.0387
Epoch [2/3], Step [2619/3236], Loss: 1.9640, Perplexity: 7.1280
Epoch [2/3], Step [2620/3236], Loss: 2.3665, Perplexity: 10.6599
Epoch [2/3], Step [2621/3236], Loss: 1.9804, Perplexity: 7.2457
Epoch [2/3], Step [2622/3236], Loss: 1.8831, Perplexity: 6.5739
Epoch [2/3], Step [2623/3236], Loss: 2.4867, Perplexity: 12.0213
Epoch [2/3], Step [2624/3236], Loss: 2.0893, Perplexity: 8.0796
Epoch [2/3], Step [2625/3236], Loss: 1.8924, Perplexity: 6.6352
Epoch [2/3], Step [2626/3236], Loss: 2.0660, Perplexity: 7.8932
Epoch [2/3], Step [2627/3236], Loss: 2.1440, Perplexity: 8.5333
Epoch [2/3], Step [2628/3236], Loss: 2.0604, Perplexity: 7.8493
Epoch [2/3], Step [2629/3236], Loss: 2.0390, Perplexity: 7.6829
Epoch [2/3], Step [2630/3236], Loss: 2.1497, Perplexity: 8.5823
Epoch [2/3], Step [2631/3236], Loss: 2.0377, Perplexity: 7.6728
Epoch [2/3], Step [2632/3236], Loss: 2.1005, Perplexity: 8.1701
Epoch [2/3], Step [2633/3236], Loss: 1.8759, Perplexity: 6.5267
Epoch [2/3], Step [2634/3236], Loss: 2.0658, Perplexity: 7.8916
Epoch [2/3], Step [2635/3236], Loss: 2.0192, Perplexity: 7.5320
Epoch [2/3], Step [2636/3236], Loss: 2.5100, Perplexity: 12.3052
Epoch [2/3], Step [2637/3236], Loss: 1.8549, Perplexity: 6.3913
Epoch [2/3], Step [2638/3236], Loss: 1.9266, Perplexity: 6.8660
Epoch [2/3], Step [2639/3236], Loss: 2.3500, Perplexity: 10.4852
Epoch [2/3], Step [2640/3236], Loss: 2.0650, Perplexity: 7.8855
Epoch [2/3], Step [2641/3236], Loss: 2.0796, Perplexity: 8.0009
Epoch [2/3], Step [2642/3236], Loss: 2.0084, Perplexity: 7.4514
Epoch [2/3], Step [2643/3236], Loss: 2.8001, Perplexity: 16.4470
Epoch [2/3], Step [2644/3236], Loss: 1.8924, Perplexity: 6.6351
Epoch [2/3], Step [2645/3236], Loss: 1.9201, Perplexity: 6.8216
Epoch [2/3], Step [2646/3236], Loss: 2.8077, Perplexity: 16.5724
Epoch [2/3], Step [2647/3236], Loss: 2.1808, Perplexity: 8.8534
Epoch [2/3], Step [2648/3236], Loss: 2.2034, Perplexity: 9.0556
Epoch [2/3], Step [2649/3236], Loss: 2.0504, Perplexity: 7.7709
Epoch [2/3], Step [2650/3236], Loss: 2.0673, Perplexity: 7.9036
Epoch [2/3], Step [2651/3236], Loss: 2.2476, Perplexity: 9.4653
Epoch [2/3], Step [2652/3236], Loss: 1.9686, Perplexity: 7.1610
Epoch [2/3], Step [2653/3236], Loss: 2.3101, Perplexity: 10.0758
Epoch [2/3], Step [2654/3236], Loss: 2.2212, Perplexity: 9.2187
Epoch [2/3], Step [2655/3236], Loss: 2.0797, Perplexity: 8.0017
Epoch [2/3], Step [2656/3236], Loss: 2.1041, Perplexity: 8.2000
Epoch [2/3], Step [2657/3236], Loss: 2.0276, Perplexity: 7.5956
Epoch [2/3], Step [2658/3236], Loss: 2.2157, Perplexity: 9.1680
Epoch [2/3], Step [2659/3236], Loss: 1.8845, Perplexity: 6.5832
Epoch [2/3], Step [2660/3236], Loss: 2.0312, Perplexity: 7.6230
Epoch [2/3], Step [2661/3236], Loss: 2.1157, Perplexity: 8.2950
Epoch [2/3], Step [2662/3236], Loss: 2.1898, Perplexity: 8.9335
Epoch [2/3], Step [2663/3236], Loss: 2.0158, Perplexity: 7.5064
Epoch [2/3], Step [2664/3236], Loss: 1.9306, Perplexity: 6.8939
Epoch [2/3], Step [2665/3236], Loss: 2.0356, Perplexity: 7.6568
Epoch [2/3], Step [2666/3236], Loss: 1.9865, Perplexity: 7.2903
Epoch [2/3], Step [2667/3236], Loss: 2.0809, Perplexity: 8.0116
Epoch [2/3], Step [2668/3236], Loss: 1.9810, Perplexity: 7.2500
Epoch [2/3], Step [2669/3236], Loss: 1.9857, Perplexity: 7.2841
Epoch [2/3], Step [2670/3236], Loss: 2.0710, Perplexity: 7.9329
Epoch [2/3], Step [2671/3236], Loss: 1.9288, Perplexity: 6.8813
Epoch [2/3], Step [2672/3236], Loss: 1.9784, Perplexity: 7.2312
Epoch [2/3], Step [2673/3236], Loss: 1.9384, Perplexity: 6.9475
Epoch [2/3], Step [2674/3236], Loss: 2.1095, Perplexity: 8.2441
Epoch [2/3], Step [2675/3236], Loss: 2.8725, Perplexity: 17.6812
Epoch [2/3], Step [2676/3236], Loss: 1.9646, Perplexity: 7.1324
Epoch [2/3], Step [2677/3236], Loss: 2.0682, Perplexity: 7.9102
Epoch [2/3], Step [2678/3236], Loss: 2.0186, Perplexity: 7.5279
Epoch [2/3], Step [2679/3236], Loss: 2.5995, Perplexity: 13.4571
Epoch [2/3], Step [2680/3236], Loss: 2.3361, Perplexity: 10.3405
Epoch [2/3], Step [2681/3236], Loss: 2.0445, Perplexity: 7.7254
Epoch [2/3], Step [2682/3236], Loss: 2.4137, Perplexity: 11.1747
Epoch [2/3], Step [2683/3236], Loss: 2.6686, Perplexity: 14.4192
Epoch [2/3], Step [2684/3236], Loss: 2.0340, Perplexity: 7.6450
Epoch [2/3], Step [2685/3236], Loss: 2.0965, Perplexity: 8.1373
Epoch [2/3], Step [2686/3236], Loss: 1.9337, Perplexity: 6.9153
Epoch [2/3], Step [2687/3236], Loss: 2.1452, Perplexity: 8.5439
Epoch [2/3], Step [2688/3236], Loss: 2.0680, Perplexity: 7.9088
Epoch [2/3], Step [2689/3236], Loss: 2.5631, Perplexity: 12.9765
Epoch [2/3], Step [2690/3236], Loss: 2.1909, Perplexity: 8.9431
Epoch [2/3], Step [2691/3236], Loss: 2.0758, Perplexity: 7.9711
Epoch [2/3], Step [2692/3236], Loss: 2.1346, Perplexity: 8.4539
Epoch [2/3], Step [2693/3236], Loss: 2.2479, Perplexity: 9.4681
Epoch [2/3], Step [2694/3236], Loss: 2.1818, Perplexity: 8.8619
Epoch [2/3], Step [2695/3236], Loss: 1.9789, Perplexity: 7.2349
Epoch [2/3], Step [2696/3236], Loss: 2.1347, Perplexity: 8.4549
Epoch [2/3], Step [2697/3236], Loss: 2.0236, Perplexity: 7.5655
Epoch [2/3], Step [2698/3236], Loss: 2.0164, Perplexity: 7.5116
Epoch [2/3], Step [2699/3236], Loss: 2.0146, Perplexity: 7.4975
Epoch [2/3], Step [2700/3236], Loss: 2.0127, Perplexity: 7.4837
Epoch [2/3], Step [2701/3236], Loss: 1.9397, Perplexity: 6.9564
Epoch [2/3], Step [2702/3236], Loss: 2.4144, Perplexity: 11.1830
Epoch [2/3], Step [2703/3236], Loss: 2.1422, Perplexity: 8.5185
Epoch [2/3], Step [2704/3236], Loss: 2.1819, Perplexity: 8.8631
Epoch [2/3], Step [2705/3236], Loss: 2.3103, Perplexity: 10.0773
Epoch [2/3], Step [2706/3236], Loss: 2.0828, Perplexity: 8.0268
Epoch [2/3], Step [2707/3236], Loss: 1.9610, Perplexity: 7.1063
Epoch [2/3], Step [2708/3236], Loss: 2.0142, Perplexity: 7.4946
Epoch [2/3], Step [2709/3236], Loss: 1.9572, Perplexity: 7.0792
Epoch [2/3], Step [2710/3236], Loss: 1.9652, Perplexity: 7.1365
Epoch [2/3], Step [2711/3236], Loss: 1.9652, Perplexity: 7.1361
Epoch [2/3], Step [2712/3236], Loss: 2.0843, Perplexity: 8.0393
Epoch [2/3], Step [2713/3236], Loss: 1.9301, Perplexity: 6.8904
Epoch [2/3], Step [2714/3236], Loss: 1.8520, Perplexity: 6.3724
Epoch [2/3], Step [2715/3236], Loss: 2.1014, Perplexity: 8.1775
Epoch [2/3], Step [2716/3236], Loss: 1.9908, Perplexity: 7.3216
Epoch [2/3], Step [2717/3236], Loss: 1.9541, Perplexity: 7.0578
Epoch [2/3], Step [2718/3236], Loss: 2.0852, Perplexity: 8.0466
Epoch [2/3], Step [2719/3236], Loss: 2.0201, Perplexity: 7.5388
Epoch [2/3], Step [2720/3236], Loss: 2.0972, Perplexity: 8.1431
Epoch [2/3], Step [2721/3236], Loss: 1.9503, Perplexity: 7.0305
Epoch [2/3], Step [2722/3236], Loss: 2.4837, Perplexity: 11.9859
Epoch [2/3], Step [2723/3236], Loss: 2.0563, Perplexity: 7.8169
Epoch [2/3], Step [2724/3236], Loss: 1.9092, Perplexity: 6.7477
Epoch [2/3], Step [2725/3236], Loss: 1.9900, Perplexity: 7.3157
Epoch [2/3], Step [2726/3236], Loss: 1.9871, Perplexity: 7.2942
Epoch [2/3], Step [2727/3236], Loss: 2.0725, Perplexity: 7.9449
Epoch [2/3], Step [2728/3236], Loss: 2.0154, Perplexity: 7.5034
Epoch [2/3], Step [2729/3236], Loss: 2.0641, Perplexity: 7.8779
Epoch [2/3], Step [2730/3236], Loss: 2.0856, Perplexity: 8.0497
Epoch [2/3], Step [2731/3236], Loss: 2.1625, Perplexity: 8.6928
Epoch [2/3], Step [2732/3236], Loss: 2.1842, Perplexity: 8.8835
Epoch [2/3], Step [2733/3236], Loss: 2.1241, Perplexity: 8.3657
Epoch [2/3], Step [2734/3236], Loss: 2.1615, Perplexity: 8.6837
Epoch [2/3], Step [2735/3236], Loss: 2.3368, Perplexity: 10.3478
Epoch [2/3], Step [2736/3236], Loss: 1.9515, Perplexity: 7.0396
Epoch [2/3], Step [2737/3236], Loss: 2.0609, Perplexity: 7.8529
Epoch [2/3], Step [2738/3236], Loss: 1.9410, Perplexity: 6.9656
Epoch [2/3], Step [2739/3236], Loss: 1.9795, Perplexity: 7.2389
Epoch [2/3], Step [2740/3236], Loss: 1.9899, Perplexity: 7.3148
Epoch [2/3], Step [2741/3236], Loss: 2.4737, Perplexity: 11.8659
Epoch [2/3], Step [2742/3236], Loss: 2.1680, Perplexity: 8.7406
Epoch [2/3], Step [2743/3236], Loss: 1.9707, Perplexity: 7.1759
Epoch [2/3], Step [2744/3236], Loss: 2.0078, Perplexity: 7.4470
Epoch [2/3], Step [2745/3236], Loss: 1.9848, Perplexity: 7.2773
Epoch [2/3], Step [2746/3236], Loss: 2.2274, Perplexity: 9.2762
Epoch [2/3], Step [2747/3236], Loss: 2.1626, Perplexity: 8.6938
Epoch [2/3], Step [2748/3236], Loss: 2.0274, Perplexity: 7.5944
Epoch [2/3], Step [2749/3236], Loss: 2.0866, Perplexity: 8.0577
Epoch [2/3], Step [2750/3236], Loss: 2.0576, Perplexity: 7.8269
Epoch [2/3], Step [2751/3236], Loss: 2.2598, Perplexity: 9.5815
Epoch [2/3], Step [2752/3236], Loss: 2.0016, Perplexity: 7.4008
Epoch [2/3], Step [2753/3236], Loss: 2.5688, Perplexity: 13.0496
Epoch [2/3], Step [2754/3236], Loss: 2.1067, Perplexity: 8.2209
Epoch [2/3], Step [2755/3236], Loss: 2.4594, Perplexity: 11.6978
Epoch [2/3], Step [2756/3236], Loss: 1.9287, Perplexity: 6.8803
Epoch [2/3], Step [2757/3236], Loss: 2.3641, Perplexity: 10.6350
Epoch [2/3], Step [2758/3236], Loss: 2.1051, Perplexity: 8.2079
Epoch [2/3], Step [2759/3236], Loss: 2.0241, Perplexity: 7.5693
Epoch [2/3], Step [2760/3236], Loss: 2.0577, Perplexity: 7.8279
Epoch [2/3], Step [2761/3236], Loss: 2.0968, Perplexity: 8.1405
Epoch [2/3], Step [2762/3236], Loss: 1.9473, Perplexity: 7.0095
Epoch [2/3], Step [2763/3236], Loss: 1.9936, Perplexity: 7.3417
Epoch [2/3], Step [2764/3236], Loss: 2.0544, Perplexity: 7.8019
Epoch [2/3], Step [2765/3236], Loss: 2.3792, Perplexity: 10.7965
Epoch [2/3], Step [2766/3236], Loss: 1.9307, Perplexity: 6.8944
Epoch [2/3], Step [2767/3236], Loss: 2.1036, Perplexity: 8.1953
Epoch [2/3], Step [2768/3236], Loss: 2.0762, Perplexity: 7.9738
Epoch [2/3], Step [2769/3236], Loss: 2.1030, Perplexity: 8.1906
Epoch [2/3], Step [2770/3236], Loss: 2.2243, Perplexity: 9.2471
Epoch [2/3], Step [2771/3236], Loss: 2.0270, Perplexity: 7.5914
Epoch [2/3], Step [2772/3236], Loss: 2.0434, Perplexity: 7.7168
Epoch [2/3], Step [2773/3236], Loss: 1.9677, Perplexity: 7.1545
Epoch [2/3], Step [2774/3236], Loss: 2.0419, Perplexity: 7.7052
Epoch [2/3], Step [2775/3236], Loss: 2.1169, Perplexity: 8.3052
Epoch [2/3], Step [2776/3236], Loss: 2.0598, Perplexity: 7.8440
Epoch [2/3], Step [2777/3236], Loss: 1.9607, Perplexity: 7.1044
Epoch [2/3], Step [2778/3236], Loss: 2.2795, Perplexity: 9.7722
Epoch [2/3], Step [2779/3236], Loss: 2.0999, Perplexity: 8.1657
Epoch [2/3], Step [2780/3236], Loss: 2.0102, Perplexity: 7.4651
Epoch [2/3], Step [2781/3236], Loss: 2.0269, Perplexity: 7.5908
Epoch [2/3], Step [2782/3236], Loss: 1.8927, Perplexity: 6.6371
Epoch [2/3], Step [2783/3236], Loss: 2.0508, Perplexity: 7.7744
Epoch [2/3], Step [2784/3236], Loss: 2.0321, Perplexity: 7.6297
Epoch [2/3], Step [2785/3236], Loss: 2.0943, Perplexity: 8.1196
Epoch [2/3], Step [2786/3236], Loss: 1.9725, Perplexity: 7.1888
Epoch [2/3], Step [2787/3236], Loss: 2.1815, Perplexity: 8.8596
Epoch [2/3], Step [2788/3236], Loss: 1.9399, Perplexity: 6.9582
Epoch [2/3], Step [2789/3236], Loss: 1.8716, Perplexity: 6.4986
Epoch [2/3], Step [2790/3236], Loss: 2.0099, Perplexity: 7.4625
Epoch [2/3], Step [2791/3236], Loss: 1.8678, Perplexity: 6.4738
Epoch [2/3], Step [2792/3236], Loss: 2.2697, Perplexity: 9.6764
Epoch [2/3], Step [2793/3236], Loss: 2.0166, Perplexity: 7.5127
Epoch [2/3], Step [2794/3236], Loss: 2.3444, Perplexity: 10.4272
Epoch [2/3], Step [2795/3236], Loss: 2.0020, Perplexity: 7.4036
Epoch [2/3], Step [2796/3236], Loss: 1.9907, Perplexity: 7.3205
Epoch [2/3], Step [2797/3236], Loss: 1.9740, Perplexity: 7.1992
Epoch [2/3], Step [2798/3236], Loss: 1.9333, Perplexity: 6.9119
Epoch [2/3], Step [2799/3236], Loss: 1.9627, Perplexity: 7.1183
Epoch [2/3], Step [2800/3236], Loss: 2.0778, Perplexity: 7.9870
Epoch [2/3], Step [2801/3236], Loss: 2.2460, Perplexity: 9.4495
Epoch [2/3], Step [2802/3236], Loss: 2.0251, Perplexity: 7.5766
Epoch [2/3], Step [2803/3236], Loss: 1.8511, Perplexity: 6.3668
Epoch [2/3], Step [2804/3236], Loss: 2.0804, Perplexity: 8.0077
Epoch [2/3], Step [2805/3236], Loss: 1.9935, Perplexity: 7.3413
Epoch [2/3], Step [2806/3236], Loss: 2.0907, Perplexity: 8.0905
Epoch [2/3], Step [2807/3236], Loss: 2.0305, Perplexity: 7.6177
Epoch [2/3], Step [2808/3236], Loss: 1.9795, Perplexity: 7.2394
Epoch [2/3], Step [2809/3236], Loss: 2.0494, Perplexity: 7.7635
Epoch [2/3], Step [2810/3236], Loss: 2.0551, Perplexity: 7.8074
Epoch [2/3], Step [2811/3236], Loss: 2.0606, Perplexity: 7.8508
Epoch [2/3], Step [2812/3236], Loss: 2.0342, Perplexity: 7.6460
Epoch [2/3], Step [2813/3236], Loss: 2.0708, Perplexity: 7.9314
Epoch [2/3], Step [2814/3236], Loss: 2.0770, Perplexity: 7.9805
Epoch [2/3], Step [2815/3236], Loss: 2.1787, Perplexity: 8.8345
Epoch [2/3], Step [2816/3236], Loss: 2.2879, Perplexity: 9.8545
Epoch [2/3], Step [2817/3236], Loss: 2.4361, Perplexity: 11.4284
Epoch [2/3], Step [2818/3236], Loss: 2.0427, Perplexity: 7.7117
Epoch [2/3], Step [2819/3236], Loss: 2.7139, Perplexity: 15.0887
Epoch [2/3], Step [2820/3236], Loss: 1.8921, Perplexity: 6.6333
Epoch [2/3], Step [2821/3236], Loss: 1.9447, Perplexity: 6.9917
Epoch [2/3], Step [2822/3236], Loss: 2.0583, Perplexity: 7.8326
Epoch [2/3], Step [2823/3236], Loss: 2.5997, Perplexity: 13.4602
Epoch [2/3], Step [2824/3236], Loss: 2.0609, Perplexity: 7.8527
Epoch [2/3], Step [2825/3236], Loss: 2.5861, Perplexity: 13.2775
Epoch [2/3], Step [2826/3236], Loss: 1.9701, Perplexity: 7.1712
Epoch [2/3], Step [2827/3236], Loss: 2.0866, Perplexity: 8.0577
Epoch [2/3], Step [2828/3236], Loss: 1.9284, Perplexity: 6.8788
Epoch [2/3], Step [2829/3236], Loss: 2.0659, Perplexity: 7.8923
Epoch [2/3], Step [2830/3236], Loss: 2.1657, Perplexity: 8.7204
Epoch [2/3], Step [2831/3236], Loss: 1.9890, Perplexity: 7.3083
Epoch [2/3], Step [2832/3236], Loss: 2.2597, Perplexity: 9.5800
Epoch [2/3], Step [2833/3236], Loss: 1.8467, Perplexity: 6.3392
Epoch [2/3], Step [2834/3236], Loss: 2.4286, Perplexity: 11.3436
Epoch [2/3], Step [2835/3236], Loss: 1.8330, Perplexity: 6.2524
Epoch [2/3], Step [2836/3236], Loss: 2.1334, Perplexity: 8.4435
Epoch [2/3], Step [2837/3236], Loss: 2.2324, Perplexity: 9.3222
Epoch [2/3], Step [2838/3236], Loss: 1.9436, Perplexity: 6.9841
Epoch [2/3], Step [2839/3236], Loss: 1.9960, Perplexity: 7.3592
Epoch [2/3], Step [2840/3236], Loss: 1.9190, Perplexity: 6.8139
Epoch [2/3], Step [2841/3236], Loss: 1.9595, Perplexity: 7.0960
Epoch [2/3], Step [2842/3236], Loss: 2.1281, Perplexity: 8.3992
Epoch [2/3], Step [2843/3236], Loss: 2.1874, Perplexity: 8.9121
Epoch [2/3], Step [2844/3236], Loss: 2.3573, Perplexity: 10.5629
Epoch [2/3], Step [2845/3236], Loss: 1.9954, Perplexity: 7.3550
Epoch [2/3], Step [2846/3236], Loss: 1.8671, Perplexity: 6.4693
Epoch [2/3], Step [2847/3236], Loss: 2.0148, Perplexity: 7.4991
Epoch [2/3], Step [2848/3236], Loss: 1.9395, Perplexity: 6.9552
Epoch [2/3], Step [2849/3236], Loss: 1.9787, Perplexity: 7.2330
Epoch [2/3], Step [2850/3236], Loss: 2.5297, Perplexity: 12.5496
Epoch [2/3], Step [2851/3236], Loss: 2.0403, Perplexity: 7.6930
Epoch [2/3], Step [2852/3236], Loss: 2.1033, Perplexity: 8.1928
Epoch [2/3], Step [2853/3236], Loss: 2.1252, Perplexity: 8.3746
Epoch [2/3], Step [2854/3236], Loss: 2.1339, Perplexity: 8.4478
Epoch [2/3], Step [2855/3236], Loss: 2.5275, Perplexity: 12.5219
Epoch [2/3], Step [2856/3236], Loss: 2.0717, Perplexity: 7.9387
Epoch [2/3], Step [2857/3236], Loss: 1.9164, Perplexity: 6.7968
Epoch [2/3], Step [2858/3236], Loss: 2.4027, Perplexity: 11.0530
Epoch [2/3], Step [2859/3236], Loss: 1.9158, Perplexity: 6.7923
Epoch [2/3], Step [2860/3236], Loss: 1.9665, Perplexity: 7.1453
Epoch [2/3], Step [2861/3236], Loss: 2.0759, Perplexity: 7.9715
Epoch [2/3], Step [2862/3236], Loss: 2.0386, Perplexity: 7.6799
Epoch [2/3], Step [2863/3236], Loss: 2.0856, Perplexity: 8.0494
Epoch [2/3], Step [2864/3236], Loss: 2.0228, Perplexity: 7.5593
Epoch [2/3], Step [2865/3236], Loss: 1.9956, Perplexity: 7.3564
Epoch [2/3], Step [2866/3236], Loss: 2.0547, Perplexity: 7.8046
Epoch [2/3], Step [2867/3236], Loss: 1.9125, Perplexity: 6.7702
Epoch [2/3], Step [2868/3236], Loss: 2.0323, Perplexity: 7.6318
Epoch [2/3], Step [2869/3236], Loss: 2.0273, Perplexity: 7.5937
Epoch [2/3], Step [2870/3236], Loss: 1.9401, Perplexity: 6.9594
Epoch [2/3], Step [2871/3236], Loss: 2.2846, Perplexity: 9.8222
Epoch [2/3], Step [2872/3236], Loss: 2.2880, Perplexity: 9.8555
Epoch [2/3], Step [2873/3236], Loss: 2.2909, Perplexity: 9.8843
Epoch [2/3], Step [2874/3236], Loss: 2.4090, Perplexity: 11.1224
Epoch [2/3], Step [2875/3236], Loss: 2.0246, Perplexity: 7.5730
Epoch [2/3], Step [2876/3236], Loss: 1.7922, Perplexity: 6.0024
Epoch [2/3], Step [2877/3236], Loss: 2.2520, Perplexity: 9.5068
Epoch [2/3], Step [2878/3236], Loss: 1.9265, Perplexity: 6.8654
Epoch [2/3], Step [2879/3236], Loss: 2.1120, Perplexity: 8.2646
Epoch [2/3], Step [2880/3236], Loss: 1.9413, Perplexity: 6.9679
Epoch [2/3], Step [2881/3236], Loss: 2.0012, Perplexity: 7.3980
Epoch [2/3], Step [2882/3236], Loss: 2.2050, Perplexity: 9.0704
Epoch [2/3], Step [2883/3236], Loss: 1.9856, Perplexity: 7.2833
Epoch [2/3], Step [2884/3236], Loss: 1.9545, Perplexity: 7.0607
Epoch [2/3], Step [2885/3236], Loss: 2.1039, Perplexity: 8.1981
Epoch [2/3], Step [2886/3236], Loss: 2.0697, Perplexity: 7.9221
Epoch [2/3], Step [2887/3236], Loss: 2.2166, Perplexity: 9.1762
Epoch [2/3], Step [2888/3236], Loss: 1.8929, Perplexity: 6.6383
Epoch [2/3], Step [2889/3236], Loss: 2.0798, Perplexity: 8.0031
Epoch [2/3], Step [2890/3236], Loss: 1.9294, Perplexity: 6.8855
Epoch [2/3], Step [2891/3236], Loss: 2.1926, Perplexity: 8.9584
Epoch [2/3], Step [2892/3236], Loss: 2.2960, Perplexity: 9.9344
Epoch [2/3], Step [2893/3236], Loss: 1.9522, Perplexity: 7.0439
Epoch [2/3], Step [2894/3236], Loss: 2.0224, Perplexity: 7.5563
Epoch [2/3], Step [2895/3236], Loss: 1.9961, Perplexity: 7.3600
Epoch [2/3], Step [2896/3236], Loss: 1.9562, Perplexity: 7.0723
Epoch [2/3], Step [2897/3236], Loss: 2.0913, Perplexity: 8.0951
Epoch [2/3], Step [2898/3236], Loss: 2.3091, Perplexity: 10.0652
Epoch [2/3], Step [2899/3236], Loss: 1.9517, Perplexity: 7.0407
Epoch [2/3], Step [2900/3236], Loss: 2.2498, Perplexity: 9.4855
Epoch [2/3], Step [2901/3236], Loss: 2.1037, Perplexity: 8.1962
Epoch [2/3], Step [2902/3236], Loss: 1.8941, Perplexity: 6.6468
Epoch [2/3], Step [2903/3236], Loss: 1.8725, Perplexity: 6.5043
Epoch [2/3], Step [2904/3236], Loss: 2.0688, Perplexity: 7.9151
Epoch [2/3], Step [2905/3236], Loss: 2.4069, Perplexity: 11.0991
Epoch [2/3], Step [2906/3236], Loss: 1.9672, Perplexity: 7.1509
Epoch [2/3], Step [2907/3236], Loss: 2.4353, Perplexity: 11.4193
Epoch [2/3], Step [2908/3236], Loss: 2.0351, Perplexity: 7.6534
Epoch [2/3], Step [2909/3236], Loss: 1.9193, Perplexity: 6.8159
Epoch [2/3], Step [2910/3236], Loss: 2.3025, Perplexity: 9.9987
Epoch [2/3], Step [2911/3236], Loss: 2.0617, Perplexity: 7.8596
Epoch [2/3], Step [2912/3236], Loss: 2.0845, Perplexity: 8.0407
Epoch [2/3], Step [2913/3236], Loss: 1.9062, Perplexity: 6.7275
Epoch [2/3], Step [2914/3236], Loss: 1.9755, Perplexity: 7.2103
Epoch [2/3], Step [2915/3236], Loss: 2.0224, Perplexity: 7.5568
Epoch [2/3], Step [2916/3236], Loss: 1.9355, Perplexity: 6.9273
Epoch [2/3], Step [2917/3236], Loss: 2.0295, Perplexity: 7.6105
Epoch [2/3], Step [2918/3236], Loss: 1.9142, Perplexity: 6.7812
Epoch [2/3], Step [2919/3236], Loss: 2.0258, Perplexity: 7.5825
Epoch [2/3], Step [2920/3236], Loss: 2.0315, Perplexity: 7.6252
Epoch [2/3], Step [2921/3236], Loss: 2.0729, Perplexity: 7.9475
Epoch [2/3], Step [2922/3236], Loss: 2.1067, Perplexity: 8.2214
Epoch [2/3], Step [2923/3236], Loss: 1.9407, Perplexity: 6.9639
Epoch [2/3], Step [2924/3236], Loss: 2.0755, Perplexity: 7.9686
Epoch [2/3], Step [2925/3236], Loss: 1.8735, Perplexity: 6.5108
Epoch [2/3], Step [2926/3236], Loss: 1.9296, Perplexity: 6.8865
Epoch [2/3], Step [2927/3236], Loss: 1.9736, Perplexity: 7.1964
Epoch [2/3], Step [2928/3236], Loss: 2.0610, Perplexity: 7.8535
Epoch [2/3], Step [2929/3236], Loss: 1.8735, Perplexity: 6.5113
Epoch [2/3], Step [2930/3236], Loss: 2.1899, Perplexity: 8.9345
Epoch [2/3], Step [2931/3236], Loss: 2.0038, Perplexity: 7.4171
Epoch [2/3], Step [2932/3236], Loss: 1.9839, Perplexity: 7.2709
Epoch [2/3], Step [2933/3236], Loss: 1.9835, Perplexity: 7.2680
Epoch [2/3], Step [2934/3236], Loss: 1.8681, Perplexity: 6.4757
Epoch [2/3], Step [2935/3236], Loss: 2.1085, Perplexity: 8.2363
Epoch [2/3], Step [2936/3236], Loss: 2.6172, Perplexity: 13.6970
Epoch [2/3], Step [2937/3236], Loss: 2.9900, Perplexity: 19.8853
Epoch [2/3], Step [2938/3236], Loss: 2.2224, Perplexity: 9.2297
Epoch [2/3], Step [2939/3236], Loss: 2.0769, Perplexity: 7.9795
Epoch [2/3], Step [2940/3236], Loss: 2.5616, Perplexity: 12.9562
Epoch [2/3], Step [2941/3236], Loss: 2.4185, Perplexity: 11.2285
Epoch [2/3], Step [2942/3236], Loss: 2.0833, Perplexity: 8.0309
Epoch [2/3], Step [2943/3236], Loss: 1.8534, Perplexity: 6.3816
Epoch [2/3], Step [2944/3236], Loss: 1.9886, Perplexity: 7.3055
Epoch [2/3], Step [2945/3236], Loss: 2.0868, Perplexity: 8.0595
Epoch [2/3], Step [2946/3236], Loss: 2.4572, Perplexity: 11.6716
Epoch [2/3], Step [2947/3236], Loss: 2.0100, Perplexity: 7.4630
Epoch [2/3], Step [2948/3236], Loss: 2.0804, Perplexity: 8.0073
Epoch [2/3], Step [2949/3236], Loss: 1.8377, Perplexity: 6.2823
Epoch [2/3], Step [2950/3236], Loss: 2.0163, Perplexity: 7.5107
Epoch [2/3], Step [2951/3236], Loss: 2.1764, Perplexity: 8.8145
Epoch [2/3], Step [2952/3236], Loss: 1.9245, Perplexity: 6.8518
Epoch [2/3], Step [2953/3236], Loss: 2.2986, Perplexity: 9.9606
Epoch [2/3], Step [2954/3236], Loss: 2.2897, Perplexity: 9.8724
Epoch [2/3], Step [2955/3236], Loss: 2.1846, Perplexity: 8.8874
Epoch [2/3], Step [2956/3236], Loss: 2.0432, Perplexity: 7.7150
Epoch [2/3], Step [2957/3236], Loss: 1.9348, Perplexity: 6.9228
Epoch [2/3], Step [2958/3236], Loss: 2.3845, Perplexity: 10.8534
Epoch [2/3], Step [2959/3236], Loss: 2.0106, Perplexity: 7.4677
Epoch [2/3], Step [2960/3236], Loss: 1.9588, Perplexity: 7.0909
Epoch [2/3], Step [2961/3236], Loss: 2.3264, Perplexity: 10.2414
Epoch [2/3], Step [2962/3236], Loss: 1.9615, Perplexity: 7.1103
Epoch [2/3], Step [2963/3236], Loss: 1.9396, Perplexity: 6.9558
Epoch [2/3], Step [2964/3236], Loss: 2.2775, Perplexity: 9.7527
Epoch [2/3], Step [2965/3236], Loss: 2.2138, Perplexity: 9.1506
Epoch [2/3], Step [2966/3236], Loss: 2.0535, Perplexity: 7.7951
Epoch [2/3], Step [2967/3236], Loss: 2.0866, Perplexity: 8.0578
Epoch [2/3], Step [2968/3236], Loss: 2.0514, Perplexity: 7.7792
Epoch [2/3], Step [2969/3236], Loss: 1.9882, Perplexity: 7.3026
Epoch [2/3], Step [2970/3236], Loss: 2.0789, Perplexity: 7.9953
Epoch [2/3], Step [2971/3236], Loss: 2.0195, Perplexity: 7.5347
Epoch [2/3], Step [2972/3236], Loss: 2.0085, Perplexity: 7.4523
Epoch [2/3], Step [2973/3236], Loss: 1.9985, Perplexity: 7.3783
Epoch [2/3], Step [2974/3236], Loss: 1.9801, Perplexity: 7.2432
Epoch [2/3], Step [2975/3236], Loss: 2.1464, Perplexity: 8.5538
Epoch [2/3], Step [2976/3236], Loss: 1.9595, Perplexity: 7.0958
Epoch [2/3], Step [2977/3236], Loss: 1.9999, Perplexity: 7.3886
Epoch [2/3], Step [2978/3236], Loss: 2.0626, Perplexity: 7.8668
Epoch [2/3], Step [2979/3236], Loss: 2.0458, Perplexity: 7.7355
Epoch [2/3], Step [2980/3236], Loss: 2.0170, Perplexity: 7.5158
Epoch [2/3], Step [2981/3236], Loss: 2.0056, Perplexity: 7.4303
Epoch [2/3], Step [2982/3236], Loss: 2.1801, Perplexity: 8.8469
Epoch [2/3], Step [2983/3236], Loss: 2.2544, Perplexity: 9.5298
Epoch [2/3], Step [2984/3236], Loss: 2.1933, Perplexity: 8.9651
Epoch [2/3], Step [2985/3236], Loss: 1.9525, Perplexity: 7.0465
Epoch [2/3], Step [2986/3236], Loss: 1.9081, Perplexity: 6.7405
Epoch [2/3], Step [2987/3236], Loss: 2.0034, Perplexity: 7.4145
Epoch [2/3], Step [2988/3236], Loss: 3.1195, Perplexity: 22.6342
Epoch [2/3], Step [2989/3236], Loss: 1.9895, Perplexity: 7.3121
Epoch [2/3], Step [2990/3236], Loss: 1.8364, Perplexity: 6.2742
Epoch [2/3], Step [2991/3236], Loss: 2.4043, Perplexity: 11.0705
Epoch [2/3], Step [2992/3236], Loss: 2.2624, Perplexity: 9.6057
Epoch [2/3], Step [2993/3236], Loss: 1.9752, Perplexity: 7.2080
Epoch [2/3], Step [2994/3236], Loss: 1.9885, Perplexity: 7.3043
Epoch [2/3], Step [2995/3236], Loss: 1.9587, Perplexity: 7.0904
Epoch [2/3], Step [2996/3236], Loss: 2.0952, Perplexity: 8.1274
Epoch [2/3], Step [2997/3236], Loss: 1.9178, Perplexity: 6.8061
Epoch [2/3], Step [2998/3236], Loss: 2.0047, Perplexity: 7.4238
Epoch [2/3], Step [2999/3236], Loss: 1.8771, Perplexity: 6.5347
Epoch [2/3], Step [3000/3236], Loss: 2.3928, Perplexity: 10.9444
Epoch [2/3], Step [3001/3236], Loss: 2.0266, Perplexity: 7.5880
Epoch [2/3], Step [3002/3236], Loss: 1.9565, Perplexity: 7.0745
Epoch [2/3], Step [3003/3236], Loss: 1.9685, Perplexity: 7.1601
Epoch [2/3], Step [3004/3236], Loss: 2.0735, Perplexity: 7.9524
Epoch [2/3], Step [3005/3236], Loss: 2.2602, Perplexity: 9.5849
Epoch [2/3], Step [3006/3236], Loss: 2.4192, Perplexity: 11.2374
Epoch [2/3], Step [3007/3236], Loss: 1.9456, Perplexity: 6.9977
Epoch [2/3], Step [3008/3236], Loss: 2.0217, Perplexity: 7.5515
Epoch [2/3], Step [3009/3236], Loss: 1.9933, Perplexity: 7.3398
Epoch [2/3], Step [3010/3236], Loss: 1.9544, Perplexity: 7.0594
Epoch [2/3], Step [3011/3236], Loss: 2.0992, Perplexity: 8.1600
Epoch [2/3], Step [3012/3236], Loss: 1.8142, Perplexity: 6.1362
Epoch [2/3], Step [3013/3236], Loss: 1.9406, Perplexity: 6.9627
Epoch [2/3], Step [3014/3236], Loss: 2.0834, Perplexity: 8.0320
Epoch [2/3], Step [3015/3236], Loss: 1.9456, Perplexity: 6.9976
Epoch [2/3], Step [3016/3236], Loss: 2.0722, Perplexity: 7.9425
Epoch [2/3], Step [3017/3236], Loss: 2.0949, Perplexity: 8.1248
Epoch [2/3], Step [3018/3236], Loss: 1.9312, Perplexity: 6.8980
Epoch [2/3], Step [3019/3236], Loss: 1.9891, Perplexity: 7.3092
Epoch [2/3], Step [3020/3236], Loss: 1.9422, Perplexity: 6.9744
Epoch [2/3], Step [3021/3236], Loss: 1.8301, Perplexity: 6.2348
Epoch [2/3], Step [3022/3236], Loss: 2.1954, Perplexity: 8.9837
Epoch [2/3], Step [3023/3236], Loss: 2.0999, Perplexity: 8.1651
Epoch [2/3], Step [3024/3236], Loss: 2.2936, Perplexity: 9.9103
Epoch [2/3], Step [3025/3236], Loss: 1.9353, Perplexity: 6.9265
Epoch [2/3], Step [3026/3236], Loss: 2.3184, Perplexity: 10.1590
Epoch [2/3], Step [3027/3236], Loss: 2.1139, Perplexity: 8.2805
Epoch [2/3], Step [3028/3236], Loss: 1.9807, Perplexity: 7.2480
Epoch [2/3], Step [3029/3236], Loss: 1.9446, Perplexity: 6.9908
Epoch [2/3], Step [3030/3236], Loss: 2.0082, Perplexity: 7.4498
Epoch [2/3], Step [3031/3236], Loss: 2.2033, Perplexity: 9.0550
Epoch [2/3], Step [3032/3236], Loss: 1.9884, Perplexity: 7.3038
Epoch [2/3], Step [3033/3236], Loss: 2.0235, Perplexity: 7.5645
Epoch [2/3], Step [3034/3236], Loss: 2.0288, Perplexity: 7.6047
Epoch [2/3], Step [3035/3236], Loss: 2.0379, Perplexity: 7.6742
Epoch [2/3], Step [3036/3236], Loss: 2.0118, Perplexity: 7.4770
Epoch [2/3], Step [3037/3236], Loss: 2.2472, Perplexity: 9.4617
Epoch [2/3], Step [3038/3236], Loss: 2.0327, Perplexity: 7.6344
Epoch [2/3], Step [3039/3236], Loss: 1.9933, Perplexity: 7.3398
Epoch [2/3], Step [3040/3236], Loss: 2.0144, Perplexity: 7.4961
Epoch [2/3], Step [3041/3236], Loss: 2.2922, Perplexity: 9.8970
Epoch [2/3], Step [3042/3236], Loss: 2.5314, Perplexity: 12.5708
Epoch [2/3], Step [3043/3236], Loss: 1.9374, Perplexity: 6.9406
Epoch [2/3], Step [3044/3236], Loss: 2.1069, Perplexity: 8.2229
Epoch [2/3], Step [3045/3236], Loss: 1.9688, Perplexity: 7.1618
Epoch [2/3], Step [3046/3236], Loss: 1.9225, Perplexity: 6.8381
Epoch [2/3], Step [3047/3236], Loss: 2.0340, Perplexity: 7.6446
Epoch [2/3], Step [3048/3236], Loss: 1.9935, Perplexity: 7.3415
Epoch [2/3], Step [3049/3236], Loss: 1.9303, Perplexity: 6.8919
Epoch [2/3], Step [3050/3236], Loss: 2.0660, Perplexity: 7.8933
Epoch [2/3], Step [3051/3236], Loss: 2.0171, Perplexity: 7.5166
Epoch [2/3], Step [3052/3236], Loss: 1.9756, Perplexity: 7.2108
Epoch [2/3], Step [3053/3236], Loss: 2.4351, Perplexity: 11.4169
Epoch [2/3], Step [3054/3236], Loss: 2.0912, Perplexity: 8.0949
Epoch [2/3], Step [3055/3236], Loss: 1.9102, Perplexity: 6.7542
Epoch [2/3], Step [3056/3236], Loss: 2.1991, Perplexity: 9.0170
Epoch [2/3], Step [3057/3236], Loss: 2.0604, Perplexity: 7.8493
Epoch [2/3], Step [3058/3236], Loss: 1.8966, Perplexity: 6.6634
Epoch [2/3], Step [3059/3236], Loss: 1.9962, Perplexity: 7.3614
Epoch [2/3], Step [3060/3236], Loss: 2.3310, Perplexity: 10.2886
Epoch [2/3], Step [3061/3236], Loss: 2.0196, Perplexity: 7.5355
Epoch [2/3], Step [3062/3236], Loss: 1.9538, Perplexity: 7.0556
Epoch [2/3], Step [3063/3236], Loss: 2.3218, Perplexity: 10.1939
Epoch [2/3], Step [3064/3236], Loss: 1.9777, Perplexity: 7.2264
Epoch [2/3], Step [3065/3236], Loss: 2.0275, Perplexity: 7.5949
Epoch [2/3], Step [3066/3236], Loss: 2.0087, Perplexity: 7.4534
Epoch [2/3], Step [3067/3236], Loss: 1.9858, Perplexity: 7.2852
Epoch [2/3], Step [3068/3236], Loss: 1.8513, Perplexity: 6.3678
Epoch [2/3], Step [3069/3236], Loss: 2.1349, Perplexity: 8.4561
Epoch [2/3], Step [3070/3236], Loss: 1.9711, Perplexity: 7.1785
Epoch [2/3], Step [3071/3236], Loss: 1.9002, Perplexity: 6.6873
Epoch [2/3], Step [3072/3236], Loss: 2.0269, Perplexity: 7.5909
Epoch [2/3], Step [3073/3236], Loss: 2.0492, Perplexity: 7.7615
Epoch [2/3], Step [3074/3236], Loss: 2.0530, Perplexity: 7.7911
Epoch [2/3], Step [3075/3236], Loss: 2.0048, Perplexity: 7.4243
Epoch [2/3], Step [3076/3236], Loss: 1.8800, Perplexity: 6.5534
Epoch [2/3], Step [3077/3236], Loss: 1.9422, Perplexity: 6.9742
Epoch [2/3], Step [3078/3236], Loss: 1.9850, Perplexity: 7.2791
Epoch [2/3], Step [3079/3236], Loss: 2.1733, Perplexity: 8.7869
Epoch [2/3], Step [3080/3236], Loss: 2.5702, Perplexity: 13.0681
Epoch [2/3], Step [3081/3236], Loss: 1.9164, Perplexity: 6.7964
Epoch [2/3], Step [3082/3236], Loss: 1.9762, Perplexity: 7.2149
Epoch [2/3], Step [3083/3236], Loss: 2.1586, Perplexity: 8.6592
Epoch [2/3], Step [3084/3236], Loss: 2.2566, Perplexity: 9.5501
Epoch [2/3], Step [3085/3236], Loss: 2.4824, Perplexity: 11.9703
Epoch [2/3], Step [3086/3236], Loss: 2.8620, Perplexity: 17.4968
Epoch [2/3], Step [3087/3236], Loss: 2.0296, Perplexity: 7.6110
Epoch [2/3], Step [3088/3236], Loss: 2.0617, Perplexity: 7.8590
Epoch [2/3], Step [3089/3236], Loss: 1.9271, Perplexity: 6.8697
Epoch [2/3], Step [3090/3236], Loss: 2.0271, Perplexity: 7.5922
Epoch [2/3], Step [3091/3236], Loss: 2.0650, Perplexity: 7.8850
Epoch [2/3], Step [3092/3236], Loss: 2.0341, Perplexity: 7.6453
Epoch [2/3], Step [3093/3236], Loss: 2.4177, Perplexity: 11.2203
Epoch [2/3], Step [3094/3236], Loss: 2.1229, Perplexity: 8.3550
Epoch [2/3], Step [3095/3236], Loss: 2.0710, Perplexity: 7.9324
Epoch [2/3], Step [3096/3236], Loss: 2.0644, Perplexity: 7.8806
Epoch [2/3], Step [3097/3236], Loss: 2.0654, Perplexity: 7.8886
Epoch [2/3], Step [3098/3236], Loss: 2.1876, Perplexity: 8.9137
Epoch [2/3], Step [3099/3236], Loss: 2.3286, Perplexity: 10.2634
Epoch [2/3], Step [3100/3236], Loss: 1.9427, Perplexity: 6.9778
Epoch [2/3], Step [3101/3236], Loss: 2.1399, Perplexity: 8.4986
Epoch [2/3], Step [3102/3236], Loss: 2.2033, Perplexity: 9.0547
Epoch [2/3], Step [3103/3236], Loss: 1.9613, Perplexity: 7.1086
Epoch [2/3], Step [3104/3236], Loss: 2.0732, Perplexity: 7.9500
Epoch [2/3], Step [3105/3236], Loss: 2.0225, Perplexity: 7.5571
Epoch [2/3], Step [3106/3236], Loss: 2.0268, Perplexity: 7.5897
Epoch [2/3], Step [3107/3236], Loss: 1.9770, Perplexity: 7.2212
Epoch [2/3], Step [3108/3236], Loss: 1.9496, Perplexity: 7.0259
Epoch [2/3], Step [3109/3236], Loss: 1.9349, Perplexity: 6.9233
Epoch [2/3], Step [3110/3236], Loss: 1.9052, Perplexity: 6.7205
Epoch [2/3], Step [3111/3236], Loss: 1.8977, Perplexity: 6.6706
Epoch [2/3], Step [3112/3236], Loss: 2.1013, Perplexity: 8.1764
Epoch [2/3], Step [3113/3236], Loss: 2.5828, Perplexity: 13.2342
Epoch [2/3], Step [3114/3236], Loss: 2.1781, Perplexity: 8.8299
Epoch [2/3], Step [3115/3236], Loss: 1.9608, Perplexity: 7.1052
Epoch [2/3], Step [3116/3236], Loss: 1.8765, Perplexity: 6.5308
Epoch [2/3], Step [3117/3236], Loss: 2.0728, Perplexity: 7.9472
Epoch [2/3], Step [3118/3236], Loss: 2.2239, Perplexity: 9.2437
Epoch [2/3], Step [3119/3236], Loss: 1.9298, Perplexity: 6.8883
Epoch [2/3], Step [3120/3236], Loss: 2.6773, Perplexity: 14.5459
Epoch [2/3], Step [3121/3236], Loss: 2.1582, Perplexity: 8.6558
Epoch [2/3], Step [3122/3236], Loss: 1.9896, Perplexity: 7.3126
Epoch [2/3], Step [3123/3236], Loss: 2.5904, Perplexity: 13.3355
Epoch [2/3], Step [3124/3236], Loss: 2.1799, Perplexity: 8.8453
Epoch [2/3], Step [3125/3236], Loss: 2.0623, Perplexity: 7.8639
Epoch [2/3], Step [3126/3236], Loss: 2.1294, Perplexity: 8.4100
Epoch [2/3], Step [3127/3236], Loss: 2.2798, Perplexity: 9.7749
Epoch [2/3], Step [3128/3236], Loss: 2.1539, Perplexity: 8.6188
Epoch [2/3], Step [3129/3236], Loss: 2.1482, Perplexity: 8.5697
Epoch [2/3], Step [3130/3236], Loss: 1.9454, Perplexity: 6.9963
Epoch [2/3], Step [3131/3236], Loss: 2.0059, Perplexity: 7.4330
Epoch [2/3], Step [3132/3236], Loss: 2.0397, Perplexity: 7.6885
Epoch [2/3], Step [3133/3236], Loss: 2.1099, Perplexity: 8.2472
Epoch [2/3], Step [3134/3236], Loss: 1.8929, Perplexity: 6.6388
Epoch [2/3], Step [3135/3236], Loss: 2.2329, Perplexity: 9.3268
Epoch [2/3], Step [3136/3236], Loss: 2.0616, Perplexity: 7.8589
Epoch [2/3], Step [3137/3236], Loss: 1.9197, Perplexity: 6.8187
Epoch [2/3], Step [3138/3236], Loss: 1.9944, Perplexity: 7.3477
Epoch [2/3], Step [3139/3236], Loss: 2.1198, Perplexity: 8.3293
Epoch [2/3], Step [3140/3236], Loss: 2.0174, Perplexity: 7.5185
Epoch [2/3], Step [3141/3236], Loss: 2.0088, Perplexity: 7.4546
Epoch [2/3], Step [3142/3236], Loss: 1.9410, Perplexity: 6.9659
Epoch [2/3], Step [3143/3236], Loss: 2.0131, Perplexity: 7.4863
Epoch [2/3], Step [3144/3236], Loss: 2.3937, Perplexity: 10.9534
Epoch [2/3], Step [3145/3236], Loss: 2.1647, Perplexity: 8.7124
Epoch [2/3], Step [3146/3236], Loss: 2.0560, Perplexity: 7.8148
Epoch [2/3], Step [3147/3236], Loss: 2.0057, Perplexity: 7.4315
Epoch [2/3], Step [3148/3236], Loss: 1.9134, Perplexity: 6.7758
Epoch [2/3], Step [3149/3236], Loss: 2.1202, Perplexity: 8.3327
Epoch [2/3], Step [3150/3236], Loss: 1.9635, Perplexity: 7.1242
Epoch [2/3], Step [3151/3236], Loss: 2.2335, Perplexity: 9.3326
Epoch [2/3], Step [3152/3236], Loss: 1.9150, Perplexity: 6.7872
Epoch [2/3], Step [3153/3236], Loss: 1.8749, Perplexity: 6.5204
Epoch [2/3], Step [3154/3236], Loss: 2.1804, Perplexity: 8.8495
Epoch [2/3], Step [3155/3236], Loss: 1.9337, Perplexity: 6.9152
Epoch [2/3], Step [3156/3236], Loss: 2.0020, Perplexity: 7.4035
Epoch [2/3], Step [3157/3236], Loss: 1.8799, Perplexity: 6.5531
Epoch [2/3], Step [3158/3236], Loss: 3.8574, Perplexity: 47.3440
Epoch [2/3], Step [3159/3236], Loss: 2.1275, Perplexity: 8.3942
Epoch [2/3], Step [3160/3236], Loss: 2.0234, Perplexity: 7.5639
Epoch [2/3], Step [3161/3236], Loss: 2.3142, Perplexity: 10.1168
Epoch [2/3], Step [3162/3236], Loss: 2.2454, Perplexity: 9.4444
Epoch [2/3], Step [3163/3236], Loss: 2.0308, Perplexity: 7.6200
Epoch [2/3], Step [3164/3236], Loss: 2.1091, Perplexity: 8.2406
Epoch [2/3], Step [3165/3236], Loss: 2.1196, Perplexity: 8.3277
Epoch [2/3], Step [3166/3236], Loss: 2.2066, Perplexity: 9.0843
Epoch [2/3], Step [3167/3236], Loss: 1.9920, Perplexity: 7.3300
Epoch [2/3], Step [3168/3236], Loss: 2.0023, Perplexity: 7.4063
Epoch [2/3], Step [3169/3236], Loss: 1.9900, Perplexity: 7.3159
Epoch [2/3], Step [3170/3236], Loss: 2.0053, Perplexity: 7.4280
Epoch [2/3], Step [3171/3236], Loss: 2.0662, Perplexity: 7.8948
Epoch [2/3], Step [3172/3236], Loss: 2.0097, Perplexity: 7.4609
Epoch [2/3], Step [3173/3236], Loss: 1.9713, Perplexity: 7.1802
Epoch [2/3], Step [3174/3236], Loss: 2.0877, Perplexity: 8.0662
Epoch [2/3], Step [3175/3236], Loss: 2.0050, Perplexity: 7.4263
Epoch [2/3], Step [3176/3236], Loss: 1.9878, Perplexity: 7.2995
Epoch [2/3], Step [3177/3236], Loss: 1.8863, Perplexity: 6.5949
Epoch [2/3], Step [3178/3236], Loss: 2.0474, Perplexity: 7.7476
Epoch [2/3], Step [3179/3236], Loss: 2.0213, Perplexity: 7.5481
Epoch [2/3], Step [3180/3236], Loss: 2.0658, Perplexity: 7.8916
Epoch [2/3], Step [3181/3236], Loss: 2.2109, Perplexity: 9.1241
Epoch [2/3], Step [3182/3236], Loss: 1.7947, Perplexity: 6.0179
Epoch [2/3], Step [3183/3236], Loss: 1.9210, Perplexity: 6.8280
Epoch [2/3], Step [3184/3236], Loss: 1.9757, Perplexity: 7.2119
Epoch [2/3], Step [3185/3236], Loss: 1.9439, Perplexity: 6.9860
Epoch [2/3], Step [3186/3236], Loss: 2.0258, Perplexity: 7.5818
Epoch [2/3], Step [3187/3236], Loss: 2.1273, Perplexity: 8.3919
Epoch [2/3], Step [3188/3236], Loss: 2.2300, Perplexity: 9.2994
Epoch [2/3], Step [3189/3236], Loss: 2.4959, Perplexity: 12.1329
Epoch [2/3], Step [3190/3236], Loss: 1.9120, Perplexity: 6.7663
Epoch [2/3], Step [3191/3236], Loss: 1.9770, Perplexity: 7.2208
Epoch [2/3], Step [3192/3236], Loss: 1.8832, Perplexity: 6.5746
Epoch [2/3], Step [3193/3236], Loss: 2.0185, Perplexity: 7.5274
Epoch [2/3], Step [3194/3236], Loss: 1.9915, Perplexity: 7.3263
Epoch [2/3], Step [3195/3236], Loss: 1.8937, Perplexity: 6.6436
Epoch [2/3], Step [3196/3236], Loss: 2.4394, Perplexity: 11.4656
Epoch [2/3], Step [3197/3236], Loss: 1.8681, Perplexity: 6.4759
Epoch [2/3], Step [3198/3236], Loss: 2.1366, Perplexity: 8.4708
Epoch [2/3], Step [3199/3236], Loss: 2.2664, Perplexity: 9.6442
Epoch [2/3], Step [3200/3236], Loss: 2.0113, Perplexity: 7.4728
Epoch [2/3], Step [3201/3236], Loss: 1.8725, Perplexity: 6.5047
Epoch [2/3], Step [3202/3236], Loss: 1.9757, Perplexity: 7.2120
Epoch [2/3], Step [3203/3236], Loss: 2.0180, Perplexity: 7.5234
Epoch [2/3], Step [3204/3236], Loss: 1.9184, Perplexity: 6.8098
Epoch [2/3], Step [3205/3236], Loss: 2.3031, Perplexity: 10.0052
Epoch [2/3], Step [3206/3236], Loss: 1.9346, Perplexity: 6.9213
Epoch [2/3], Step [3207/3236], Loss: 2.1746, Perplexity: 8.7984
Epoch [2/3], Step [3208/3236], Loss: 1.9342, Perplexity: 6.9186
Epoch [2/3], Step [3209/3236], Loss: 1.8810, Perplexity: 6.5598
Epoch [2/3], Step [3210/3236], Loss: 2.0217, Perplexity: 7.5513
Epoch [2/3], Step [3211/3236], Loss: 2.1206, Perplexity: 8.3361
Epoch [2/3], Step [3212/3236], Loss: 1.9284, Perplexity: 6.8782
Epoch [2/3], Step [3213/3236], Loss: 2.0229, Perplexity: 7.5603
Epoch [2/3], Step [3214/3236], Loss: 1.9621, Perplexity: 7.1143
Epoch [2/3], Step [3215/3236], Loss: 2.3003, Perplexity: 9.9768
Epoch [2/3], Step [3216/3236], Loss: 2.0725, Perplexity: 7.9443
Epoch [2/3], Step [3217/3236], Loss: 2.0257, Perplexity: 7.5812
Epoch [2/3], Step [3218/3236], Loss: 2.0056, Perplexity: 7.4309
Epoch [2/3], Step [3219/3236], Loss: 2.0300, Perplexity: 7.6139
Epoch [2/3], Step [3220/3236], Loss: 1.9807, Perplexity: 7.2477
Epoch [2/3], Step [3221/3236], Loss: 2.0828, Perplexity: 8.0272
Epoch [2/3], Step [3222/3236], Loss: 1.8914, Perplexity: 6.6286
Epoch [2/3], Step [3223/3236], Loss: 1.9020, Perplexity: 6.6994
Epoch [2/3], Step [3224/3236], Loss: 2.0257, Perplexity: 7.5817
Epoch [2/3], Step [3225/3236], Loss: 2.2536, Perplexity: 9.5222
Epoch [2/3], Step [3226/3236], Loss: 2.0426, Perplexity: 7.7104
Epoch [2/3], Step [3227/3236], Loss: 1.9313, Perplexity: 6.8984
Epoch [2/3], Step [3228/3236], Loss: 1.9901, Perplexity: 7.3164
Epoch [2/3], Step [3229/3236], Loss: 1.9641, Perplexity: 7.1282
Epoch [2/3], Step [3230/3236], Loss: 2.2057, Perplexity: 9.0762
Epoch [2/3], Step [3231/3236], Loss: 2.0824, Perplexity: 8.0236
Epoch [2/3], Step [3232/3236], Loss: 2.4491, Perplexity: 11.5774
Epoch [2/3], Step [3233/3236], Loss: 1.8792, Perplexity: 6.5484
Epoch [2/3], Step [3234/3236], Loss: 2.3245, Perplexity: 10.2215
Epoch [2/3], Step [3235/3236], Loss: 1.9966, Perplexity: 7.3640
Epoch [2/3], Step [3236/3236], Loss: 1.9451, Perplexity: 6.9944
Epoch [3/3], Step [1/3236], Loss: 2.0387, Perplexity: 7.6804
Epoch [3/3], Step [2/3236], Loss: 1.9357, Perplexity: 6.9286
Epoch [3/3], Step [3/3236], Loss: 2.3055, Perplexity: 10.0288
Epoch [3/3], Step [4/3236], Loss: 1.8921, Perplexity: 6.6334
Epoch [3/3], Step [5/3236], Loss: 1.9332, Perplexity: 6.9116
Epoch [3/3], Step [6/3236], Loss: 2.0228, Perplexity: 7.5596
Epoch [3/3], Step [7/3236], Loss: 2.0223, Perplexity: 7.5555
Epoch [3/3], Step [8/3236], Loss: 2.0303, Perplexity: 7.6166
Epoch [3/3], Step [9/3236], Loss: 1.9276, Perplexity: 6.8733
Epoch [3/3], Step [10/3236], Loss: 2.2417, Perplexity: 9.4098
Epoch [3/3], Step [11/3236], Loss: 1.9077, Perplexity: 6.7377
Epoch [3/3], Step [12/3236], Loss: 2.2157, Perplexity: 9.1677
Epoch [3/3], Step [13/3236], Loss: 1.9010, Perplexity: 6.6924
Epoch [3/3], Step [14/3236], Loss: 2.1047, Perplexity: 8.2044
Epoch [3/3], Step [15/3236], Loss: 1.9533, Perplexity: 7.0516
Epoch [3/3], Step [16/3236], Loss: 1.9775, Perplexity: 7.2245
Epoch [3/3], Step [17/3236], Loss: 2.0399, Perplexity: 7.6895
Epoch [3/3], Step [18/3236], Loss: 1.9556, Perplexity: 7.0683
Epoch [3/3], Step [19/3236], Loss: 2.5618, Perplexity: 12.9588
Epoch [3/3], Step [20/3236], Loss: 2.2211, Perplexity: 9.2178
Epoch [3/3], Step [21/3236], Loss: 1.8879, Perplexity: 6.6053
Epoch [3/3], Step [22/3236], Loss: 2.0372, Perplexity: 7.6693
Epoch [3/3], Step [23/3236], Loss: 2.0685, Perplexity: 7.9127
Epoch [3/3], Step [24/3236], Loss: 2.0758, Perplexity: 7.9708
Epoch [3/3], Step [25/3236], Loss: 1.9690, Perplexity: 7.1635
Epoch [3/3], Step [26/3236], Loss: 1.8945, Perplexity: 6.6494
Epoch [3/3], Step [27/3236], Loss: 1.9299, Perplexity: 6.8886
Epoch [3/3], Step [28/3236], Loss: 2.0555, Perplexity: 7.8111
Epoch [3/3], Step [29/3236], Loss: 2.1151, Perplexity: 8.2903
Epoch [3/3], Step [30/3236], Loss: 1.9818, Perplexity: 7.2560
Epoch [3/3], Step [31/3236], Loss: 2.0628, Perplexity: 7.8683
Epoch [3/3], Step [32/3236], Loss: 2.0167, Perplexity: 7.5136
Epoch [3/3], Step [33/3236], Loss: 1.9572, Perplexity: 7.0795
Epoch [3/3], Step [34/3236], Loss: 1.9402, Perplexity: 6.9601
Epoch [3/3], Step [35/3236], Loss: 2.2913, Perplexity: 9.8880
Epoch [3/3], Step [36/3236], Loss: 2.0130, Perplexity: 7.4861
Epoch [3/3], Step [37/3236], Loss: 2.2324, Perplexity: 9.3225
Epoch [3/3], Step [38/3236], Loss: 1.9762, Perplexity: 7.2154
Epoch [3/3], Step [39/3236], Loss: 2.1506, Perplexity: 8.5904
Epoch [3/3], Step [40/3236], Loss: 2.8478, Perplexity: 17.2499
Epoch [3/3], Step [41/3236], Loss: 2.3246, Perplexity: 10.2230
Epoch [3/3], Step [42/3236], Loss: 1.9570, Perplexity: 7.0780
Epoch [3/3], Step [43/3236], Loss: 2.0120, Perplexity: 7.4782
Epoch [3/3], Step [44/3236], Loss: 1.8832, Perplexity: 6.5747
Epoch [3/3], Step [45/3236], Loss: 1.9966, Perplexity: 7.3638
Epoch [3/3], Step [46/3236], Loss: 1.9927, Perplexity: 7.3353
Epoch [3/3], Step [47/3236], Loss: 1.8779, Perplexity: 6.5400
Epoch [3/3], Step [48/3236], Loss: 1.9731, Perplexity: 7.1926
Epoch [3/3], Step [49/3236], Loss: 2.0009, Perplexity: 7.3959
Epoch [3/3], Step [50/3236], Loss: 2.0542, Perplexity: 7.8006
Epoch [3/3], Step [51/3236], Loss: 1.9376, Perplexity: 6.9423
Epoch [3/3], Step [52/3236], Loss: 1.9417, Perplexity: 6.9708
Epoch [3/3], Step [53/3236], Loss: 1.8988, Perplexity: 6.6781
Epoch [3/3], Step [54/3236], Loss: 2.3075, Perplexity: 10.0490
Epoch [3/3], Step [55/3236], Loss: 1.9640, Perplexity: 7.1279
Epoch [3/3], Step [56/3236], Loss: 2.6972, Perplexity: 14.8384
Epoch [3/3], Step [57/3236], Loss: 2.2464, Perplexity: 9.4537
Epoch [3/3], Step [58/3236], Loss: 1.8834, Perplexity: 6.5761
Epoch [3/3], Step [59/3236], Loss: 1.9509, Perplexity: 7.0351
Epoch [3/3], Step [60/3236], Loss: 2.0036, Perplexity: 7.4159
Epoch [3/3], Step [61/3236], Loss: 1.9191, Perplexity: 6.8151
Epoch [3/3], Step [62/3236], Loss: 2.0014, Perplexity: 7.3996
Epoch [3/3], Step [63/3236], Loss: 2.0723, Perplexity: 7.9429
Epoch [3/3], Step [64/3236], Loss: 1.9609, Perplexity: 7.1054
Epoch [3/3], Step [65/3236], Loss: 1.8390, Perplexity: 6.2901
Epoch [3/3], Step [66/3236], Loss: 2.0283, Perplexity: 7.6012
Epoch [3/3], Step [67/3236], Loss: 2.0401, Perplexity: 7.6914
Epoch [3/3], Step [68/3236], Loss: 1.9350, Perplexity: 6.9243
Epoch [3/3], Step [69/3236], Loss: 1.9105, Perplexity: 6.7568
Epoch [3/3], Step [70/3236], Loss: 2.1982, Perplexity: 9.0086
Epoch [3/3], Step [71/3236], Loss: 2.0015, Perplexity: 7.3999
Epoch [3/3], Step [72/3236], Loss: 1.9549, Perplexity: 7.0634
Epoch [3/3], Step [73/3236], Loss: 2.0662, Perplexity: 7.8951
Epoch [3/3], Step [74/3236], Loss: 2.1301, Perplexity: 8.4155
Epoch [3/3], Step [75/3236], Loss: 2.0342, Perplexity: 7.6461
Epoch [3/3], Step [76/3236], Loss: 1.8935, Perplexity: 6.6424
Epoch [3/3], Step [77/3236], Loss: 2.0794, Perplexity: 7.9996
Epoch [3/3], Step [78/3236], Loss: 1.8538, Perplexity: 6.3839
Epoch [3/3], Step [79/3236], Loss: 2.0793, Perplexity: 7.9990
Epoch [3/3], Step [80/3236], Loss: 1.9964, Perplexity: 7.3626
Epoch [3/3], Step [81/3236], Loss: 1.9691, Perplexity: 7.1639
Epoch [3/3], Step [82/3236], Loss: 2.8804, Perplexity: 17.8209
Epoch [3/3], Step [83/3236], Loss: 2.0486, Perplexity: 7.7569
Epoch [3/3], Step [84/3236], Loss: 2.1385, Perplexity: 8.4866
Epoch [3/3], Step [85/3236], Loss: 2.5282, Perplexity: 12.5305
Epoch [3/3], Step [86/3236], Loss: 1.9391, Perplexity: 6.9528
Epoch [3/3], Step [87/3236], Loss: 2.0187, Perplexity: 7.5286
Epoch [3/3], Step [88/3236], Loss: 2.0615, Perplexity: 7.8577
Epoch [3/3], Step [89/3236], Loss: 1.9792, Perplexity: 7.2373
Epoch [3/3], Step [90/3236], Loss: 1.8853, Perplexity: 6.5885
Epoch [3/3], Step [91/3236], Loss: 1.8187, Perplexity: 6.1640
Epoch [3/3], Step [92/3236], Loss: 2.0242, Perplexity: 7.5701
Epoch [3/3], Step [93/3236], Loss: 1.9544, Perplexity: 7.0600
Epoch [3/3], Step [94/3236], Loss: 2.0029, Perplexity: 7.4103
Epoch [3/3], Step [95/3236], Loss: 1.9548, Perplexity: 7.0628
Epoch [3/3], Step [96/3236], Loss: 1.9233, Perplexity: 6.8437
Epoch [3/3], Step [97/3236], Loss: 1.8481, Perplexity: 6.3479
Epoch [3/3], Step [98/3236], Loss: 2.7604, Perplexity: 15.8054
Epoch [3/3], Step [99/3236], Loss: 2.0722, Perplexity: 7.9419
Epoch [3/3], Step [100/3236], Loss: 2.1431, Perplexity: 8.5260
Epoch [3/3], Step [101/3236], Loss: 2.2578, Perplexity: 9.5616
Epoch [3/3], Step [102/3236], Loss: 2.3848, Perplexity: 10.8574
Epoch [3/3], Step [103/3236], Loss: 1.8946, Perplexity: 6.6500
Epoch [3/3], Step [104/3236], Loss: 2.1390, Perplexity: 8.4907
Epoch [3/3], Step [105/3236], Loss: 1.8943, Perplexity: 6.6481
Epoch [3/3], Step [106/3236], Loss: 1.8723, Perplexity: 6.5034
Epoch [3/3], Step [107/3236], Loss: 1.9152, Perplexity: 6.7880
Epoch [3/3], Step [108/3236], Loss: 2.1052, Perplexity: 8.2091
Epoch [3/3], Step [109/3236], Loss: 2.0652, Perplexity: 7.8868
Epoch [3/3], Step [110/3236], Loss: 1.9123, Perplexity: 6.7689
Epoch [3/3], Step [111/3236], Loss: 2.0339, Perplexity: 7.6438
Epoch [3/3], Step [112/3236], Loss: 2.1065, Perplexity: 8.2194
Epoch [3/3], Step [113/3236], Loss: 2.2763, Perplexity: 9.7401
Epoch [3/3], Step [114/3236], Loss: 1.9617, Perplexity: 7.1115
Epoch [3/3], Step [115/3236], Loss: 1.8409, Perplexity: 6.3023
Epoch [3/3], Step [116/3236], Loss: 1.8566, Perplexity: 6.4021
Epoch [3/3], Step [117/3236], Loss: 1.9256, Perplexity: 6.8595
Epoch [3/3], Step [118/3236], Loss: 2.1958, Perplexity: 8.9868
Epoch [3/3], Step [119/3236], Loss: 2.0465, Perplexity: 7.7404
Epoch [3/3], Step [120/3236], Loss: 2.7557, Perplexity: 15.7323
Epoch [3/3], Step [121/3236], Loss: 1.8949, Perplexity: 6.6516
Epoch [3/3], Step [122/3236], Loss: 1.9148, Perplexity: 6.7853
Epoch [3/3], Step [123/3236], Loss: 2.4329, Perplexity: 11.3916
Epoch [3/3], Step [124/3236], Loss: 2.0234, Perplexity: 7.5641
Epoch [3/3], Step [125/3236], Loss: 2.0138, Perplexity: 7.4917
Epoch [3/3], Step [126/3236], Loss: 2.0926, Perplexity: 8.1059
Epoch [3/3], Step [127/3236], Loss: 1.9164, Perplexity: 6.7968
Epoch [3/3], Step [128/3236], Loss: 1.9970, Perplexity: 7.3667
Epoch [3/3], Step [129/3236], Loss: 2.0606, Perplexity: 7.8503
Epoch [3/3], Step [130/3236], Loss: 1.9294, Perplexity: 6.8852
Epoch [3/3], Step [131/3236], Loss: 2.1167, Perplexity: 8.3035
Epoch [3/3], Step [132/3236], Loss: 1.9936, Perplexity: 7.3422
Epoch [3/3], Step [133/3236], Loss: 2.3741, Perplexity: 10.7414
Epoch [3/3], Step [134/3236], Loss: 1.9926, Perplexity: 7.3349
Epoch [3/3], Step [135/3236], Loss: 1.8764, Perplexity: 6.5298
Epoch [3/3], Step [136/3236], Loss: 2.0192, Perplexity: 7.5326
Epoch [3/3], Step [137/3236], Loss: 1.9415, Perplexity: 6.9695
Epoch [3/3], Step [138/3236], Loss: 2.0047, Perplexity: 7.4237
Epoch [3/3], Step [139/3236], Loss: 2.0008, Perplexity: 7.3952
Epoch [3/3], Step [140/3236], Loss: 1.9516, Perplexity: 7.0402
Epoch [3/3], Step [141/3236], Loss: 1.8765, Perplexity: 6.5307
Epoch [3/3], Step [142/3236], Loss: 2.1179, Perplexity: 8.3135
Epoch [3/3], Step [143/3236], Loss: 1.8678, Perplexity: 6.4743
Epoch [3/3], Step [144/3236], Loss: 1.9759, Perplexity: 7.2133
Epoch [3/3], Step [145/3236], Loss: 1.8777, Perplexity: 6.5383
Epoch [3/3], Step [146/3236], Loss: 2.0624, Perplexity: 7.8645
Epoch [3/3], Step [147/3236], Loss: 2.0309, Perplexity: 7.6209
Epoch [3/3], Step [148/3236], Loss: 1.9514, Perplexity: 7.0387
Epoch [3/3], Step [149/3236], Loss: 2.0194, Perplexity: 7.5336
Epoch [3/3], Step [150/3236], Loss: 1.9597, Perplexity: 7.0969
Epoch [3/3], Step [151/3236], Loss: 2.2306, Perplexity: 9.3052
Epoch [3/3], Step [152/3236], Loss: 1.9714, Perplexity: 7.1811
Epoch [3/3], Step [153/3236], Loss: 1.8136, Perplexity: 6.1325
Epoch [3/3], Step [154/3236], Loss: 1.9322, Perplexity: 6.9045
Epoch [3/3], Step [155/3236], Loss: 2.1365, Perplexity: 8.4700
Epoch [3/3], Step [156/3236], Loss: 2.0431, Perplexity: 7.7147
Epoch [3/3], Step [157/3236], Loss: 2.6914, Perplexity: 14.7526
Epoch [3/3], Step [158/3236], Loss: 1.8596, Perplexity: 6.4210
Epoch [3/3], Step [159/3236], Loss: 2.1052, Perplexity: 8.2089
Epoch [3/3], Step [160/3236], Loss: 1.8573, Perplexity: 6.4061
Epoch [3/3], Step [161/3236], Loss: 1.9261, Perplexity: 6.8629
Epoch [3/3], Step [162/3236], Loss: 1.9313, Perplexity: 6.8988
Epoch [3/3], Step [163/3236], Loss: 2.3403, Perplexity: 10.3842
Epoch [3/3], Step [164/3236], Loss: 1.9786, Perplexity: 7.2324
Epoch [3/3], Step [165/3236], Loss: 2.0565, Perplexity: 7.8188
Epoch [3/3], Step [166/3236], Loss: 1.9300, Perplexity: 6.8898
Epoch [3/3], Step [167/3236], Loss: 1.9794, Perplexity: 7.2382
Epoch [3/3], Step [168/3236], Loss: 1.9024, Perplexity: 6.7022
Epoch [3/3], Step [169/3236], Loss: 2.1655, Perplexity: 8.7189
Epoch [3/3], Step [170/3236], Loss: 1.8786, Perplexity: 6.5442
Epoch [3/3], Step [171/3236], Loss: 2.2373, Perplexity: 9.3683
Epoch [3/3], Step [172/3236], Loss: 1.8207, Perplexity: 6.1761
Epoch [3/3], Step [173/3236], Loss: 1.9507, Perplexity: 7.0340
Epoch [3/3], Step [174/3236], Loss: 1.7387, Perplexity: 5.6900
Epoch [3/3], Step [175/3236], Loss: 1.9355, Perplexity: 6.9278
Epoch [3/3], Step [176/3236], Loss: 2.2479, Perplexity: 9.4677
Epoch [3/3], Step [177/3236], Loss: 1.9303, Perplexity: 6.8918
Epoch [3/3], Step [178/3236], Loss: 1.9039, Perplexity: 6.7118
Epoch [3/3], Step [179/3236], Loss: 2.0933, Perplexity: 8.1116
Epoch [3/3], Step [180/3236], Loss: 1.8668, Perplexity: 6.4674
Epoch [3/3], Step [181/3236], Loss: 1.9311, Perplexity: 6.8968
Epoch [3/3], Step [182/3236], Loss: 1.9292, Perplexity: 6.8842
Epoch [3/3], Step [183/3236], Loss: 1.8753, Perplexity: 6.5227
Epoch [3/3], Step [184/3236], Loss: 2.0492, Perplexity: 7.7619
Epoch [3/3], Step [185/3236], Loss: 2.8033, Perplexity: 16.4984
Epoch [3/3], Step [186/3236], Loss: 1.9069, Perplexity: 6.7321
Epoch [3/3], Step [187/3236], Loss: 1.9786, Perplexity: 7.2327
Epoch [3/3], Step [188/3236], Loss: 2.0870, Perplexity: 8.0606
Epoch [3/3], Step [189/3236], Loss: 1.9278, Perplexity: 6.8741
Epoch [3/3], Step [190/3236], Loss: 1.9632, Perplexity: 7.1224
Epoch [3/3], Step [191/3236], Loss: 1.9898, Perplexity: 7.3141
Epoch [3/3], Step [192/3236], Loss: 1.9641, Perplexity: 7.1283
Epoch [3/3], Step [193/3236], Loss: 1.9343, Perplexity: 6.9192
Epoch [3/3], Step [194/3236], Loss: 1.9893, Perplexity: 7.3102
Epoch [3/3], Step [195/3236], Loss: 2.2232, Perplexity: 9.2366
Epoch [3/3], Step [196/3236], Loss: 2.0082, Perplexity: 7.4496
Epoch [3/3], Step [197/3236], Loss: 2.0248, Perplexity: 7.5746
Epoch [3/3], Step [198/3236], Loss: 1.8480, Perplexity: 6.3472
Epoch [3/3], Step [199/3236], Loss: 1.9084, Perplexity: 6.7421
Epoch [3/3], Step [200/3236], Loss: 2.1918, Perplexity: 8.9514
Epoch [3/3], Step [201/3236], Loss: 1.9781, Perplexity: 7.2289
Epoch [3/3], Step [202/3236], Loss: 2.2346, Perplexity: 9.3423
Epoch [3/3], Step [203/3236], Loss: 2.0690, Perplexity: 7.9170
Epoch [3/3], Step [204/3236], Loss: 1.9384, Perplexity: 6.9474
Epoch [3/3], Step [205/3236], Loss: 1.9881, Perplexity: 7.3014
Epoch [3/3], Step [206/3236], Loss: 2.2048, Perplexity: 9.0682
Epoch [3/3], Step [207/3236], Loss: 2.0360, Perplexity: 7.6599
Epoch [3/3], Step [208/3236], Loss: 1.9331, Perplexity: 6.9107
Epoch [3/3], Step [209/3236], Loss: 2.2773, Perplexity: 9.7502
Epoch [3/3], Step [210/3236], Loss: 1.9294, Perplexity: 6.8855
Epoch [3/3], Step [211/3236], Loss: 1.9854, Perplexity: 7.2821
Epoch [3/3], Step [212/3236], Loss: 1.9636, Perplexity: 7.1252
Epoch [3/3], Step [213/3236], Loss: 1.9643, Perplexity: 7.1299
Epoch [3/3], Step [214/3236], Loss: 2.0271, Perplexity: 7.5921
Epoch [3/3], Step [215/3236], Loss: 1.9724, Perplexity: 7.1883
Epoch [3/3], Step [216/3236], Loss: 1.8779, Perplexity: 6.5400
Epoch [3/3], Step [217/3236], Loss: 2.0433, Perplexity: 7.7159
Epoch [3/3], Step [218/3236], Loss: 1.8528, Perplexity: 6.3779
Epoch [3/3], Step [219/3236], Loss: 2.0410, Perplexity: 7.6981
Epoch [3/3], Step [220/3236], Loss: 1.8689, Perplexity: 6.4809
Epoch [3/3], Step [221/3236], Loss: 1.8288, Perplexity: 6.2265
Epoch [3/3], Step [222/3236], Loss: 1.9953, Perplexity: 7.3542
Epoch [3/3], Step [223/3236], Loss: 2.1638, Perplexity: 8.7041
Epoch [3/3], Step [224/3236], Loss: 2.1334, Perplexity: 8.4437
Epoch [3/3], Step [225/3236], Loss: 1.9197, Perplexity: 6.8186
Epoch [3/3], Step [226/3236], Loss: 1.8707, Perplexity: 6.4927
Epoch [3/3], Step [227/3236], Loss: 1.8915, Perplexity: 6.6293
Epoch [3/3], Step [228/3236], Loss: 2.2907, Perplexity: 9.8817
Epoch [3/3], Step [229/3236], Loss: 1.8481, Perplexity: 6.3480
Epoch [3/3], Step [230/3236], Loss: 2.0604, Perplexity: 7.8489
Epoch [3/3], Step [231/3236], Loss: 1.9933, Perplexity: 7.3397
Epoch [3/3], Step [232/3236], Loss: 1.8496, Perplexity: 6.3573
Epoch [3/3], Step [233/3236], Loss: 1.8866, Perplexity: 6.5967
Epoch [3/3], Step [234/3236], Loss: 1.8228, Perplexity: 6.1892
Epoch [3/3], Step [235/3236], Loss: 2.5999, Perplexity: 13.4626
Epoch [3/3], Step [236/3236], Loss: 1.9668, Perplexity: 7.1479
Epoch [3/3], Step [237/3236], Loss: 1.9840, Perplexity: 7.2718
Epoch [3/3], Step [238/3236], Loss: 2.3387, Perplexity: 10.3680
Epoch [3/3], Step [239/3236], Loss: 1.9492, Perplexity: 7.0229
Epoch [3/3], Step [240/3236], Loss: 2.1115, Perplexity: 8.2610
Epoch [3/3], Step [241/3236], Loss: 2.2344, Perplexity: 9.3405
Epoch [3/3], Step [242/3236], Loss: 2.2294, Perplexity: 9.2939
Epoch [3/3], Step [243/3236], Loss: 2.1263, Perplexity: 8.3840
Epoch [3/3], Step [244/3236], Loss: 1.8984, Perplexity: 6.6755
Epoch [3/3], Step [245/3236], Loss: 1.9153, Perplexity: 6.7886
Epoch [3/3], Step [246/3236], Loss: 2.7878, Perplexity: 16.2449
Epoch [3/3], Step [247/3236], Loss: 1.9930, Perplexity: 7.3374
Epoch [3/3], Step [248/3236], Loss: 1.9604, Perplexity: 7.1021
Epoch [3/3], Step [249/3236], Loss: 1.9613, Perplexity: 7.1086
Epoch [3/3], Step [250/3236], Loss: 1.8212, Perplexity: 6.1793
Epoch [3/3], Step [251/3236], Loss: 2.1289, Perplexity: 8.4057
Epoch [3/3], Step [252/3236], Loss: 1.9411, Perplexity: 6.9664
Epoch [3/3], Step [253/3236], Loss: 2.1412, Perplexity: 8.5096
Epoch [3/3], Step [254/3236], Loss: 2.2423, Perplexity: 9.4151
Epoch [3/3], Step [255/3236], Loss: 1.8974, Perplexity: 6.6688
Epoch [3/3], Step [256/3236], Loss: 1.9273, Perplexity: 6.8712
Epoch [3/3], Step [257/3236], Loss: 1.9626, Perplexity: 7.1180
Epoch [3/3], Step [258/3236], Loss: 2.0700, Perplexity: 7.9250
Epoch [3/3], Step [259/3236], Loss: 2.0730, Perplexity: 7.9488
Epoch [3/3], Step [260/3236], Loss: 2.0424, Perplexity: 7.7094
Epoch [3/3], Step [261/3236], Loss: 2.0546, Perplexity: 7.8035
Epoch [3/3], Step [262/3236], Loss: 2.1570, Perplexity: 8.6450
Epoch [3/3], Step [263/3236], Loss: 1.8792, Perplexity: 6.5481
Epoch [3/3], Step [264/3236], Loss: 2.0158, Perplexity: 7.5064
Epoch [3/3], Step [265/3236], Loss: 1.9533, Perplexity: 7.0521
Epoch [3/3], Step [266/3236], Loss: 2.0339, Perplexity: 7.6440
Epoch [3/3], Step [267/3236], Loss: 2.1672, Perplexity: 8.7338
Epoch [3/3], Step [268/3236], Loss: 2.1387, Perplexity: 8.4882
Epoch [3/3], Step [269/3236], Loss: 1.8890, Perplexity: 6.6125
Epoch [3/3], Step [270/3236], Loss: 2.0098, Perplexity: 7.4615
Epoch [3/3], Step [271/3236], Loss: 2.5023, Perplexity: 12.2106
Epoch [3/3], Step [272/3236], Loss: 1.9117, Perplexity: 6.7648
Epoch [3/3], Step [273/3236], Loss: 2.0685, Perplexity: 7.9128
Epoch [3/3], Step [274/3236], Loss: 2.4946, Perplexity: 12.1175
Epoch [3/3], Step [275/3236], Loss: 1.9176, Perplexity: 6.8049
Epoch [3/3], Step [276/3236], Loss: 2.2477, Perplexity: 9.4661
Epoch [3/3], Step [277/3236], Loss: 2.2343, Perplexity: 9.3396
Epoch [3/3], Step [278/3236], Loss: 1.9130, Perplexity: 6.7731
Epoch [3/3], Step [279/3236], Loss: 1.9934, Perplexity: 7.3404
Epoch [3/3], Step [280/3236], Loss: 1.8983, Perplexity: 6.6742
Epoch [3/3], Step [281/3236], Loss: 2.0287, Perplexity: 7.6043
Epoch [3/3], Step [282/3236], Loss: 2.1023, Perplexity: 8.1852
Epoch [3/3], Step [283/3236], Loss: 2.6867, Perplexity: 14.6824
Epoch [3/3], Step [284/3236], Loss: 1.9262, Perplexity: 6.8631
Epoch [3/3], Step [285/3236], Loss: 2.3830, Perplexity: 10.8372
Epoch [3/3], Step [286/3236], Loss: 2.7230, Perplexity: 15.2256
Epoch [3/3], Step [287/3236], Loss: 1.9742, Perplexity: 7.2009
Epoch [3/3], Step [288/3236], Loss: 1.9486, Perplexity: 7.0190
Epoch [3/3], Step [289/3236], Loss: 2.0387, Perplexity: 7.6809
Epoch [3/3], Step [290/3236], Loss: 2.2924, Perplexity: 9.8987
Epoch [3/3], Step [291/3236], Loss: 1.8817, Perplexity: 6.5645
Epoch [3/3], Step [292/3236], Loss: 1.9557, Perplexity: 7.0686
Epoch [3/3], Step [293/3236], Loss: 2.4698, Perplexity: 11.8204
Epoch [3/3], Step [294/3236], Loss: 1.9657, Perplexity: 7.1401
Epoch [3/3], Step [295/3236], Loss: 1.8091, Perplexity: 6.1047
Epoch [3/3], Step [296/3236], Loss: 2.0262, Perplexity: 7.5852
Epoch [3/3], Step [297/3236], Loss: 1.9800, Perplexity: 7.2425
Epoch [3/3], Step [298/3236], Loss: 2.0679, Perplexity: 7.9082
Epoch [3/3], Step [299/3236], Loss: 2.0367, Perplexity: 7.6649
Epoch [3/3], Step [300/3236], Loss: 1.9111, Perplexity: 6.7608
Epoch [3/3], Step [301/3236], Loss: 1.9723, Perplexity: 7.1871
Epoch [3/3], Step [302/3236], Loss: 2.1027, Perplexity: 8.1882
Epoch [3/3], Step [303/3236], Loss: 2.5047, Perplexity: 12.2402
Epoch [3/3], Step [304/3236], Loss: 1.8562, Perplexity: 6.3993
Epoch [3/3], Step [305/3236], Loss: 1.9248, Perplexity: 6.8540
Epoch [3/3], Step [306/3236], Loss: 2.3961, Perplexity: 10.9807
Epoch [3/3], Step [307/3236], Loss: 1.9197, Perplexity: 6.8192
Epoch [3/3], Step [308/3236], Loss: 1.9175, Perplexity: 6.8036
Epoch [3/3], Step [309/3236], Loss: 1.9228, Perplexity: 6.8404
Epoch [3/3], Step [310/3236], Loss: 1.9570, Perplexity: 7.0777
Epoch [3/3], Step [311/3236], Loss: 2.0722, Perplexity: 7.9420
Epoch [3/3], Step [312/3236], Loss: 2.1661, Perplexity: 8.7241
Epoch [3/3], Step [313/3236], Loss: 1.9984, Perplexity: 7.3771
Epoch [3/3], Step [314/3236], Loss: 1.9531, Perplexity: 7.0509
Epoch [3/3], Step [315/3236], Loss: 2.0261, Perplexity: 7.5845
Epoch [3/3], Step [316/3236], Loss: 1.9957, Perplexity: 7.3572
Epoch [3/3], Step [317/3236], Loss: 2.2465, Perplexity: 9.4548
Epoch [3/3], Step [318/3236], Loss: 1.9524, Perplexity: 7.0458
Epoch [3/3], Step [319/3236], Loss: 2.0205, Perplexity: 7.5423
Epoch [3/3], Step [320/3236], Loss: 2.6085, Perplexity: 13.5793
Epoch [3/3], Step [321/3236], Loss: 1.8540, Perplexity: 6.3855
Epoch [3/3], Step [322/3236], Loss: 2.0120, Perplexity: 7.4785
Epoch [3/3], Step [323/3236], Loss: 2.7094, Perplexity: 15.0203
Epoch [3/3], Step [324/3236], Loss: 1.9332, Perplexity: 6.9118
Epoch [3/3], Step [325/3236], Loss: 1.9311, Perplexity: 6.8973
Epoch [3/3], Step [326/3236], Loss: 2.1480, Perplexity: 8.5674
Epoch [3/3], Step [327/3236], Loss: 2.0354, Perplexity: 7.6554
Epoch [3/3], Step [328/3236], Loss: 1.8155, Perplexity: 6.1439
Epoch [3/3], Step [329/3236], Loss: 1.9308, Perplexity: 6.8949
Epoch [3/3], Step [330/3236], Loss: 1.9316, Perplexity: 6.9006
Epoch [3/3], Step [331/3236], Loss: 2.0239, Perplexity: 7.5674
Epoch [3/3], Step [332/3236], Loss: 1.9117, Perplexity: 6.7644
Epoch [3/3], Step [333/3236], Loss: 1.9445, Perplexity: 6.9902
Epoch [3/3], Step [334/3236], Loss: 1.8977, Perplexity: 6.6704
Epoch [3/3], Step [335/3236], Loss: 1.7780, Perplexity: 5.9177
Epoch [3/3], Step [336/3236], Loss: 1.9006, Perplexity: 6.6896
Epoch [3/3], Step [337/3236], Loss: 1.9666, Perplexity: 7.1462
Epoch [3/3], Step [338/3236], Loss: 1.9568, Perplexity: 7.0770
Epoch [3/3], Step [339/3236], Loss: 1.9131, Perplexity: 6.7740
Epoch [3/3], Step [340/3236], Loss: 1.8386, Perplexity: 6.2875
Epoch [3/3], Step [341/3236], Loss: 2.0671, Perplexity: 7.9021
Epoch [3/3], Step [342/3236], Loss: 1.9052, Perplexity: 6.7206
Epoch [3/3], Step [343/3236], Loss: 2.5455, Perplexity: 12.7498
Epoch [3/3], Step [344/3236], Loss: 1.8632, Perplexity: 6.4445
Epoch [3/3], Step [345/3236], Loss: 2.2579, Perplexity: 9.5625
Epoch [3/3], Step [346/3236], Loss: 1.8677, Perplexity: 6.4731
Epoch [3/3], Step [347/3236], Loss: 2.0957, Perplexity: 8.1315
Epoch [3/3], Step [348/3236], Loss: 1.9071, Perplexity: 6.7333
Epoch [3/3], Step [349/3236], Loss: 2.1724, Perplexity: 8.7794
Epoch [3/3], Step [350/3236], Loss: 1.8299, Perplexity: 6.2331
Epoch [3/3], Step [351/3236], Loss: 2.0522, Perplexity: 7.7847
Epoch [3/3], Step [352/3236], Loss: 1.8432, Perplexity: 6.3169
Epoch [3/3], Step [353/3236], Loss: 1.7448, Perplexity: 5.7249
Epoch [3/3], Step [354/3236], Loss: 2.1116, Perplexity: 8.2615
Epoch [3/3], Step [355/3236], Loss: 2.2154, Perplexity: 9.1648
Epoch [3/3], Step [356/3236], Loss: 1.8601, Perplexity: 6.4244
Epoch [3/3], Step [357/3236], Loss: 1.8847, Perplexity: 6.5841
Epoch [3/3], Step [358/3236], Loss: 1.8552, Perplexity: 6.3933
Epoch [3/3], Step [359/3236], Loss: 1.9604, Perplexity: 7.1021
Epoch [3/3], Step [360/3236], Loss: 2.0067, Perplexity: 7.4384
Epoch [3/3], Step [361/3236], Loss: 1.9637, Perplexity: 7.1255
Epoch [3/3], Step [362/3236], Loss: 1.9894, Perplexity: 7.3108
Epoch [3/3], Step [363/3236], Loss: 2.0527, Perplexity: 7.7891
Epoch [3/3], Step [364/3236], Loss: 2.0312, Perplexity: 7.6235
Epoch [3/3], Step [365/3236], Loss: 2.2398, Perplexity: 9.3919
Epoch [3/3], Step [366/3236], Loss: 2.1368, Perplexity: 8.4720
Epoch [3/3], Step [367/3236], Loss: 2.2804, Perplexity: 9.7806
Epoch [3/3], Step [368/3236], Loss: 2.2108, Perplexity: 9.1227
Epoch [3/3], Step [369/3236], Loss: 1.9532, Perplexity: 7.0509
Epoch [3/3], Step [370/3236], Loss: 1.9618, Perplexity: 7.1119
Epoch [3/3], Step [371/3236], Loss: 1.9317, Perplexity: 6.9014
Epoch [3/3], Step [372/3236], Loss: 2.2584, Perplexity: 9.5682
Epoch [3/3], Step [373/3236], Loss: 1.9682, Perplexity: 7.1577
Epoch [3/3], Step [374/3236], Loss: 1.9674, Perplexity: 7.1518
Epoch [3/3], Step [375/3236], Loss: 2.0142, Perplexity: 7.4946
Epoch [3/3], Step [376/3236], Loss: 2.0120, Perplexity: 7.4784
Epoch [3/3], Step [377/3236], Loss: 1.9658, Perplexity: 7.1405
Epoch [3/3], Step [378/3236], Loss: 2.0394, Perplexity: 7.6864
Epoch [3/3], Step [379/3236], Loss: 1.9858, Perplexity: 7.2848
Epoch [3/3], Step [380/3236], Loss: 1.9591, Perplexity: 7.0928
Epoch [3/3], Step [381/3236], Loss: 2.2161, Perplexity: 9.1718
Epoch [3/3], Step [382/3236], Loss: 1.9212, Perplexity: 6.8289
Epoch [3/3], Step [383/3236], Loss: 1.9742, Perplexity: 7.2009
Epoch [3/3], Step [384/3236], Loss: 2.0662, Perplexity: 7.8946
Epoch [3/3], Step [385/3236], Loss: 1.8174, Perplexity: 6.1556
Epoch [3/3], Step [386/3236], Loss: 1.9757, Perplexity: 7.2117
Epoch [3/3], Step [387/3236], Loss: 1.9263, Perplexity: 6.8643
Epoch [3/3], Step [388/3236], Loss: 2.1128, Perplexity: 8.2717
Epoch [3/3], Step [389/3236], Loss: 1.8767, Perplexity: 6.5322
Epoch [3/3], Step [390/3236], Loss: 2.8239, Perplexity: 16.8425
Epoch [3/3], Step [391/3236], Loss: 1.8501, Perplexity: 6.3604
Epoch [3/3], Step [392/3236], Loss: 2.1245, Perplexity: 8.3683
Epoch [3/3], Step [393/3236], Loss: 1.9925, Perplexity: 7.3342
Epoch [3/3], Step [394/3236], Loss: 2.0539, Perplexity: 7.7986
Epoch [3/3], Step [395/3236], Loss: 2.0346, Perplexity: 7.6494
Epoch [3/3], Step [396/3236], Loss: 1.8990, Perplexity: 6.6789
Epoch [3/3], Step [397/3236], Loss: 1.9821, Perplexity: 7.2580
Epoch [3/3], Step [398/3236], Loss: 2.0545, Perplexity: 7.8031
Epoch [3/3], Step [399/3236], Loss: 1.9397, Perplexity: 6.9563
Epoch [3/3], Step [400/3236], Loss: 2.1842, Perplexity: 8.8833
Epoch [3/3], Step [401/3236], Loss: 1.8551, Perplexity: 6.3925
Epoch [3/3], Step [402/3236], Loss: 1.9004, Perplexity: 6.6886
Epoch [3/3], Step [403/3236], Loss: 1.9309, Perplexity: 6.8954
Epoch [3/3], Step [404/3236], Loss: 1.8340, Perplexity: 6.2586
Epoch [3/3], Step [405/3236], Loss: 2.1806, Perplexity: 8.8515
Epoch [3/3], Step [406/3236], Loss: 1.8550, Perplexity: 6.3917
Epoch [3/3], Step [407/3236], Loss: 2.1064, Perplexity: 8.2189
Epoch [3/3], Step [408/3236], Loss: 1.9741, Perplexity: 7.2004
Epoch [3/3], Step [409/3236], Loss: 2.2339, Perplexity: 9.3364
Epoch [3/3], Step [410/3236], Loss: 2.2030, Perplexity: 9.0520
Epoch [3/3], Step [411/3236], Loss: 2.0526, Perplexity: 7.7884
Epoch [3/3], Step [412/3236], Loss: 1.9732, Perplexity: 7.1933
Epoch [3/3], Step [413/3236], Loss: 2.5282, Perplexity: 12.5306
Epoch [3/3], Step [414/3236], Loss: 1.8292, Perplexity: 6.2291
Epoch [3/3], Step [415/3236], Loss: 1.9023, Perplexity: 6.7013
Epoch [3/3], Step [416/3236], Loss: 1.9879, Perplexity: 7.3000
Epoch [3/3], Step [417/3236], Loss: 1.9336, Perplexity: 6.9143
Epoch [3/3], Step [418/3236], Loss: 1.7914, Perplexity: 5.9976
Epoch [3/3], Step [419/3236], Loss: 1.9739, Perplexity: 7.1985
Epoch [3/3], Step [420/3236], Loss: 1.9035, Perplexity: 6.7094
Epoch [3/3], Step [421/3236], Loss: 1.9646, Perplexity: 7.1318
Epoch [3/3], Step [422/3236], Loss: 1.9451, Perplexity: 6.9940
Epoch [3/3], Step [423/3236], Loss: 2.0236, Perplexity: 7.5658
Epoch [3/3], Step [424/3236], Loss: 2.3590, Perplexity: 10.5801
Epoch [3/3], Step [425/3236], Loss: 2.0062, Perplexity: 7.4350
Epoch [3/3], Step [426/3236], Loss: 1.9520, Perplexity: 7.0428
Epoch [3/3], Step [427/3236], Loss: 2.0258, Perplexity: 7.5821
Epoch [3/3], Step [428/3236], Loss: 1.9000, Perplexity: 6.6856
Epoch [3/3], Step [429/3236], Loss: 2.3218, Perplexity: 10.1939
Epoch [3/3], Step [430/3236], Loss: 2.0105, Perplexity: 7.4668
Epoch [3/3], Step [431/3236], Loss: 2.0216, Perplexity: 7.5507
Epoch [3/3], Step [432/3236], Loss: 2.0755, Perplexity: 7.9689
Epoch [3/3], Step [433/3236], Loss: 2.0344, Perplexity: 7.6477
Epoch [3/3], Step [434/3236], Loss: 1.9837, Perplexity: 7.2693
Epoch [3/3], Step [435/3236], Loss: 2.3226, Perplexity: 10.2026
Epoch [3/3], Step [436/3236], Loss: 1.9006, Perplexity: 6.6899
Epoch [3/3], Step [437/3236], Loss: 1.9504, Perplexity: 7.0313
Epoch [3/3], Step [438/3236], Loss: 2.2687, Perplexity: 9.6672
Epoch [3/3], Step [439/3236], Loss: 2.1720, Perplexity: 8.7756
Epoch [3/3], Step [440/3236], Loss: 2.0397, Perplexity: 7.6880
Epoch [3/3], Step [441/3236], Loss: 1.9111, Perplexity: 6.7603
Epoch [3/3], Step [442/3236], Loss: 1.9333, Perplexity: 6.9123
Epoch [3/3], Step [443/3236], Loss: 2.0466, Perplexity: 7.7413
Epoch [3/3], Step [444/3236], Loss: 1.9720, Perplexity: 7.1849
Epoch [3/3], Step [445/3236], Loss: 1.9911, Perplexity: 7.3236
Epoch [3/3], Step [446/3236], Loss: 1.9431, Perplexity: 6.9801
Epoch [3/3], Step [447/3236], Loss: 2.3824, Perplexity: 10.8305
Epoch [3/3], Step [448/3236], Loss: 1.9267, Perplexity: 6.8668
Epoch [3/3], Step [449/3236], Loss: 2.0601, Perplexity: 7.8468
Epoch [3/3], Step [450/3236], Loss: 2.1580, Perplexity: 8.6541
Epoch [3/3], Step [451/3236], Loss: 2.1675, Perplexity: 8.7363
Epoch [3/3], Step [452/3236], Loss: 2.1006, Perplexity: 8.1714
Epoch [3/3], Step [453/3236], Loss: 2.0381, Perplexity: 7.6759
Epoch [3/3], Step [454/3236], Loss: 1.9740, Perplexity: 7.1993
Epoch [3/3], Step [455/3236], Loss: 2.2457, Perplexity: 9.4470
Epoch [3/3], Step [456/3236], Loss: 2.0257, Perplexity: 7.5815
Epoch [3/3], Step [457/3236], Loss: 2.0884, Perplexity: 8.0717
Epoch [3/3], Step [458/3236], Loss: 1.9420, Perplexity: 6.9727
Epoch [3/3], Step [459/3236], Loss: 1.8285, Perplexity: 6.2245
Epoch [3/3], Step [460/3236], Loss: 2.0801, Perplexity: 8.0051
Epoch [3/3], Step [461/3236], Loss: 1.8615, Perplexity: 6.4333
Epoch [3/3], Step [462/3236], Loss: 1.9412, Perplexity: 6.9672
Epoch [3/3], Step [463/3236], Loss: 1.9381, Perplexity: 6.9453
Epoch [3/3], Step [464/3236], Loss: 1.9171, Perplexity: 6.8013
Epoch [3/3], Step [465/3236], Loss: 2.0059, Perplexity: 7.4326
Epoch [3/3], Step [466/3236], Loss: 1.9141, Perplexity: 6.7809
Epoch [3/3], Step [467/3236], Loss: 2.2903, Perplexity: 9.8778
Epoch [3/3], Step [468/3236], Loss: 2.4215, Perplexity: 11.2624
Epoch [3/3], Step [469/3236], Loss: 1.9105, Perplexity: 6.7565
Epoch [3/3], Step [470/3236], Loss: 2.2118, Perplexity: 9.1323
Epoch [3/3], Step [471/3236], Loss: 1.9276, Perplexity: 6.8732
Epoch [3/3], Step [472/3236], Loss: 2.0600, Perplexity: 7.8462
Epoch [3/3], Step [473/3236], Loss: 2.8495, Perplexity: 17.2787
Epoch [3/3], Step [474/3236], Loss: 2.0153, Perplexity: 7.5030
Epoch [3/3], Step [475/3236], Loss: 1.9214, Perplexity: 6.8304
Epoch [3/3], Step [476/3236], Loss: 2.0352, Perplexity: 7.6539
Epoch [3/3], Step [477/3236], Loss: 2.0268, Perplexity: 7.5896
Epoch [3/3], Step [478/3236], Loss: 1.9265, Perplexity: 6.8657
Epoch [3/3], Step [479/3236], Loss: 2.4641, Perplexity: 11.7528
Epoch [3/3], Step [480/3236], Loss: 1.9092, Perplexity: 6.7477
Epoch [3/3], Step [481/3236], Loss: 1.9296, Perplexity: 6.8870
Epoch [3/3], Step [482/3236], Loss: 1.9053, Perplexity: 6.7213
Epoch [3/3], Step [483/3236], Loss: 2.0355, Perplexity: 7.6559
Epoch [3/3], Step [484/3236], Loss: 1.9001, Perplexity: 6.6864
Epoch [3/3], Step [485/3236], Loss: 2.0205, Perplexity: 7.5420
Epoch [3/3], Step [486/3236], Loss: 1.9556, Perplexity: 7.0683
Epoch [3/3], Step [487/3236], Loss: 1.9751, Perplexity: 7.2070
Epoch [3/3], Step [488/3236], Loss: 1.9301, Perplexity: 6.8901
Epoch [3/3], Step [489/3236], Loss: 2.1198, Perplexity: 8.3295
Epoch [3/3], Step [490/3236], Loss: 1.8888, Perplexity: 6.6112
Epoch [3/3], Step [491/3236], Loss: 2.0219, Perplexity: 7.5529
Epoch [3/3], Step [492/3236], Loss: 1.9583, Perplexity: 7.0873
Epoch [3/3], Step [493/3236], Loss: 1.9105, Perplexity: 6.7566
Epoch [3/3], Step [494/3236], Loss: 2.5355, Perplexity: 12.6228
Epoch [3/3], Step [495/3236], Loss: 1.8459, Perplexity: 6.3338
Epoch [3/3], Step [496/3236], Loss: 2.2404, Perplexity: 9.3967
Epoch [3/3], Step [497/3236], Loss: 2.3192, Perplexity: 10.1678
Epoch [3/3], Step [498/3236], Loss: 2.9498, Perplexity: 19.1025
Epoch [3/3], Step [499/3236], Loss: 2.0521, Perplexity: 7.7842
Epoch [3/3], Step [500/3236], Loss: 2.0024, Perplexity: 7.4071
Epoch [3/3], Step [501/3236], Loss: 2.0908, Perplexity: 8.0911
Epoch [3/3], Step [502/3236], Loss: 1.8828, Perplexity: 6.5717
Epoch [3/3], Step [503/3236], Loss: 1.9347, Perplexity: 6.9218
Epoch [3/3], Step [504/3236], Loss: 2.0981, Perplexity: 8.1511
Epoch [3/3], Step [505/3236], Loss: 1.9989, Perplexity: 7.3807
Epoch [3/3], Step [506/3236], Loss: 1.9475, Perplexity: 7.0112
Epoch [3/3], Step [507/3236], Loss: 2.0450, Perplexity: 7.7292
Epoch [3/3], Step [508/3236], Loss: 2.0039, Perplexity: 7.4177
Epoch [3/3], Step [509/3236], Loss: 1.9044, Perplexity: 6.7152
Epoch [3/3], Step [510/3236], Loss: 1.9941, Perplexity: 7.3453
Epoch [3/3], Step [511/3236], Loss: 2.0516, Perplexity: 7.7806
Epoch [3/3], Step [512/3236], Loss: 1.9012, Perplexity: 6.6938
Epoch [3/3], Step [513/3236], Loss: 2.0087, Perplexity: 7.4540
Epoch [3/3], Step [514/3236], Loss: 2.2128, Perplexity: 9.1409
Epoch [3/3], Step [515/3236], Loss: 1.9353, Perplexity: 6.9262
Epoch [3/3], Step [516/3236], Loss: 1.9834, Perplexity: 7.2675
Epoch [3/3], Step [517/3236], Loss: 2.5695, Perplexity: 13.0596
Epoch [3/3], Step [518/3236], Loss: 2.0553, Perplexity: 7.8089
Epoch [3/3], Step [519/3236], Loss: 1.9909, Perplexity: 7.3222
Epoch [3/3], Step [520/3236], Loss: 1.8380, Perplexity: 6.2842
Epoch [3/3], Step [521/3236], Loss: 1.9849, Perplexity: 7.2784
Epoch [3/3], Step [522/3236], Loss: 2.0310, Perplexity: 7.6219
Epoch [3/3], Step [523/3236], Loss: 2.0630, Perplexity: 7.8696
Epoch [3/3], Step [524/3236], Loss: 1.7953, Perplexity: 6.0216
Epoch [3/3], Step [525/3236], Loss: 1.8497, Perplexity: 6.3580
Epoch [3/3], Step [526/3236], Loss: 1.9836, Perplexity: 7.2688
Epoch [3/3], Step [527/3236], Loss: 2.1507, Perplexity: 8.5905
Epoch [3/3], Step [528/3236], Loss: 2.1961, Perplexity: 8.9897
Epoch [3/3], Step [529/3236], Loss: 2.0808, Perplexity: 8.0109
Epoch [3/3], Step [530/3236], Loss: 1.8882, Perplexity: 6.6074
Epoch [3/3], Step [531/3236], Loss: 1.9587, Perplexity: 7.0904
Epoch [3/3], Step [532/3236], Loss: 2.0455, Perplexity: 7.7329
Epoch [3/3], Step [533/3236], Loss: 2.3384, Perplexity: 10.3644
Epoch [3/3], Step [534/3236], Loss: 2.0711, Perplexity: 7.9332
Epoch [3/3], Step [535/3236], Loss: 1.9117, Perplexity: 6.7646
Epoch [3/3], Step [536/3236], Loss: 2.0586, Perplexity: 7.8347
Epoch [3/3], Step [537/3236], Loss: 1.9796, Perplexity: 7.2401
Epoch [3/3], Step [538/3236], Loss: 2.0047, Perplexity: 7.4239
Epoch [3/3], Step [539/3236], Loss: 2.0490, Perplexity: 7.7601
Epoch [3/3], Step [540/3236], Loss: 1.9365, Perplexity: 6.9346
Epoch [3/3], Step [541/3236], Loss: 1.9376, Perplexity: 6.9419
Epoch [3/3], Step [542/3236], Loss: 1.9464, Perplexity: 7.0033
Epoch [3/3], Step [543/3236], Loss: 1.9262, Perplexity: 6.8636
Epoch [3/3], Step [544/3236], Loss: 1.9584, Perplexity: 7.0877
Epoch [3/3], Step [545/3236], Loss: 2.0513, Perplexity: 7.7783
Epoch [3/3], Step [546/3236], Loss: 2.0288, Perplexity: 7.6046
Epoch [3/3], Step [547/3236], Loss: 2.0889, Perplexity: 8.0761
Epoch [3/3], Step [548/3236], Loss: 1.8207, Perplexity: 6.1762
Epoch [3/3], Step [549/3236], Loss: 1.9942, Perplexity: 7.3462
Epoch [3/3], Step [550/3236], Loss: 2.0110, Perplexity: 7.4709
Epoch [3/3], Step [551/3236], Loss: 2.3171, Perplexity: 10.1460
Epoch [3/3], Step [552/3236], Loss: 2.2304, Perplexity: 9.3039
Epoch [3/3], Step [553/3236], Loss: 2.1157, Perplexity: 8.2951
Epoch [3/3], Step [554/3236], Loss: 1.9584, Perplexity: 7.0883
Epoch [3/3], Step [555/3236], Loss: 1.9615, Perplexity: 7.1097
Epoch [3/3], Step [556/3236], Loss: 1.9131, Perplexity: 6.7741
Epoch [3/3], Step [557/3236], Loss: 1.9532, Perplexity: 7.0514
Epoch [3/3], Step [558/3236], Loss: 1.9317, Perplexity: 6.9014
Epoch [3/3], Step [559/3236], Loss: 1.8526, Perplexity: 6.3763
Epoch [3/3], Step [560/3236], Loss: 2.1494, Perplexity: 8.5797
Epoch [3/3], Step [561/3236], Loss: 2.0828, Perplexity: 8.0266
Epoch [3/3], Step [562/3236], Loss: 1.9353, Perplexity: 6.9264
Epoch [3/3], Step [563/3236], Loss: 2.0299, Perplexity: 7.6135
Epoch [3/3], Step [564/3236], Loss: 1.8925, Perplexity: 6.6357
Epoch [3/3], Step [565/3236], Loss: 2.2602, Perplexity: 9.5846
Epoch [3/3], Step [566/3236], Loss: 1.7989, Perplexity: 6.0429
Epoch [3/3], Step [567/3236], Loss: 2.4128, Perplexity: 11.1652
Epoch [3/3], Step [568/3236], Loss: 1.9384, Perplexity: 6.9475
Epoch [3/3], Step [569/3236], Loss: 1.9582, Perplexity: 7.0865
Epoch [3/3], Step [570/3236], Loss: 1.8677, Perplexity: 6.4734
Epoch [3/3], Step [571/3236], Loss: 1.8777, Perplexity: 6.5384
Epoch [3/3], Step [572/3236], Loss: 1.9810, Perplexity: 7.2497
Epoch [3/3], Step [573/3236], Loss: 2.2587, Perplexity: 9.5706
Epoch [3/3], Step [574/3236], Loss: 1.9906, Perplexity: 7.3201
Epoch [3/3], Step [575/3236], Loss: 1.9265, Perplexity: 6.8656
Epoch [3/3], Step [576/3236], Loss: 2.2361, Perplexity: 9.3567
Epoch [3/3], Step [577/3236], Loss: 2.1491, Perplexity: 8.5776
Epoch [3/3], Step [578/3236], Loss: 1.9010, Perplexity: 6.6926
Epoch [3/3], Step [579/3236], Loss: 1.9382, Perplexity: 6.9460
Epoch [3/3], Step [580/3236], Loss: 2.0833, Perplexity: 8.0313
Epoch [3/3], Step [581/3236], Loss: 2.2102, Perplexity: 9.1171
Epoch [3/3], Step [582/3236], Loss: 2.2721, Perplexity: 9.6993
Epoch [3/3], Step [583/3236], Loss: 1.8222, Perplexity: 6.1856
Epoch [3/3], Step [584/3236], Loss: 2.1329, Perplexity: 8.4392
Epoch [3/3], Step [585/3236], Loss: 2.0834, Perplexity: 8.0319
Epoch [3/3], Step [586/3236], Loss: 1.9893, Perplexity: 7.3104
Epoch [3/3], Step [587/3236], Loss: 2.0756, Perplexity: 7.9696
Epoch [3/3], Step [588/3236], Loss: 1.9409, Perplexity: 6.9652
Epoch [3/3], Step [589/3236], Loss: 1.8949, Perplexity: 6.6520
Epoch [3/3], Step [590/3236], Loss: 2.0071, Perplexity: 7.4418
Epoch [3/3], Step [591/3236], Loss: 2.0247, Perplexity: 7.5740
Epoch [3/3], Step [592/3236], Loss: 2.0331, Perplexity: 7.6378
Epoch [3/3], Step [593/3236], Loss: 1.9337, Perplexity: 6.9148
Epoch [3/3], Step [594/3236], Loss: 2.0118, Perplexity: 7.4767
Epoch [3/3], Step [595/3236], Loss: 2.0319, Perplexity: 7.6283
Epoch [3/3], Step [596/3236], Loss: 1.8570, Perplexity: 6.4046
Epoch [3/3], Step [597/3236], Loss: 1.9319, Perplexity: 6.9025
Epoch [3/3], Step [598/3236], Loss: 1.9624, Perplexity: 7.1165
Epoch [3/3], Step [599/3236], Loss: 2.0438, Perplexity: 7.7203
Epoch [3/3], Step [600/3236], Loss: 2.0476, Perplexity: 7.7497
Epoch [3/3], Step [601/3236], Loss: 2.5592, Perplexity: 12.9253
Epoch [3/3], Step [602/3236], Loss: 1.8558, Perplexity: 6.3967
Epoch [3/3], Step [603/3236], Loss: 2.2024, Perplexity: 9.0468
Epoch [3/3], Step [604/3236], Loss: 1.8323, Perplexity: 6.2480
Epoch [3/3], Step [605/3236], Loss: 2.5382, Perplexity: 12.6565
Epoch [3/3], Step [606/3236], Loss: 2.1943, Perplexity: 8.9734
Epoch [3/3], Step [607/3236], Loss: 1.9953, Perplexity: 7.3541
Epoch [3/3], Step [608/3236], Loss: 2.1134, Perplexity: 8.2761
Epoch [3/3], Step [609/3236], Loss: 2.0514, Perplexity: 7.7788
Epoch [3/3], Step [610/3236], Loss: 1.9560, Perplexity: 7.0709
Epoch [3/3], Step [611/3236], Loss: 2.0832, Perplexity: 8.0302
Epoch [3/3], Step [612/3236], Loss: 2.0313, Perplexity: 7.6238
Epoch [3/3], Step [613/3236], Loss: 1.9329, Perplexity: 6.9092
Epoch [3/3], Step [614/3236], Loss: 1.9801, Perplexity: 7.2436
Epoch [3/3], Step [615/3236], Loss: 2.1811, Perplexity: 8.8557
Epoch [3/3], Step [616/3236], Loss: 1.8385, Perplexity: 6.2874
Epoch [3/3], Step [617/3236], Loss: 1.8671, Perplexity: 6.4693
Epoch [3/3], Step [618/3236], Loss: 2.0334, Perplexity: 7.6400
Epoch [3/3], Step [619/3236], Loss: 2.1006, Perplexity: 8.1712
Epoch [3/3], Step [620/3236], Loss: 1.9319, Perplexity: 6.9026
Epoch [3/3], Step [621/3236], Loss: 1.9520, Perplexity: 7.0428
Epoch [3/3], Step [622/3236], Loss: 1.9900, Perplexity: 7.3152
Epoch [3/3], Step [623/3236], Loss: 2.0086, Perplexity: 7.4529
Epoch [3/3], Step [624/3236], Loss: 1.9563, Perplexity: 7.0732
Epoch [3/3], Step [625/3236], Loss: 2.0994, Perplexity: 8.1611
Epoch [3/3], Step [626/3236], Loss: 2.3957, Perplexity: 10.9763
Epoch [3/3], Step [627/3236], Loss: 2.2535, Perplexity: 9.5211
Epoch [3/3], Step [628/3236], Loss: 1.8607, Perplexity: 6.4282
Epoch [3/3], Step [629/3236], Loss: 2.0428, Perplexity: 7.7125
Epoch [3/3], Step [630/3236], Loss: 2.0536, Perplexity: 7.7963
Epoch [3/3], Step [631/3236], Loss: 2.1174, Perplexity: 8.3092
Epoch [3/3], Step [632/3236], Loss: 2.0434, Perplexity: 7.7168
Epoch [3/3], Step [633/3236], Loss: 2.2366, Perplexity: 9.3614
Epoch [3/3], Step [634/3236], Loss: 1.9635, Perplexity: 7.1240
Epoch [3/3], Step [635/3236], Loss: 1.9285, Perplexity: 6.8790
Epoch [3/3], Step [636/3236], Loss: 1.7873, Perplexity: 5.9735
Epoch [3/3], Step [637/3236], Loss: 1.9164, Perplexity: 6.7966
Epoch [3/3], Step [638/3236], Loss: 1.8412, Perplexity: 6.3044
Epoch [3/3], Step [639/3236], Loss: 1.9725, Perplexity: 7.1885
Epoch [3/3], Step [640/3236], Loss: 1.8645, Perplexity: 6.4529
Epoch [3/3], Step [641/3236], Loss: 2.3595, Perplexity: 10.5856
Epoch [3/3], Step [642/3236], Loss: 1.9662, Perplexity: 7.1436
Epoch [3/3], Step [643/3236], Loss: 1.9544, Perplexity: 7.0595
Epoch [3/3], Step [644/3236], Loss: 1.8864, Perplexity: 6.5959
Epoch [3/3], Step [645/3236], Loss: 1.9102, Perplexity: 6.7546
Epoch [3/3], Step [646/3236], Loss: 1.9683, Perplexity: 7.1586
Epoch [3/3], Step [647/3236], Loss: 2.0328, Perplexity: 7.6352
Epoch [3/3], Step [648/3236], Loss: 1.9095, Perplexity: 6.7500
Epoch [3/3], Step [649/3236], Loss: 2.0091, Perplexity: 7.4565
Epoch [3/3], Step [650/3236], Loss: 2.0350, Perplexity: 7.6519
Epoch [3/3], Step [651/3236], Loss: 1.8617, Perplexity: 6.4344
Epoch [3/3], Step [652/3236], Loss: 1.9693, Perplexity: 7.1657
Epoch [3/3], Step [653/3236], Loss: 1.8195, Perplexity: 6.1686
Epoch [3/3], Step [654/3236], Loss: 1.9353, Perplexity: 6.9262
Epoch [3/3], Step [655/3236], Loss: 2.0107, Perplexity: 7.4684
Epoch [3/3], Step [656/3236], Loss: 1.9547, Perplexity: 7.0620
Epoch [3/3], Step [657/3236], Loss: 1.8216, Perplexity: 6.1816
Epoch [3/3], Step [658/3236], Loss: 1.9781, Perplexity: 7.2286
Epoch [3/3], Step [659/3236], Loss: 1.8734, Perplexity: 6.5102
Epoch [3/3], Step [660/3236], Loss: 2.3029, Perplexity: 10.0032
Epoch [3/3], Step [661/3236], Loss: 1.9446, Perplexity: 6.9910
Epoch [3/3], Step [662/3236], Loss: 2.0104, Perplexity: 7.4660
Epoch [3/3], Step [663/3236], Loss: 1.9276, Perplexity: 6.8727
Epoch [3/3], Step [664/3236], Loss: 2.1510, Perplexity: 8.5936
Epoch [3/3], Step [665/3236], Loss: 1.8690, Perplexity: 6.4820
Epoch [3/3], Step [666/3236], Loss: 1.8890, Perplexity: 6.6130
Epoch [3/3], Step [667/3236], Loss: 2.6409, Perplexity: 14.0259
Epoch [3/3], Step [668/3236], Loss: 1.9381, Perplexity: 6.9455
Epoch [3/3], Step [669/3236], Loss: 2.0837, Perplexity: 8.0345
Epoch [3/3], Step [670/3236], Loss: 1.9730, Perplexity: 7.1921
Epoch [3/3], Step [671/3236], Loss: 2.0190, Perplexity: 7.5307
Epoch [3/3], Step [672/3236], Loss: 1.9696, Perplexity: 7.1676
Epoch [3/3], Step [673/3236], Loss: 2.3178, Perplexity: 10.1530
Epoch [3/3], Step [674/3236], Loss: 2.0519, Perplexity: 7.7827
Epoch [3/3], Step [675/3236], Loss: 2.0006, Perplexity: 7.3937
Epoch [3/3], Step [676/3236], Loss: 2.4233, Perplexity: 11.2832
Epoch [3/3], Step [677/3236], Loss: 2.0011, Perplexity: 7.3976
Epoch [3/3], Step [678/3236], Loss: 2.1212, Perplexity: 8.3408
Epoch [3/3], Step [679/3236], Loss: 1.9664, Perplexity: 7.1452
Epoch [3/3], Step [680/3236], Loss: 2.1258, Perplexity: 8.3799
Epoch [3/3], Step [681/3236], Loss: 2.2109, Perplexity: 9.1236
Epoch [3/3], Step [682/3236], Loss: 1.8587, Perplexity: 6.4151
Epoch [3/3], Step [683/3236], Loss: 2.1200, Perplexity: 8.3315
Epoch [3/3], Step [684/3236], Loss: 1.9073, Perplexity: 6.7351
Epoch [3/3], Step [685/3236], Loss: 1.8145, Perplexity: 6.1382
Epoch [3/3], Step [686/3236], Loss: 2.1123, Perplexity: 8.2673
Epoch [3/3], Step [687/3236], Loss: 1.9088, Perplexity: 6.7449
Epoch [3/3], Step [688/3236], Loss: 1.9526, Perplexity: 7.0472
Epoch [3/3], Step [689/3236], Loss: 1.8760, Perplexity: 6.5275
Epoch [3/3], Step [690/3236], Loss: 2.0441, Perplexity: 7.7223
Epoch [3/3], Step [691/3236], Loss: 2.0111, Perplexity: 7.4717
Epoch [3/3], Step [692/3236], Loss: 2.2103, Perplexity: 9.1184
Epoch [3/3], Step [693/3236], Loss: 2.0096, Perplexity: 7.4604
Epoch [3/3], Step [694/3236], Loss: 1.9574, Perplexity: 7.0809
Epoch [3/3], Step [695/3236], Loss: 1.8930, Perplexity: 6.6395
Epoch [3/3], Step [696/3236], Loss: 1.9904, Perplexity: 7.3184
Epoch [3/3], Step [697/3236], Loss: 1.8096, Perplexity: 6.1082
Epoch [3/3], Step [698/3236], Loss: 2.3288, Perplexity: 10.2651
Epoch [3/3], Step [699/3236], Loss: 2.5469, Perplexity: 12.7673
Epoch [3/3], Step [700/3236], Loss: 2.4844, Perplexity: 11.9944
Epoch [3/3], Step [701/3236], Loss: 2.0329, Perplexity: 7.6362
Epoch [3/3], Step [702/3236], Loss: 2.0754, Perplexity: 7.9674
Epoch [3/3], Step [703/3236], Loss: 1.9170, Perplexity: 6.8004
Epoch [3/3], Step [704/3236], Loss: 2.0038, Perplexity: 7.4172
Epoch [3/3], Step [705/3236], Loss: 2.0321, Perplexity: 7.6303
Epoch [3/3], Step [706/3236], Loss: 1.8247, Perplexity: 6.2008
Epoch [3/3], Step [707/3236], Loss: 1.9302, Perplexity: 6.8912
Epoch [3/3], Step [708/3236], Loss: 2.2288, Perplexity: 9.2883
Epoch [3/3], Step [709/3236], Loss: 2.0040, Perplexity: 7.4184
Epoch [3/3], Step [710/3236], Loss: 1.9070, Perplexity: 6.7332
Epoch [3/3], Step [711/3236], Loss: 1.8479, Perplexity: 6.3464
Epoch [3/3], Step [712/3236], Loss: 1.8340, Perplexity: 6.2589
Epoch [3/3], Step [713/3236], Loss: 2.0148, Perplexity: 7.4994
Epoch [3/3], Step [714/3236], Loss: 1.9425, Perplexity: 6.9765
Epoch [3/3], Step [715/3236], Loss: 2.1816, Perplexity: 8.8608
Epoch [3/3], Step [716/3236], Loss: 1.9489, Perplexity: 7.0213
Epoch [3/3], Step [717/3236], Loss: 2.1897, Perplexity: 8.9325
Epoch [3/3], Step [718/3236], Loss: 2.0401, Perplexity: 7.6912
Epoch [3/3], Step [719/3236], Loss: 2.0564, Perplexity: 7.8177
Epoch [3/3], Step [720/3236], Loss: 1.9147, Perplexity: 6.7848
Epoch [3/3], Step [721/3236], Loss: 2.0672, Perplexity: 7.9029
Epoch [3/3], Step [722/3236], Loss: 1.8963, Perplexity: 6.6613
Epoch [3/3], Step [723/3236], Loss: 1.9038, Perplexity: 6.7113
Epoch [3/3], Step [724/3236], Loss: 2.1530, Perplexity: 8.6106
Epoch [3/3], Step [725/3236], Loss: 1.8838, Perplexity: 6.5785
Epoch [3/3], Step [726/3236], Loss: 2.2069, Perplexity: 9.0874
Epoch [3/3], Step [727/3236], Loss: 1.9968, Perplexity: 7.3652
Epoch [3/3], Step [728/3236], Loss: 2.2058, Perplexity: 9.0776
Epoch [3/3], Step [729/3236], Loss: 1.8849, Perplexity: 6.5854
Epoch [3/3], Step [730/3236], Loss: 2.0389, Perplexity: 7.6819
Epoch [3/3], Step [731/3236], Loss: 1.9559, Perplexity: 7.0700
Epoch [3/3], Step [732/3236], Loss: 1.8520, Perplexity: 6.3724
Epoch [3/3], Step [733/3236], Loss: 1.8843, Perplexity: 6.5817
Epoch [3/3], Step [734/3236], Loss: 2.1103, Perplexity: 8.2508
Epoch [3/3], Step [735/3236], Loss: 1.8425, Perplexity: 6.3123
Epoch [3/3], Step [736/3236], Loss: 1.8259, Perplexity: 6.2083
Epoch [3/3], Step [737/3236], Loss: 2.1819, Perplexity: 8.8631
Epoch [3/3], Step [738/3236], Loss: 2.0597, Perplexity: 7.8439
Epoch [3/3], Step [739/3236], Loss: 1.9876, Perplexity: 7.2981
Epoch [3/3], Step [740/3236], Loss: 2.5936, Perplexity: 13.3775
Epoch [3/3], Step [741/3236], Loss: 2.2070, Perplexity: 9.0881
Epoch [3/3], Step [742/3236], Loss: 1.9154, Perplexity: 6.7896
Epoch [3/3], Step [743/3236], Loss: 2.2587, Perplexity: 9.5702
Epoch [3/3], Step [744/3236], Loss: 2.1592, Perplexity: 8.6643
Epoch [3/3], Step [745/3236], Loss: 2.0030, Perplexity: 7.4114
Epoch [3/3], Step [746/3236], Loss: 2.3300, Perplexity: 10.2778
Epoch [3/3], Step [747/3236], Loss: 2.0419, Perplexity: 7.7055
Epoch [3/3], Step [748/3236], Loss: 1.8477, Perplexity: 6.3450
Epoch [3/3], Step [749/3236], Loss: 1.9072, Perplexity: 6.7343
Epoch [3/3], Step [750/3236], Loss: 2.0280, Perplexity: 7.5989
Epoch [3/3], Step [751/3236], Loss: 1.8876, Perplexity: 6.6033
Epoch [3/3], Step [752/3236], Loss: 2.2124, Perplexity: 9.1379
Epoch [3/3], Step [753/3236], Loss: 2.1188, Perplexity: 8.3211
Epoch [3/3], Step [754/3236], Loss: 1.8922, Perplexity: 6.6341
Epoch [3/3], Step [755/3236], Loss: 2.2698, Perplexity: 9.6771
Epoch [3/3], Step [756/3236], Loss: 1.9018, Perplexity: 6.6982
Epoch [3/3], Step [757/3236], Loss: 2.0852, Perplexity: 8.0460
Epoch [3/3], Step [758/3236], Loss: 1.9080, Perplexity: 6.7397
Epoch [3/3], Step [759/3236], Loss: 1.8783, Perplexity: 6.5425
Epoch [3/3], Step [760/3236], Loss: 2.0695, Perplexity: 7.9207
Epoch [3/3], Step [761/3236], Loss: 2.1082, Perplexity: 8.2335
Epoch [3/3], Step [762/3236], Loss: 1.9248, Perplexity: 6.8541
Epoch [3/3], Step [763/3236], Loss: 1.9087, Perplexity: 6.7441
Epoch [3/3], Step [764/3236], Loss: 2.1996, Perplexity: 9.0214
Epoch [3/3], Step [765/3236], Loss: 2.0389, Perplexity: 7.6818
Epoch [3/3], Step [766/3236], Loss: 2.0223, Perplexity: 7.5559
Epoch [3/3], Step [767/3236], Loss: 2.0359, Perplexity: 7.6592
Epoch [3/3], Step [768/3236], Loss: 1.9616, Perplexity: 7.1105
Epoch [3/3], Step [769/3236], Loss: 2.1734, Perplexity: 8.7878
Epoch [3/3], Step [770/3236], Loss: 2.1746, Perplexity: 8.7985
Epoch [3/3], Step [771/3236], Loss: 2.1784, Perplexity: 8.8322
Epoch [3/3], Step [772/3236], Loss: 2.1274, Perplexity: 8.3932
Epoch [3/3], Step [773/3236], Loss: 2.1005, Perplexity: 8.1705
Epoch [3/3], Step [774/3236], Loss: 2.0179, Perplexity: 7.5222
Epoch [3/3], Step [775/3236], Loss: 2.1502, Perplexity: 8.5866
Epoch [3/3], Step [776/3236], Loss: 1.9243, Perplexity: 6.8501
Epoch [3/3], Step [777/3236], Loss: 1.9282, Perplexity: 6.8769
Epoch [3/3], Step [778/3236], Loss: 1.7643, Perplexity: 5.8376
Epoch [3/3], Step [779/3236], Loss: 2.0006, Perplexity: 7.3938
Epoch [3/3], Step [780/3236], Loss: 2.1602, Perplexity: 8.6731
Epoch [3/3], Step [781/3236], Loss: 2.1101, Perplexity: 8.2490
Epoch [3/3], Step [782/3236], Loss: 2.1191, Perplexity: 8.3237
Epoch [3/3], Step [783/3236], Loss: 1.9739, Perplexity: 7.1987
Epoch [3/3], Step [784/3236], Loss: 1.9269, Perplexity: 6.8684
Epoch [3/3], Step [785/3236], Loss: 1.9265, Perplexity: 6.8652
Epoch [3/3], Step [786/3236], Loss: 1.8973, Perplexity: 6.6680
Epoch [3/3], Step [787/3236], Loss: 2.0273, Perplexity: 7.5936
Epoch [3/3], Step [788/3236], Loss: 1.8460, Perplexity: 6.3347
Epoch [3/3], Step [789/3236], Loss: 2.2561, Perplexity: 9.5454
Epoch [3/3], Step [790/3236], Loss: 1.9163, Perplexity: 6.7960
Epoch [3/3], Step [791/3236], Loss: 1.9666, Perplexity: 7.1466
Epoch [3/3], Step [792/3236], Loss: 1.9228, Perplexity: 6.8397
Epoch [3/3], Step [793/3236], Loss: 2.0515, Perplexity: 7.7793
Epoch [3/3], Step [794/3236], Loss: 1.8011, Perplexity: 6.0565
Epoch [3/3], Step [795/3236], Loss: 1.9322, Perplexity: 6.9047
Epoch [3/3], Step [796/3236], Loss: 1.9957, Perplexity: 7.3570
Epoch [3/3], Step [797/3236], Loss: 2.2621, Perplexity: 9.6030
Epoch [3/3], Step [798/3236], Loss: 1.9622, Perplexity: 7.1153
Epoch [3/3], Step [799/3236], Loss: 2.2664, Perplexity: 9.6445
Epoch [3/3], Step [800/3236], Loss: 2.1581, Perplexity: 8.6549
Epoch [3/3], Step [801/3236], Loss: 2.0305, Perplexity: 7.6182
Epoch [3/3], Step [802/3236], Loss: 2.2674, Perplexity: 9.6539
Epoch [3/3], Step [803/3236], Loss: 2.1757, Perplexity: 8.8083
Epoch [3/3], Step [804/3236], Loss: 1.8933, Perplexity: 6.6412
Epoch [3/3], Step [805/3236], Loss: 1.9190, Perplexity: 6.8141
Epoch [3/3], Step [806/3236], Loss: 1.8793, Perplexity: 6.5491
Epoch [3/3], Step [807/3236], Loss: 1.9657, Perplexity: 7.1396
Epoch [3/3], Step [808/3236], Loss: 1.9312, Perplexity: 6.8978
Epoch [3/3], Step [809/3236], Loss: 1.9959, Perplexity: 7.3592
Epoch [3/3], Step [810/3236], Loss: 1.9306, Perplexity: 6.8936
Epoch [3/3], Step [811/3236], Loss: 1.9224, Perplexity: 6.8376
Epoch [3/3], Step [812/3236], Loss: 1.9337, Perplexity: 6.9147
Epoch [3/3], Step [813/3236], Loss: 2.3098, Perplexity: 10.0720
Epoch [3/3], Step [814/3236], Loss: 2.2168, Perplexity: 9.1779
Epoch [3/3], Step [815/3236], Loss: 2.0553, Perplexity: 7.8093
Epoch [3/3], Step [816/3236], Loss: 1.8205, Perplexity: 6.1749
Epoch [3/3], Step [817/3236], Loss: 2.0331, Perplexity: 7.6377
Epoch [3/3], Step [818/3236], Loss: 1.8839, Perplexity: 6.5793
Epoch [3/3], Step [819/3236], Loss: 2.2374, Perplexity: 9.3687
Epoch [3/3], Step [820/3236], Loss: 2.0231, Perplexity: 7.5618
Epoch [3/3], Step [821/3236], Loss: 2.5352, Perplexity: 12.6185
Epoch [3/3], Step [822/3236], Loss: 2.0061, Perplexity: 7.4341
Epoch [3/3], Step [823/3236], Loss: 2.1006, Perplexity: 8.1710
Epoch [3/3], Step [824/3236], Loss: 1.8197, Perplexity: 6.1700
Epoch [3/3], Step [825/3236], Loss: 1.8506, Perplexity: 6.3637
Epoch [3/3], Step [826/3236], Loss: 1.9627, Perplexity: 7.1186
Epoch [3/3], Step [827/3236], Loss: 2.2169, Perplexity: 9.1790
Epoch [3/3], Step [828/3236], Loss: 1.9455, Perplexity: 6.9972
Epoch [3/3], Step [829/3236], Loss: 1.9444, Perplexity: 6.9892
Epoch [3/3], Step [830/3236], Loss: 2.0186, Perplexity: 7.5281
Epoch [3/3], Step [831/3236], Loss: 1.9953, Perplexity: 7.3543
Epoch [3/3], Step [832/3236], Loss: 2.0221, Perplexity: 7.5545
Epoch [3/3], Step [833/3236], Loss: 1.8825, Perplexity: 6.5701
Epoch [3/3], Step [834/3236], Loss: 2.0250, Perplexity: 7.5763
Epoch [3/3], Step [835/3236], Loss: 1.9118, Perplexity: 6.7654
Epoch [3/3], Step [836/3236], Loss: 1.8015, Perplexity: 6.0587
Epoch [3/3], Step [837/3236], Loss: 2.0964, Perplexity: 8.1367
Epoch [3/3], Step [838/3236], Loss: 2.1243, Perplexity: 8.3671
Epoch [3/3], Step [839/3236], Loss: 1.9340, Perplexity: 6.9171
Epoch [3/3], Step [840/3236], Loss: 2.6398, Perplexity: 14.0107
Epoch [3/3], Step [841/3236], Loss: 1.9048, Perplexity: 6.7182
Epoch [3/3], Step [842/3236], Loss: 2.0584, Perplexity: 7.8336
Epoch [3/3], Step [843/3236], Loss: 1.8748, Perplexity: 6.5197
Epoch [3/3], Step [844/3236], Loss: 2.3074, Perplexity: 10.0486
Epoch [3/3], Step [845/3236], Loss: 2.0381, Perplexity: 7.6761
Epoch [3/3], Step [846/3236], Loss: 1.8693, Perplexity: 6.4839
Epoch [3/3], Step [847/3236], Loss: 2.7454, Perplexity: 15.5712
Epoch [3/3], Step [848/3236], Loss: 1.8967, Perplexity: 6.6637
Epoch [3/3], Step [849/3236], Loss: 1.9708, Perplexity: 7.1762
Epoch [3/3], Step [850/3236], Loss: 1.9946, Perplexity: 7.3493
Epoch [3/3], Step [851/3236], Loss: 2.0280, Perplexity: 7.5992
Epoch [3/3], Step [852/3236], Loss: 2.2294, Perplexity: 9.2942
Epoch [3/3], Step [853/3236], Loss: 1.9770, Perplexity: 7.2208
Epoch [3/3], Step [854/3236], Loss: 2.0013, Perplexity: 7.3983
Epoch [3/3], Step [855/3236], Loss: 2.2254, Perplexity: 9.2573
Epoch [3/3], Step [856/3236], Loss: 2.0789, Perplexity: 7.9954
Epoch [3/3], Step [857/3236], Loss: 1.9083, Perplexity: 6.7417
Epoch [3/3], Step [858/3236], Loss: 2.0448, Perplexity: 7.7273
Epoch [3/3], Step [859/3236], Loss: 2.0076, Perplexity: 7.4456
Epoch [3/3], Step [860/3236], Loss: 1.9721, Perplexity: 7.1857
Epoch [3/3], Step [861/3236], Loss: 2.3144, Perplexity: 10.1190
Epoch [3/3], Step [862/3236], Loss: 2.0977, Perplexity: 8.1472
Epoch [3/3], Step [863/3236], Loss: 1.8734, Perplexity: 6.5102
Epoch [3/3], Step [864/3236], Loss: 2.0099, Perplexity: 7.4627
Epoch [3/3], Step [865/3236], Loss: 1.8997, Perplexity: 6.6840
Epoch [3/3], Step [866/3236], Loss: 2.2076, Perplexity: 9.0935
Epoch [3/3], Step [867/3236], Loss: 2.0596, Perplexity: 7.8427
Epoch [3/3], Step [868/3236], Loss: 2.0797, Perplexity: 8.0022
Epoch [3/3], Step [869/3236], Loss: 2.1705, Perplexity: 8.7622
Epoch [3/3], Step [870/3236], Loss: 2.1127, Perplexity: 8.2706
Epoch [3/3], Step [871/3236], Loss: 1.8485, Perplexity: 6.3500
Epoch [3/3], Step [872/3236], Loss: 2.0069, Perplexity: 7.4404
Epoch [3/3], Step [873/3236], Loss: 2.1479, Perplexity: 8.5672
Epoch [3/3], Step [874/3236], Loss: 2.0611, Perplexity: 7.8544
Epoch [3/3], Step [875/3236], Loss: 1.9099, Perplexity: 6.7525
Epoch [3/3], Step [876/3236], Loss: 1.9831, Perplexity: 7.2655
Epoch [3/3], Step [877/3236], Loss: 1.9561, Perplexity: 7.0716
Epoch [3/3], Step [878/3236], Loss: 2.1393, Perplexity: 8.4936
Epoch [3/3], Step [879/3236], Loss: 2.0833, Perplexity: 8.0312
Epoch [3/3], Step [880/3236], Loss: 1.9664, Perplexity: 7.1449
Epoch [3/3], Step [881/3236], Loss: 1.9522, Perplexity: 7.0439
Epoch [3/3], Step [882/3236], Loss: 1.9211, Perplexity: 6.8284
Epoch [3/3], Step [883/3236], Loss: 1.9340, Perplexity: 6.9173
Epoch [3/3], Step [884/3236], Loss: 2.4080, Perplexity: 11.1114
Epoch [3/3], Step [885/3236], Loss: 1.8852, Perplexity: 6.5876
Epoch [3/3], Step [886/3236], Loss: 1.9678, Perplexity: 7.1548
Epoch [3/3], Step [887/3236], Loss: 2.9987, Perplexity: 20.0592
Epoch [3/3], Step [888/3236], Loss: 2.0319, Perplexity: 7.6289
Epoch [3/3], Step [889/3236], Loss: 2.2498, Perplexity: 9.4856
Epoch [3/3], Step [890/3236], Loss: 2.0319, Perplexity: 7.6286
Epoch [3/3], Step [891/3236], Loss: 3.0685, Perplexity: 21.5100
Epoch [3/3], Step [892/3236], Loss: 2.0618, Perplexity: 7.8601
Epoch [3/3], Step [893/3236], Loss: 1.9577, Perplexity: 7.0831
Epoch [3/3], Step [894/3236], Loss: 2.2449, Perplexity: 9.4396
Epoch [3/3], Step [895/3236], Loss: 2.4756, Perplexity: 11.8890
Epoch [3/3], Step [896/3236], Loss: 2.1775, Perplexity: 8.8245
Epoch [3/3], Step [897/3236], Loss: 1.8500, Perplexity: 6.3598
Epoch [3/3], Step [898/3236], Loss: 2.0038, Perplexity: 7.4173
Epoch [3/3], Step [899/3236], Loss: 1.8717, Perplexity: 6.4993
Epoch [3/3], Step [900/3236], Loss: 1.8004, Perplexity: 6.0518
Epoch [3/3], Step [901/3236], Loss: 2.0794, Perplexity: 7.9993
Epoch [3/3], Step [902/3236], Loss: 2.0591, Perplexity: 7.8388
Epoch [3/3], Step [903/3236], Loss: 1.9614, Perplexity: 7.1095
Epoch [3/3], Step [904/3236], Loss: 2.3048, Perplexity: 10.0225
Epoch [3/3], Step [905/3236], Loss: 2.0411, Perplexity: 7.6994
Epoch [3/3], Step [906/3236], Loss: 1.8963, Perplexity: 6.6615
Epoch [3/3], Step [907/3236], Loss: 2.2353, Perplexity: 9.3492
Epoch [3/3], Step [908/3236], Loss: 1.9302, Perplexity: 6.8907
Epoch [3/3], Step [909/3236], Loss: 2.1522, Perplexity: 8.6035
Epoch [3/3], Step [910/3236], Loss: 1.9498, Perplexity: 7.0270
Epoch [3/3], Step [911/3236], Loss: 1.9865, Perplexity: 7.2900
Epoch [3/3], Step [912/3236], Loss: 2.0481, Perplexity: 7.7530
Epoch [3/3], Step [913/3236], Loss: 1.9152, Perplexity: 6.7886
Epoch [3/3], Step [914/3236], Loss: 1.9192, Perplexity: 6.8157
Epoch [3/3], Step [915/3236], Loss: 2.0777, Perplexity: 7.9863
Epoch [3/3], Step [916/3236], Loss: 1.8778, Perplexity: 6.5391
Epoch [3/3], Step [917/3236], Loss: 2.1024, Perplexity: 8.1855
Epoch [3/3], Step [918/3236], Loss: 2.2031, Perplexity: 9.0529
Epoch [3/3], Step [919/3236], Loss: 2.0911, Perplexity: 8.0940
Epoch [3/3], Step [920/3236], Loss: 2.0680, Perplexity: 7.9087
Epoch [3/3], Step [921/3236], Loss: 1.9612, Perplexity: 7.1079
Epoch [3/3], Step [922/3236], Loss: 1.8893, Perplexity: 6.6145
Epoch [3/3], Step [923/3236], Loss: 1.8793, Perplexity: 6.5492
Epoch [3/3], Step [924/3236], Loss: 1.9665, Perplexity: 7.1454
Epoch [3/3], Step [925/3236], Loss: 1.9540, Perplexity: 7.0569
Epoch [3/3], Step [926/3236], Loss: 2.3163, Perplexity: 10.1384
Epoch [3/3], Step [927/3236], Loss: 2.0715, Perplexity: 7.9368
Epoch [3/3], Step [928/3236], Loss: 1.9949, Perplexity: 7.3514
Epoch [3/3], Step [929/3236], Loss: 2.0116, Perplexity: 7.4751
Epoch [3/3], Step [930/3236], Loss: 1.9665, Perplexity: 7.1458
Epoch [3/3], Step [931/3236], Loss: 1.9237, Perplexity: 6.8465
Epoch [3/3], Step [932/3236], Loss: 2.1713, Perplexity: 8.7699
Epoch [3/3], Step [933/3236], Loss: 1.9451, Perplexity: 6.9942
Epoch [3/3], Step [934/3236], Loss: 1.8948, Perplexity: 6.6510
Epoch [3/3], Step [935/3236], Loss: 1.9063, Perplexity: 6.7278
Epoch [3/3], Step [936/3236], Loss: 1.9187, Perplexity: 6.8123
Epoch [3/3], Step [937/3236], Loss: 2.0611, Perplexity: 7.8546
Epoch [3/3], Step [938/3236], Loss: 2.1438, Perplexity: 8.5318
Epoch [3/3], Step [939/3236], Loss: 1.9304, Perplexity: 6.8923
Epoch [3/3], Step [940/3236], Loss: 1.9455, Perplexity: 6.9969
Epoch [3/3], Step [941/3236], Loss: 1.8875, Perplexity: 6.6027
Epoch [3/3], Step [942/3236], Loss: 2.2214, Perplexity: 9.2202
Epoch [3/3], Step [943/3236], Loss: 2.0050, Perplexity: 7.4259
Epoch [3/3], Step [944/3236], Loss: 1.8881, Perplexity: 6.6070
Epoch [3/3], Step [945/3236], Loss: 2.1783, Perplexity: 8.8312
Epoch [3/3], Step [946/3236], Loss: 1.9391, Perplexity: 6.9528
Epoch [3/3], Step [947/3236], Loss: 2.1279, Perplexity: 8.3970
Epoch [3/3], Step [948/3236], Loss: 1.9214, Perplexity: 6.8303
Epoch [3/3], Step [949/3236], Loss: 1.9794, Perplexity: 7.2387
Epoch [3/3], Step [950/3236], Loss: 1.9978, Perplexity: 7.3728
Epoch [3/3], Step [951/3236], Loss: 2.0730, Perplexity: 7.9483
Epoch [3/3], Step [952/3236], Loss: 1.9425, Perplexity: 6.9763
Epoch [3/3], Step [953/3236], Loss: 1.9751, Perplexity: 7.2073
Epoch [3/3], Step [954/3236], Loss: 1.9279, Perplexity: 6.8749
Epoch [3/3], Step [955/3236], Loss: 1.9823, Perplexity: 7.2592
Epoch [3/3], Step [956/3236], Loss: 1.9481, Perplexity: 7.0155
Epoch [3/3], Step [957/3236], Loss: 2.0646, Perplexity: 7.8825
Epoch [3/3], Step [958/3236], Loss: 2.2898, Perplexity: 9.8730
Epoch [3/3], Step [959/3236], Loss: 1.8930, Perplexity: 6.6394
Epoch [3/3], Step [960/3236], Loss: 2.0365, Perplexity: 7.6641
Epoch [3/3], Step [961/3236], Loss: 2.6246, Perplexity: 13.7997
Epoch [3/3], Step [962/3236], Loss: 1.9649, Perplexity: 7.1340
Epoch [3/3], Step [963/3236], Loss: 1.8307, Perplexity: 6.2382
Epoch [3/3], Step [964/3236], Loss: 2.2526, Perplexity: 9.5127
Epoch [3/3], Step [965/3236], Loss: 1.9002, Perplexity: 6.6869
Epoch [3/3], Step [966/3236], Loss: 2.1221, Perplexity: 8.3488
Epoch [3/3], Step [967/3236], Loss: 1.9329, Perplexity: 6.9095
Epoch [3/3], Step [968/3236], Loss: 1.9769, Perplexity: 7.2203
Epoch [3/3], Step [969/3236], Loss: 1.9347, Perplexity: 6.9219
Epoch [3/3], Step [970/3236], Loss: 1.8289, Perplexity: 6.2273
Epoch [3/3], Step [971/3236], Loss: 2.0595, Perplexity: 7.8418
Epoch [3/3], Step [972/3236], Loss: 2.0133, Perplexity: 7.4877
Epoch [3/3], Step [973/3236], Loss: 1.9738, Perplexity: 7.1983
Epoch [3/3], Step [974/3236], Loss: 2.0287, Perplexity: 7.6046
Epoch [3/3], Step [975/3236], Loss: 1.8688, Perplexity: 6.4804
Epoch [3/3], Step [976/3236], Loss: 1.9594, Perplexity: 7.0950
Epoch [3/3], Step [977/3236], Loss: 1.9570, Perplexity: 7.0777
Epoch [3/3], Step [978/3236], Loss: 2.0147, Perplexity: 7.4984
Epoch [3/3], Step [979/3236], Loss: 1.9687, Perplexity: 7.1610
Epoch [3/3], Step [980/3236], Loss: 2.0804, Perplexity: 8.0076
Epoch [3/3], Step [981/3236], Loss: 1.9679, Perplexity: 7.1555
Epoch [3/3], Step [982/3236], Loss: 1.8010, Perplexity: 6.0559
Epoch [3/3], Step [983/3236], Loss: 2.2048, Perplexity: 9.0682
Epoch [3/3], Step [984/3236], Loss: 1.8544, Perplexity: 6.3877
Epoch [3/3], Step [985/3236], Loss: 2.0062, Perplexity: 7.4350
Epoch [3/3], Step [986/3236], Loss: 1.8833, Perplexity: 6.5748
Epoch [3/3], Step [987/3236], Loss: 1.9321, Perplexity: 6.9037
Epoch [3/3], Step [988/3236], Loss: 1.9882, Perplexity: 7.3020
Epoch [3/3], Step [989/3236], Loss: 1.9266, Perplexity: 6.8658
Epoch [3/3], Step [990/3236], Loss: 1.8906, Perplexity: 6.6234
Epoch [3/3], Step [991/3236], Loss: 2.0279, Perplexity: 7.5982
Epoch [3/3], Step [992/3236], Loss: 2.1090, Perplexity: 8.2397
Epoch [3/3], Step [993/3236], Loss: 2.0218, Perplexity: 7.5516
Epoch [3/3], Step [994/3236], Loss: 2.1779, Perplexity: 8.8274
Epoch [3/3], Step [995/3236], Loss: 1.9386, Perplexity: 6.9488
Epoch [3/3], Step [996/3236], Loss: 2.1975, Perplexity: 9.0027
Epoch [3/3], Step [997/3236], Loss: 2.0896, Perplexity: 8.0817
Epoch [3/3], Step [998/3236], Loss: 1.8655, Perplexity: 6.4591
Epoch [3/3], Step [999/3236], Loss: 1.9186, Perplexity: 6.8115
Epoch [3/3], Step [1000/3236], Loss: 1.9173, Perplexity: 6.8026
Epoch [3/3], Step [1001/3236], Loss: 1.8475, Perplexity: 6.3441
Epoch [3/3], Step [1002/3236], Loss: 1.9608, Perplexity: 7.1053
Epoch [3/3], Step [1003/3236], Loss: 1.9493, Perplexity: 7.0237
Epoch [3/3], Step [1004/3236], Loss: 1.9695, Perplexity: 7.1673
Epoch [3/3], Step [1005/3236], Loss: 2.0446, Perplexity: 7.7261
Epoch [3/3], Step [1006/3236], Loss: 2.0447, Perplexity: 7.7267
Epoch [3/3], Step [1007/3236], Loss: 1.9189, Perplexity: 6.8136
Epoch [3/3], Step [1008/3236], Loss: 1.8578, Perplexity: 6.4093
Epoch [3/3], Step [1009/3236], Loss: 2.3201, Perplexity: 10.1766
Epoch [3/3], Step [1010/3236], Loss: 1.9122, Perplexity: 6.7677
Epoch [3/3], Step [1011/3236], Loss: 2.0183, Perplexity: 7.5258
Epoch [3/3], Step [1012/3236], Loss: 1.9320, Perplexity: 6.9031
Epoch [3/3], Step [1013/3236], Loss: 2.0501, Perplexity: 7.7687
Epoch [3/3], Step [1014/3236], Loss: 2.1764, Perplexity: 8.8146
Epoch [3/3], Step [1015/3236], Loss: 1.9037, Perplexity: 6.7108
Epoch [3/3], Step [1016/3236], Loss: 2.0647, Perplexity: 7.8829
Epoch [3/3], Step [1017/3236], Loss: 2.1207, Perplexity: 8.3366
Epoch [3/3], Step [1018/3236], Loss: 2.0625, Perplexity: 7.8654
Epoch [3/3], Step [1019/3236], Loss: 2.2919, Perplexity: 9.8934
Epoch [3/3], Step [1020/3236], Loss: 2.2065, Perplexity: 9.0838
Epoch [3/3], Step [1021/3236], Loss: 1.9585, Perplexity: 7.0887
Epoch [3/3], Step [1022/3236], Loss: 2.2865, Perplexity: 9.8405
Epoch [3/3], Step [1023/3236], Loss: 2.0513, Perplexity: 7.7780
Epoch [3/3], Step [1024/3236], Loss: 1.9992, Perplexity: 7.3834
Epoch [3/3], Step [1025/3236], Loss: 2.2176, Perplexity: 9.1851
Epoch [3/3], Step [1026/3236], Loss: 1.9624, Perplexity: 7.1162
Epoch [3/3], Step [1027/3236], Loss: 2.0978, Perplexity: 8.1479
Epoch [3/3], Step [1028/3236], Loss: 1.8916, Perplexity: 6.6301
Epoch [3/3], Step [1029/3236], Loss: 2.0076, Perplexity: 7.4451
Epoch [3/3], Step [1030/3236], Loss: 1.9544, Perplexity: 7.0597
Epoch [3/3], Step [1031/3236], Loss: 2.7322, Perplexity: 15.3660
Epoch [3/3], Step [1032/3236], Loss: 1.9664, Perplexity: 7.1449
Epoch [3/3], Step [1033/3236], Loss: 1.9717, Perplexity: 7.1832
Epoch [3/3], Step [1034/3236], Loss: 2.0106, Perplexity: 7.4677
Epoch [3/3], Step [1035/3236], Loss: 1.9079, Perplexity: 6.7386
Epoch [3/3], Step [1036/3236], Loss: 1.9590, Perplexity: 7.0923
Epoch [3/3], Step [1037/3236], Loss: 2.1628, Perplexity: 8.6957
Epoch [3/3], Step [1038/3236], Loss: 2.1325, Perplexity: 8.4358
Epoch [3/3], Step [1039/3236], Loss: 1.7450, Perplexity: 5.7260
Epoch [3/3], Step [1040/3236], Loss: 1.9310, Perplexity: 6.8962
Epoch [3/3], Step [1041/3236], Loss: 2.1882, Perplexity: 8.9187
Epoch [3/3], Step [1042/3236], Loss: 1.8742, Perplexity: 6.5156
Epoch [3/3], Step [1043/3236], Loss: 2.2175, Perplexity: 9.1847
Epoch [3/3], Step [1044/3236], Loss: 2.1447, Perplexity: 8.5397
Epoch [3/3], Step [1045/3236], Loss: 1.8357, Perplexity: 6.2698
Epoch [3/3], Step [1046/3236], Loss: 1.9263, Perplexity: 6.8643
Epoch [3/3], Step [1047/3236], Loss: 2.2360, Perplexity: 9.3559
Epoch [3/3], Step [1048/3236], Loss: 1.9773, Perplexity: 7.2234
Epoch [3/3], Step [1049/3236], Loss: 1.8739, Perplexity: 6.5138
Epoch [3/3], Step [1050/3236], Loss: 1.8631, Perplexity: 6.4439
Epoch [3/3], Step [1051/3236], Loss: 1.9648, Perplexity: 7.1333
Epoch [3/3], Step [1052/3236], Loss: 2.0901, Perplexity: 8.0859
Epoch [3/3], Step [1053/3236], Loss: 2.0120, Perplexity: 7.4785
Epoch [3/3], Step [1054/3236], Loss: 2.0572, Perplexity: 7.8243
Epoch [3/3], Step [1055/3236], Loss: 2.0412, Perplexity: 7.6998
Epoch [3/3], Step [1056/3236], Loss: 2.0566, Perplexity: 7.8197
Epoch [3/3], Step [1057/3236], Loss: 1.9547, Perplexity: 7.0615
Epoch [3/3], Step [1058/3236], Loss: 1.9555, Perplexity: 7.0675
Epoch [3/3], Step [1059/3236], Loss: 2.1273, Perplexity: 8.3923
Epoch [3/3], Step [1060/3236], Loss: 2.2380, Perplexity: 9.3750
Epoch [3/3], Step [1061/3236], Loss: 2.1259, Perplexity: 8.3807
Epoch [3/3], Step [1062/3236], Loss: 1.9239, Perplexity: 6.8477
Epoch [3/3], Step [1063/3236], Loss: 2.0330, Perplexity: 7.6367
Epoch [3/3], Step [1064/3236], Loss: 2.5177, Perplexity: 12.4002
Epoch [3/3], Step [1065/3236], Loss: 1.9489, Perplexity: 7.0213
Epoch [3/3], Step [1066/3236], Loss: 2.1636, Perplexity: 8.7021
Epoch [3/3], Step [1067/3236], Loss: 1.9492, Perplexity: 7.0233
Epoch [3/3], Step [1068/3236], Loss: 1.9324, Perplexity: 6.9059
Epoch [3/3], Step [1069/3236], Loss: 1.9874, Perplexity: 7.2965
Epoch [3/3], Step [1070/3236], Loss: 2.0022, Perplexity: 7.4055
Epoch [3/3], Step [1071/3236], Loss: 2.6236, Perplexity: 13.7847
Epoch [3/3], Step [1072/3236], Loss: 1.9808, Perplexity: 7.2483
Epoch [3/3], Step [1073/3236], Loss: 1.9327, Perplexity: 6.9081
Epoch [3/3], Step [1074/3236], Loss: 1.9502, Perplexity: 7.0303
Epoch [3/3], Step [1075/3236], Loss: 1.8361, Perplexity: 6.2718
Epoch [3/3], Step [1076/3236], Loss: 2.0080, Perplexity: 7.4484
Epoch [3/3], Step [1077/3236], Loss: 2.0558, Perplexity: 7.8130
Epoch [3/3], Step [1078/3236], Loss: 2.3745, Perplexity: 10.7456
Epoch [3/3], Step [1079/3236], Loss: 1.9258, Perplexity: 6.8605
Epoch [3/3], Step [1080/3236], Loss: 2.1628, Perplexity: 8.6958
Epoch [3/3], Step [1081/3236], Loss: 1.8915, Perplexity: 6.6292
Epoch [3/3], Step [1082/3236], Loss: 1.9320, Perplexity: 6.9031
Epoch [3/3], Step [1083/3236], Loss: 2.0167, Perplexity: 7.5136
Epoch [3/3], Step [1084/3236], Loss: 1.9597, Perplexity: 7.0974
Epoch [3/3], Step [1085/3236], Loss: 2.1498, Perplexity: 8.5830
Epoch [3/3], Step [1086/3236], Loss: 2.1857, Perplexity: 8.8972
Epoch [3/3], Step [1087/3236], Loss: 1.9506, Perplexity: 7.0329
Epoch [3/3], Step [1088/3236], Loss: 1.9065, Perplexity: 6.7294
Epoch [3/3], Step [1089/3236], Loss: 2.0496, Perplexity: 7.7645
Epoch [3/3], Step [1090/3236], Loss: 1.9744, Perplexity: 7.2024
Epoch [3/3], Step [1091/3236], Loss: 1.9715, Perplexity: 7.1814
Epoch [3/3], Step [1092/3236], Loss: 2.1016, Perplexity: 8.1795
Epoch [3/3], Step [1093/3236], Loss: 1.9001, Perplexity: 6.6867
Epoch [3/3], Step [1094/3236], Loss: 2.0632, Perplexity: 7.8712
Epoch [3/3], Step [1095/3236], Loss: 1.9552, Perplexity: 7.0653
Epoch [3/3], Step [1096/3236], Loss: 1.9185, Perplexity: 6.8109
Epoch [3/3], Step [1097/3236], Loss: 2.0526, Perplexity: 7.7883
Epoch [3/3], Step [1098/3236], Loss: 1.9457, Perplexity: 6.9988
Epoch [3/3], Step [1099/3236], Loss: 1.9175, Perplexity: 6.8042
Epoch [3/3], Step [1100/3236], Loss: 2.0419, Perplexity: 7.7053
Epoch [3/3], Step [1101/3236], Loss: 1.9058, Perplexity: 6.7250
Epoch [3/3], Step [1102/3236], Loss: 1.8986, Perplexity: 6.6767
Epoch [3/3], Step [1103/3236], Loss: 1.9525, Perplexity: 7.0464
Epoch [3/3], Step [1104/3236], Loss: 2.0243, Perplexity: 7.5708
Epoch [3/3], Step [1105/3236], Loss: 1.9829, Perplexity: 7.2638
Epoch [3/3], Step [1106/3236], Loss: 1.9909, Perplexity: 7.3218
Epoch [3/3], Step [1107/3236], Loss: 1.9047, Perplexity: 6.7173
Epoch [3/3], Step [1108/3236], Loss: 2.1444, Perplexity: 8.5372
Epoch [3/3], Step [1109/3236], Loss: 2.1867, Perplexity: 8.9056
Epoch [3/3], Step [1110/3236], Loss: 2.0563, Perplexity: 7.8172
Epoch [3/3], Step [1111/3236], Loss: 2.0937, Perplexity: 8.1147
Epoch [3/3], Step [1112/3236], Loss: 1.8709, Perplexity: 6.4942
Epoch [3/3], Step [1113/3236], Loss: 2.1225, Perplexity: 8.3521
Epoch [3/3], Step [1114/3236], Loss: 2.0804, Perplexity: 8.0075
Epoch [3/3], Step [1115/3236], Loss: 2.0378, Perplexity: 7.6740
Epoch [3/3], Step [1116/3236], Loss: 1.9615, Perplexity: 7.1100
Epoch [3/3], Step [1117/3236], Loss: 1.9986, Perplexity: 7.3790
Epoch [3/3], Step [1118/3236], Loss: 2.0390, Perplexity: 7.6830
Epoch [3/3], Step [1119/3236], Loss: 1.9731, Perplexity: 7.1928
Epoch [3/3], Step [1120/3236], Loss: 1.9685, Perplexity: 7.1599
Epoch [3/3], Step [1121/3236], Loss: 1.9631, Perplexity: 7.1214
Epoch [3/3], Step [1122/3236], Loss: 1.9789, Perplexity: 7.2347
Epoch [3/3], Step [1123/3236], Loss: 2.1027, Perplexity: 8.1879
Epoch [3/3], Step [1124/3236], Loss: 1.8333, Perplexity: 6.2547
Epoch [3/3], Step [1125/3236], Loss: 1.9705, Perplexity: 7.1743
Epoch [3/3], Step [1126/3236], Loss: 1.8776, Perplexity: 6.5377
Epoch [3/3], Step [1127/3236], Loss: 2.0510, Perplexity: 7.7760
Epoch [3/3], Step [1128/3236], Loss: 2.0308, Perplexity: 7.6204
Epoch [3/3], Step [1129/3236], Loss: 1.9815, Perplexity: 7.2533
Epoch [3/3], Step [1130/3236], Loss: 2.4938, Perplexity: 12.1066
Epoch [3/3], Step [1131/3236], Loss: 1.9521, Perplexity: 7.0433
Epoch [3/3], Step [1132/3236], Loss: 1.7882, Perplexity: 5.9789
Epoch [3/3], Step [1133/3236], Loss: 1.9455, Perplexity: 6.9974
Epoch [3/3], Step [1134/3236], Loss: 2.1848, Perplexity: 8.8885
Epoch [3/3], Step [1135/3236], Loss: 1.8737, Perplexity: 6.5122
Epoch [3/3], Step [1136/3236], Loss: 2.3251, Perplexity: 10.2281
Epoch [3/3], Step [1137/3236], Loss: 2.0250, Perplexity: 7.5762
Epoch [3/3], Step [1138/3236], Loss: 1.9626, Perplexity: 7.1179
Epoch [3/3], Step [1139/3236], Loss: 2.2259, Perplexity: 9.2618
Epoch [3/3], Step [1140/3236], Loss: 2.1586, Perplexity: 8.6593
Epoch [3/3], Step [1141/3236], Loss: 1.9325, Perplexity: 6.9070
Epoch [3/3], Step [1142/3236], Loss: 1.9385, Perplexity: 6.9485
Epoch [3/3], Step [1143/3236], Loss: 2.3739, Perplexity: 10.7396
Epoch [3/3], Step [1144/3236], Loss: 2.0301, Perplexity: 7.6147
Epoch [3/3], Step [1145/3236], Loss: 2.0236, Perplexity: 7.5653
Epoch [3/3], Step [1146/3236], Loss: 2.2513, Perplexity: 9.5003
Epoch [3/3], Step [1147/3236], Loss: 2.1639, Perplexity: 8.7054
Epoch [3/3], Step [1148/3236], Loss: 2.1840, Perplexity: 8.8819
Epoch [3/3], Step [1149/3236], Loss: 1.9691, Perplexity: 7.1642
Epoch [3/3], Step [1150/3236], Loss: 1.9071, Perplexity: 6.7334
Epoch [3/3], Step [1151/3236], Loss: 2.3248, Perplexity: 10.2243
Epoch [3/3], Step [1152/3236], Loss: 1.9553, Perplexity: 7.0658
Epoch [3/3], Step [1153/3236], Loss: 2.0339, Perplexity: 7.6437
Epoch [3/3], Step [1154/3236], Loss: 1.8431, Perplexity: 6.3161
Epoch [3/3], Step [1155/3236], Loss: 2.1943, Perplexity: 8.9740
Epoch [3/3], Step [1156/3236], Loss: 2.3805, Perplexity: 10.8103
Epoch [3/3], Step [1157/3236], Loss: 1.9863, Perplexity: 7.2887
Epoch [3/3], Step [1158/3236], Loss: 1.8810, Perplexity: 6.5600
Epoch [3/3], Step [1159/3236], Loss: 1.9722, Perplexity: 7.1862
Epoch [3/3], Step [1160/3236], Loss: 1.8297, Perplexity: 6.2319
Epoch [3/3], Step [1161/3236], Loss: 1.9257, Perplexity: 6.8601
Epoch [3/3], Step [1162/3236], Loss: 2.1141, Perplexity: 8.2823
Epoch [3/3], Step [1163/3236], Loss: 2.1935, Perplexity: 8.9667
Epoch [3/3], Step [1164/3236], Loss: 2.0234, Perplexity: 7.5641
Epoch [3/3], Step [1165/3236], Loss: 2.0134, Perplexity: 7.4891
Epoch [3/3], Step [1166/3236], Loss: 1.9845, Perplexity: 7.2754
Epoch [3/3], Step [1167/3236], Loss: 1.8735, Perplexity: 6.5110
Epoch [3/3], Step [1168/3236], Loss: 2.1127, Perplexity: 8.2702
Epoch [3/3], Step [1169/3236], Loss: 1.9802, Perplexity: 7.2441
Epoch [3/3], Step [1170/3236], Loss: 2.1405, Perplexity: 8.5035
Epoch [3/3], Step [1171/3236], Loss: 1.8741, Perplexity: 6.5149
Epoch [3/3], Step [1172/3236], Loss: 1.9540, Perplexity: 7.0568
Epoch [3/3], Step [1173/3236], Loss: 1.7946, Perplexity: 6.0169
Epoch [3/3], Step [1174/3236], Loss: 1.9227, Perplexity: 6.8392
Epoch [3/3], Step [1175/3236], Loss: 1.9240, Perplexity: 6.8486
Epoch [3/3], Step [1176/3236], Loss: 1.9584, Perplexity: 7.0879
Epoch [3/3], Step [1177/3236], Loss: 3.3185, Perplexity: 27.6177
Epoch [3/3], Step [1178/3236], Loss: 1.9570, Perplexity: 7.0781
Epoch [3/3], Step [1179/3236], Loss: 1.8996, Perplexity: 6.6834
Epoch [3/3], Step [1180/3236], Loss: 2.0681, Perplexity: 7.9101
Epoch [3/3], Step [1181/3236], Loss: 1.9329, Perplexity: 6.9095
Epoch [3/3], Step [1182/3236], Loss: 1.9374, Perplexity: 6.9409
Epoch [3/3], Step [1183/3236], Loss: 1.8610, Perplexity: 6.4300
Epoch [3/3], Step [1184/3236], Loss: 1.9238, Perplexity: 6.8466
Epoch [3/3], Step [1185/3236], Loss: 1.9856, Perplexity: 7.2836
Epoch [3/3], Step [1186/3236], Loss: 1.9152, Perplexity: 6.7881
Epoch [3/3], Step [1187/3236], Loss: 2.0423, Perplexity: 7.7083
Epoch [3/3], Step [1188/3236], Loss: 2.0346, Perplexity: 7.6493
Epoch [3/3], Step [1189/3236], Loss: 1.9718, Perplexity: 7.1837
Epoch [3/3], Step [1190/3236], Loss: 2.2064, Perplexity: 9.0826
Epoch [3/3], Step [1191/3236], Loss: 1.9362, Perplexity: 6.9326
Epoch [3/3], Step [1192/3236], Loss: 1.9508, Perplexity: 7.0344
Epoch [3/3], Step [1193/3236], Loss: 1.9650, Perplexity: 7.1352
Epoch [3/3], Step [1194/3236], Loss: 1.9567, Perplexity: 7.0762
Epoch [3/3], Step [1195/3236], Loss: 2.0468, Perplexity: 7.7431
Epoch [3/3], Step [1196/3236], Loss: 2.0746, Perplexity: 7.9611
Epoch [3/3], Step [1197/3236], Loss: 2.5799, Perplexity: 13.1965
Epoch [3/3], Step [1198/3236], Loss: 1.9469, Perplexity: 7.0067
Epoch [3/3], Step [1199/3236], Loss: 2.0707, Perplexity: 7.9305
Epoch [3/3], Step [1200/3236], Loss: 2.2003, Perplexity: 9.0275
Epoch [3/3], Step [1201/3236], Loss: 2.0350, Perplexity: 7.6520
Epoch [3/3], Step [1202/3236], Loss: 1.9657, Perplexity: 7.1399
Epoch [3/3], Step [1203/3236], Loss: 1.9851, Perplexity: 7.2799
Epoch [3/3], Step [1204/3236], Loss: 2.1349, Perplexity: 8.4560
Epoch [3/3], Step [1205/3236], Loss: 2.0976, Perplexity: 8.1469
Epoch [3/3], Step [1206/3236], Loss: 1.9516, Perplexity: 7.0402
Epoch [3/3], Step [1207/3236], Loss: 2.5240, Perplexity: 12.4783
Epoch [3/3], Step [1208/3236], Loss: 2.1267, Perplexity: 8.3867
Epoch [3/3], Step [1209/3236], Loss: 1.9006, Perplexity: 6.6900
Epoch [3/3], Step [1210/3236], Loss: 1.9421, Perplexity: 6.9732
Epoch [3/3], Step [1211/3236], Loss: 2.1628, Perplexity: 8.6954
Epoch [3/3], Step [1212/3236], Loss: 1.9000, Perplexity: 6.6862
Epoch [3/3], Step [1213/3236], Loss: 2.2792, Perplexity: 9.7690
Epoch [3/3], Step [1214/3236], Loss: 2.1768, Perplexity: 8.8179
Epoch [3/3], Step [1215/3236], Loss: 1.9015, Perplexity: 6.6962
Epoch [3/3], Step [1216/3236], Loss: 2.1649, Perplexity: 8.7140
Epoch [3/3], Step [1217/3236], Loss: 2.0244, Perplexity: 7.5717
Epoch [3/3], Step [1218/3236], Loss: 2.1343, Perplexity: 8.4509
Epoch [3/3], Step [1219/3236], Loss: 2.0762, Perplexity: 7.9745
Epoch [3/3], Step [1220/3236], Loss: 1.9611, Perplexity: 7.1068
Epoch [3/3], Step [1221/3236], Loss: 1.9314, Perplexity: 6.8991
Epoch [3/3], Step [1222/3236], Loss: 1.9921, Perplexity: 7.3306
Epoch [3/3], Step [1223/3236], Loss: 1.9446, Perplexity: 6.9911
Epoch [3/3], Step [1224/3236], Loss: 1.9005, Perplexity: 6.6895
Epoch [3/3], Step [1225/3236], Loss: 2.0859, Perplexity: 8.0519
Epoch [3/3], Step [1226/3236], Loss: 2.0507, Perplexity: 7.7737
Epoch [3/3], Step [1227/3236], Loss: 1.9600, Perplexity: 7.0991
Epoch [3/3], Step [1228/3236], Loss: 1.9772, Perplexity: 7.2226
Epoch [3/3], Step [1229/3236], Loss: 1.9803, Perplexity: 7.2448
Epoch [3/3], Step [1230/3236], Loss: 2.0442, Perplexity: 7.7229
Epoch [3/3], Step [1231/3236], Loss: 1.9988, Perplexity: 7.3801
Epoch [3/3], Step [1232/3236], Loss: 1.8992, Perplexity: 6.6803
Epoch [3/3], Step [1233/3236], Loss: 2.6818, Perplexity: 14.6110
Epoch [3/3], Step [1234/3236], Loss: 1.7980, Perplexity: 6.0375
Epoch [3/3], Step [1235/3236], Loss: 1.9673, Perplexity: 7.1512
Epoch [3/3], Step [1236/3236], Loss: 2.0356, Perplexity: 7.6570
Epoch [3/3], Step [1237/3236], Loss: 2.0248, Perplexity: 7.5746
Epoch [3/3], Step [1238/3236], Loss: 2.1857, Perplexity: 8.8965
Epoch [3/3], Step [1239/3236], Loss: 1.8791, Perplexity: 6.5476
Epoch [3/3], Step [1240/3236], Loss: 1.8644, Perplexity: 6.4518
Epoch [3/3], Step [1241/3236], Loss: 1.9087, Perplexity: 6.7446
Epoch [3/3], Step [1242/3236], Loss: 1.8604, Perplexity: 6.4263
Epoch [3/3], Step [1243/3236], Loss: 1.8873, Perplexity: 6.6013
Epoch [3/3], Step [1244/3236], Loss: 1.9345, Perplexity: 6.9208
Epoch [3/3], Step [1245/3236], Loss: 1.9117, Perplexity: 6.7647
Epoch [3/3], Step [1246/3236], Loss: 1.9504, Perplexity: 7.0316
Epoch [3/3], Step [1247/3236], Loss: 2.6106, Perplexity: 13.6073
Epoch [3/3], Step [1248/3236], Loss: 1.9788, Perplexity: 7.2339
Epoch [3/3], Step [1249/3236], Loss: 1.9226, Perplexity: 6.8387
Epoch [3/3], Step [1250/3236], Loss: 1.8855, Perplexity: 6.5896
Epoch [3/3], Step [1251/3236], Loss: 1.9351, Perplexity: 6.9251
Epoch [3/3], Step [1252/3236], Loss: 2.1871, Perplexity: 8.9098
Epoch [3/3], Step [1253/3236], Loss: 2.0125, Perplexity: 7.4823
Epoch [3/3], Step [1254/3236], Loss: 1.8670, Perplexity: 6.4690
Epoch [3/3], Step [1255/3236], Loss: 1.8778, Perplexity: 6.5390
Epoch [3/3], Step [1256/3236], Loss: 2.1858, Perplexity: 8.8977
Epoch [3/3], Step [1257/3236], Loss: 1.9529, Perplexity: 7.0493
Epoch [3/3], Step [1258/3236], Loss: 2.0972, Perplexity: 8.1432
Epoch [3/3], Step [1259/3236], Loss: 2.0610, Perplexity: 7.8537
Epoch [3/3], Step [1260/3236], Loss: 1.9885, Perplexity: 7.3043
Epoch [3/3], Step [1261/3236], Loss: 1.9832, Perplexity: 7.2658
Epoch [3/3], Step [1262/3236], Loss: 1.8748, Perplexity: 6.5196
Epoch [3/3], Step [1263/3236], Loss: 1.8778, Perplexity: 6.5390
Epoch [3/3], Step [1264/3236], Loss: 1.8949, Perplexity: 6.6520
Epoch [3/3], Step [1265/3236], Loss: 1.9047, Perplexity: 6.7175
Epoch [3/3], Step [1266/3236], Loss: 1.8344, Perplexity: 6.2613
Epoch [3/3], Step [1267/3236], Loss: 1.9265, Perplexity: 6.8658
Epoch [3/3], Step [1268/3236], Loss: 1.8822, Perplexity: 6.5680
Epoch [3/3], Step [1269/3236], Loss: 1.9664, Perplexity: 7.1448
Epoch [3/3], Step [1270/3236], Loss: 2.1279, Perplexity: 8.3971
Epoch [3/3], Step [1271/3236], Loss: 1.9499, Perplexity: 7.0283
Epoch [3/3], Step [1272/3236], Loss: 2.1253, Perplexity: 8.3751
Epoch [3/3], Step [1273/3236], Loss: 1.9709, Perplexity: 7.1768
Epoch [3/3], Step [1274/3236], Loss: 1.9308, Perplexity: 6.8949
Epoch [3/3], Step [1275/3236], Loss: 1.9443, Perplexity: 6.9887
Epoch [3/3], Step [1276/3236], Loss: 2.3031, Perplexity: 10.0053
Epoch [3/3], Step [1277/3236], Loss: 1.8922, Perplexity: 6.6342
Epoch [3/3], Step [1278/3236], Loss: 1.9099, Perplexity: 6.7526
Epoch [3/3], Step [1279/3236], Loss: 2.2217, Perplexity: 9.2233
Epoch [3/3], Step [1280/3236], Loss: 2.0097, Perplexity: 7.4611
Epoch [3/3], Step [1281/3236], Loss: 2.1844, Perplexity: 8.8854
Epoch [3/3], Step [1282/3236], Loss: 2.1567, Perplexity: 8.6430
Epoch [3/3], Step [1283/3236], Loss: 2.0038, Perplexity: 7.4169
Epoch [3/3], Step [1284/3236], Loss: 1.9910, Perplexity: 7.3231
Epoch [3/3], Step [1285/3236], Loss: 2.1188, Perplexity: 8.3212
Epoch [3/3], Step [1286/3236], Loss: 2.0101, Perplexity: 7.4639
Epoch [3/3], Step [1287/3236], Loss: 1.9180, Perplexity: 6.8072
Epoch [3/3], Step [1288/3236], Loss: 2.4479, Perplexity: 11.5636
Epoch [3/3], Step [1289/3236], Loss: 1.9267, Perplexity: 6.8667
Epoch [3/3], Step [1290/3236], Loss: 2.0078, Perplexity: 7.4469
Epoch [3/3], Step [1291/3236], Loss: 1.9405, Perplexity: 6.9623
Epoch [3/3], Step [1292/3236], Loss: 1.8838, Perplexity: 6.5787
Epoch [3/3], Step [1293/3236], Loss: 1.9806, Perplexity: 7.2474
Epoch [3/3], Step [1294/3236], Loss: 1.7898, Perplexity: 5.9884
Epoch [3/3], Step [1295/3236], Loss: 2.4622, Perplexity: 11.7311
Epoch [3/3], Step [1296/3236], Loss: 1.8866, Perplexity: 6.5966
Epoch [3/3], Step [1297/3236], Loss: 1.9869, Perplexity: 7.2932
Epoch [3/3], Step [1298/3236], Loss: 2.0621, Perplexity: 7.8624
Epoch [3/3], Step [1299/3236], Loss: 2.1677, Perplexity: 8.7378
Epoch [3/3], Step [1300/3236], Loss: 2.1380, Perplexity: 8.4822
Epoch [3/3], Step [1301/3236], Loss: 1.9325, Perplexity: 6.9064
Epoch [3/3], Step [1302/3236], Loss: 1.9045, Perplexity: 6.7160
Epoch [3/3], Step [1303/3236], Loss: 1.8485, Perplexity: 6.3505
Epoch [3/3], Step [1304/3236], Loss: 1.8833, Perplexity: 6.5752
Epoch [3/3], Step [1305/3236], Loss: 2.3086, Perplexity: 10.0606
Epoch [3/3], Step [1306/3236], Loss: 1.8171, Perplexity: 6.1538
Epoch [3/3], Step [1307/3236], Loss: 1.9031, Perplexity: 6.7068
Epoch [3/3], Step [1308/3236], Loss: 2.0313, Perplexity: 7.6240
Epoch [3/3], Step [1309/3236], Loss: 2.0125, Perplexity: 7.4820
Epoch [3/3], Step [1310/3236], Loss: 1.9416, Perplexity: 6.9696
Epoch [3/3], Step [1311/3236], Loss: 1.9973, Perplexity: 7.3692
Epoch [3/3], Step [1312/3236], Loss: 2.0466, Perplexity: 7.7412
Epoch [3/3], Step [1313/3236], Loss: 2.0784, Perplexity: 7.9915
Epoch [3/3], Step [1314/3236], Loss: 2.0776, Perplexity: 7.9856
Epoch [3/3], Step [1315/3236], Loss: 2.1094, Perplexity: 8.2437
Epoch [3/3], Step [1316/3236], Loss: 2.5443, Perplexity: 12.7337
Epoch [3/3], Step [1317/3236], Loss: 1.9099, Perplexity: 6.7523
Epoch [3/3], Step [1318/3236], Loss: 2.0038, Perplexity: 7.4173
Epoch [3/3], Step [1319/3236], Loss: 1.8996, Perplexity: 6.6830
Epoch [3/3], Step [1320/3236], Loss: 2.0492, Perplexity: 7.7621
Epoch [3/3], Step [1321/3236], Loss: 2.0141, Perplexity: 7.4938
Epoch [3/3], Step [1322/3236], Loss: 1.8708, Perplexity: 6.4935
Epoch [3/3], Step [1323/3236], Loss: 2.3734, Perplexity: 10.7337
Epoch [3/3], Step [1324/3236], Loss: 2.0195, Perplexity: 7.5349
Epoch [3/3], Step [1325/3236], Loss: 2.0590, Perplexity: 7.8378
Epoch [3/3], Step [1326/3236], Loss: 1.9389, Perplexity: 6.9512
Epoch [3/3], Step [1327/3236], Loss: 2.0876, Perplexity: 8.0659
Epoch [3/3], Step [1328/3236], Loss: 2.3372, Perplexity: 10.3520
Epoch [3/3], Step [1329/3236], Loss: 1.8996, Perplexity: 6.6829
Epoch [3/3], Step [1330/3236], Loss: 2.1432, Perplexity: 8.5269
Epoch [3/3], Step [1331/3236], Loss: 2.1031, Perplexity: 8.1918
Epoch [3/3], Step [1332/3236], Loss: 1.9574, Perplexity: 7.0810
Epoch [3/3], Step [1333/3236], Loss: 2.0540, Perplexity: 7.7987
Epoch [3/3], Step [1334/3236], Loss: 2.1571, Perplexity: 8.6463
Epoch [3/3], Step [1335/3236], Loss: 1.8424, Perplexity: 6.3116
Epoch [3/3], Step [1336/3236], Loss: 1.8858, Perplexity: 6.5915
Epoch [3/3], Step [1337/3236], Loss: 2.0232, Perplexity: 7.5622
Epoch [3/3], Step [1338/3236], Loss: 1.8940, Perplexity: 6.6457
Epoch [3/3], Step [1339/3236], Loss: 1.8381, Perplexity: 6.2843
Epoch [3/3], Step [1340/3236], Loss: 1.8539, Perplexity: 6.3844
Epoch [3/3], Step [1341/3236], Loss: 1.8507, Perplexity: 6.3645
Epoch [3/3], Step [1342/3236], Loss: 1.9368, Perplexity: 6.9362
Epoch [3/3], Step [1343/3236], Loss: 2.2196, Perplexity: 9.2035
Epoch [3/3], Step [1344/3236], Loss: 1.9522, Perplexity: 7.0442
Epoch [3/3], Step [1345/3236], Loss: 2.1405, Perplexity: 8.5037
Epoch [3/3], Step [1346/3236], Loss: 2.0351, Perplexity: 7.6527
Epoch [3/3], Step [1347/3236], Loss: 2.4648, Perplexity: 11.7606
Epoch [3/3], Step [1348/3236], Loss: 1.9638, Perplexity: 7.1264
Epoch [3/3], Step [1349/3236], Loss: 2.0633, Perplexity: 7.8716
Epoch [3/3], Step [1350/3236], Loss: 2.0737, Perplexity: 7.9540
Epoch [3/3], Step [1351/3236], Loss: 2.0004, Perplexity: 7.3921
Epoch [3/3], Step [1352/3236], Loss: 1.9737, Perplexity: 7.1972
Epoch [3/3], Step [1353/3236], Loss: 1.9818, Perplexity: 7.2561
Epoch [3/3], Step [1354/3236], Loss: 1.9632, Perplexity: 7.1222
Epoch [3/3], Step [1355/3236], Loss: 2.3223, Perplexity: 10.1995
Epoch [3/3], Step [1356/3236], Loss: 1.9199, Perplexity: 6.8203
Epoch [3/3], Step [1357/3236], Loss: 1.8794, Perplexity: 6.5497
Epoch [3/3], Step [1358/3236], Loss: 1.9221, Perplexity: 6.8355
Epoch [3/3], Step [1359/3236], Loss: 1.8744, Perplexity: 6.5168
Epoch [3/3], Step [1360/3236], Loss: 1.8702, Perplexity: 6.4893
Epoch [3/3], Step [1361/3236], Loss: 2.4489, Perplexity: 11.5760
Epoch [3/3], Step [1362/3236], Loss: 1.9221, Perplexity: 6.8356
Epoch [3/3], Step [1363/3236], Loss: 2.0413, Perplexity: 7.7006
Epoch [3/3], Step [1364/3236], Loss: 1.9322, Perplexity: 6.9048
Epoch [3/3], Step [1365/3236], Loss: 2.1357, Perplexity: 8.4628
Epoch [3/3], Step [1366/3236], Loss: 1.9420, Perplexity: 6.9727
Epoch [3/3], Step [1367/3236], Loss: 2.2388, Perplexity: 9.3820
Epoch [3/3], Step [1368/3236], Loss: 2.0066, Perplexity: 7.4379
Epoch [3/3], Step [1369/3236], Loss: 1.8805, Perplexity: 6.5571
Epoch [3/3], Step [1370/3236], Loss: 1.8672, Perplexity: 6.4701
Epoch [3/3], Step [1371/3236], Loss: 2.0128, Perplexity: 7.4846
Epoch [3/3], Step [1372/3236], Loss: 1.7424, Perplexity: 5.7111
Epoch [3/3], Step [1373/3236], Loss: 1.9714, Perplexity: 7.1807
Epoch [3/3], Step [1374/3236], Loss: 1.8009, Perplexity: 6.0549
Epoch [3/3], Step [1375/3236], Loss: 2.0010, Perplexity: 7.3964
Epoch [3/3], Step [1376/3236], Loss: 1.8605, Perplexity: 6.4267
Epoch [3/3], Step [1377/3236], Loss: 1.9928, Perplexity: 7.3363
Epoch [3/3], Step [1378/3236], Loss: 2.0475, Perplexity: 7.7483
Epoch [3/3], Step [1379/3236], Loss: 2.0569, Perplexity: 7.8218
Epoch [3/3], Step [1380/3236], Loss: 2.0284, Perplexity: 7.6018
Epoch [3/3], Step [1381/3236], Loss: 1.9215, Perplexity: 6.8313
Epoch [3/3], Step [1382/3236], Loss: 1.8312, Perplexity: 6.2414
Epoch [3/3], Step [1383/3236], Loss: 2.2267, Perplexity: 9.2688
Epoch [3/3], Step [1384/3236], Loss: 1.8246, Perplexity: 6.2000
Epoch [3/3], Step [1385/3236], Loss: 2.0470, Perplexity: 7.7445
Epoch [3/3], Step [1386/3236], Loss: 1.9350, Perplexity: 6.9239
Epoch [3/3], Step [1387/3236], Loss: 2.3114, Perplexity: 10.0890
Epoch [3/3], Step [1388/3236], Loss: 1.9884, Perplexity: 7.3036
Epoch [3/3], Step [1389/3236], Loss: 1.9231, Perplexity: 6.8422
Epoch [3/3], Step [1390/3236], Loss: 1.9673, Perplexity: 7.1514
Epoch [3/3], Step [1391/3236], Loss: 1.8800, Perplexity: 6.5533
Epoch [3/3], Step [1392/3236], Loss: 1.8578, Perplexity: 6.4094
Epoch [3/3], Step [1393/3236], Loss: 1.9427, Perplexity: 6.9773
Epoch [3/3], Step [1394/3236], Loss: 2.1046, Perplexity: 8.2038
Epoch [3/3], Step [1395/3236], Loss: 1.9422, Perplexity: 6.9742
Epoch [3/3], Step [1396/3236], Loss: 1.9391, Perplexity: 6.9524
Epoch [3/3], Step [1397/3236], Loss: 2.0202, Perplexity: 7.5397
Epoch [3/3], Step [1398/3236], Loss: 1.9207, Perplexity: 6.8259
Epoch [3/3], Step [1399/3236], Loss: 1.8678, Perplexity: 6.4743
Epoch [3/3], Step [1400/3236], Loss: 2.8509, Perplexity: 17.3027
Epoch [3/3], Step [1401/3236], Loss: 2.1130, Perplexity: 8.2729
Epoch [3/3], Step [1402/3236], Loss: 2.2844, Perplexity: 9.8198
Epoch [3/3], Step [1403/3236], Loss: 1.9176, Perplexity: 6.8048
Epoch [3/3], Step [1404/3236], Loss: 2.0360, Perplexity: 7.6602
Epoch [3/3], Step [1405/3236], Loss: 1.9570, Perplexity: 7.0777
Epoch [3/3], Step [1406/3236], Loss: 2.0439, Perplexity: 7.7205
Epoch [3/3], Step [1407/3236], Loss: 1.9181, Perplexity: 6.8081
Epoch [3/3], Step [1408/3236], Loss: 2.3135, Perplexity: 10.1095
Epoch [3/3], Step [1409/3236], Loss: 1.8795, Perplexity: 6.5501
Epoch [3/3], Step [1410/3236], Loss: 1.9227, Perplexity: 6.8396
Epoch [3/3], Step [1411/3236], Loss: 2.1423, Perplexity: 8.5189
Epoch [3/3], Step [1412/3236], Loss: 1.9759, Perplexity: 7.2129
Epoch [3/3], Step [1413/3236], Loss: 2.0817, Perplexity: 8.0184
Epoch [3/3], Step [1414/3236], Loss: 1.9385, Perplexity: 6.9481
Epoch [3/3], Step [1415/3236], Loss: 1.9172, Perplexity: 6.8017
Epoch [3/3], Step [1416/3236], Loss: 2.1657, Perplexity: 8.7208
Epoch [3/3], Step [1417/3236], Loss: 1.8821, Perplexity: 6.5671
Epoch [3/3], Step [1418/3236], Loss: 1.8017, Perplexity: 6.0600
Epoch [3/3], Step [1419/3236], Loss: 1.9033, Perplexity: 6.7079
Epoch [3/3], Step [1420/3236], Loss: 2.2003, Perplexity: 9.0273
Epoch [3/3], Step [1421/3236], Loss: 1.9292, Perplexity: 6.8842
Epoch [3/3], Step [1422/3236], Loss: 1.9847, Perplexity: 7.2769
Epoch [3/3], Step [1423/3236], Loss: 1.8768, Perplexity: 6.5327
Epoch [3/3], Step [1424/3236], Loss: 1.8773, Perplexity: 6.5358
Epoch [3/3], Step [1425/3236], Loss: 2.0745, Perplexity: 7.9605
Epoch [3/3], Step [1426/3236], Loss: 2.2217, Perplexity: 9.2230
Epoch [3/3], Step [1427/3236], Loss: 1.8729, Perplexity: 6.5070
Epoch [3/3], Step [1428/3236], Loss: 2.1530, Perplexity: 8.6103
Epoch [3/3], Step [1429/3236], Loss: 2.0038, Perplexity: 7.4174
Epoch [3/3], Step [1430/3236], Loss: 2.0086, Perplexity: 7.4532
Epoch [3/3], Step [1431/3236], Loss: 2.5338, Perplexity: 12.6015
Epoch [3/3], Step [1432/3236], Loss: 1.9673, Perplexity: 7.1514
Epoch [3/3], Step [1433/3236], Loss: 2.0753, Perplexity: 7.9669
Epoch [3/3], Step [1434/3236], Loss: 2.1188, Perplexity: 8.3207
Epoch [3/3], Step [1435/3236], Loss: 2.1091, Perplexity: 8.2408
Epoch [3/3], Step [1436/3236], Loss: 2.1128, Perplexity: 8.2717
Epoch [3/3], Step [1437/3236], Loss: 1.9538, Perplexity: 7.0553
Epoch [3/3], Step [1438/3236], Loss: 2.0230, Perplexity: 7.5611
Epoch [3/3], Step [1439/3236], Loss: 1.9346, Perplexity: 6.9216
Epoch [3/3], Step [1440/3236], Loss: 1.8299, Perplexity: 6.2331
Epoch [3/3], Step [1441/3236], Loss: 2.1551, Perplexity: 8.6289
Epoch [3/3], Step [1442/3236], Loss: 2.0017, Perplexity: 7.4013
Epoch [3/3], Step [1443/3236], Loss: 2.0751, Perplexity: 7.9651
Epoch [3/3], Step [1444/3236], Loss: 2.3225, Perplexity: 10.2016
Epoch [3/3], Step [1445/3236], Loss: 2.1284, Perplexity: 8.4013
Epoch [3/3], Step [1446/3236], Loss: 1.9157, Perplexity: 6.7917
Epoch [3/3], Step [1447/3236], Loss: 2.0059, Perplexity: 7.4325
Epoch [3/3], Step [1448/3236], Loss: 1.9011, Perplexity: 6.6933
Epoch [3/3], Step [1449/3236], Loss: 1.8912, Perplexity: 6.6271
Epoch [3/3], Step [1450/3236], Loss: 2.1918, Perplexity: 8.9510
Epoch [3/3], Step [1451/3236], Loss: 1.9919, Perplexity: 7.3293
Epoch [3/3], Step [1452/3236], Loss: 2.0179, Perplexity: 7.5223
Epoch [3/3], Step [1453/3236], Loss: 2.0708, Perplexity: 7.9311
Epoch [3/3], Step [1454/3236], Loss: 1.9612, Perplexity: 7.1076
Epoch [3/3], Step [1455/3236], Loss: 1.9857, Perplexity: 7.2843
Epoch [3/3], Step [1456/3236], Loss: 1.8917, Perplexity: 6.6303
Epoch [3/3], Step [1457/3236], Loss: 1.9198, Perplexity: 6.8198
Epoch [3/3], Step [1458/3236], Loss: 1.9047, Perplexity: 6.7172
Epoch [3/3], Step [1459/3236], Loss: 1.8977, Perplexity: 6.6702
Epoch [3/3], Step [1460/3236], Loss: 1.9642, Perplexity: 7.1290
Epoch [3/3], Step [1461/3236], Loss: 1.9505, Perplexity: 7.0324
Epoch [3/3], Step [1462/3236], Loss: 1.9412, Perplexity: 6.9669
Epoch [3/3], Step [1463/3236], Loss: 2.2123, Perplexity: 9.1365
Epoch [3/3], Step [1464/3236], Loss: 2.0913, Perplexity: 8.0955
Epoch [3/3], Step [1465/3236], Loss: 1.8935, Perplexity: 6.6426
Epoch [3/3], Step [1466/3236], Loss: 1.9379, Perplexity: 6.9441
Epoch [3/3], Step [1467/3236], Loss: 1.9121, Perplexity: 6.7676
Epoch [3/3], Step [1468/3236], Loss: 1.8371, Perplexity: 6.2783
Epoch [3/3], Step [1469/3236], Loss: 2.0550, Perplexity: 7.8071
Epoch [3/3], Step [1470/3236], Loss: 2.0402, Perplexity: 7.6919
Epoch [3/3], Step [1471/3236], Loss: 1.8406, Perplexity: 6.3000
Epoch [3/3], Step [1472/3236], Loss: 1.9803, Perplexity: 7.2448
Epoch [3/3], Step [1473/3236], Loss: 1.8774, Perplexity: 6.5363
Epoch [3/3], Step [1474/3236], Loss: 2.0953, Perplexity: 8.1275
Epoch [3/3], Step [1475/3236], Loss: 2.0896, Perplexity: 8.0821
Epoch [3/3], Step [1476/3236], Loss: 1.9595, Perplexity: 7.0956
Epoch [3/3], Step [1477/3236], Loss: 2.2132, Perplexity: 9.1449
Epoch [3/3], Step [1478/3236], Loss: 1.8312, Perplexity: 6.2413
Epoch [3/3], Step [1479/3236], Loss: 1.9726, Perplexity: 7.1895
Epoch [3/3], Step [1480/3236], Loss: 1.9051, Perplexity: 6.7203
Epoch [3/3], Step [1481/3236], Loss: 1.9798, Perplexity: 7.2416
Epoch [3/3], Step [1482/3236], Loss: 2.2199, Perplexity: 9.2068
Epoch [3/3], Step [1483/3236], Loss: 2.0882, Perplexity: 8.0701
Epoch [3/3], Step [1484/3236], Loss: 2.2121, Perplexity: 9.1351
Epoch [3/3], Step [1485/3236], Loss: 1.8115, Perplexity: 6.1193
Epoch [3/3], Step [1486/3236], Loss: 2.0435, Perplexity: 7.7180
Epoch [3/3], Step [1487/3236], Loss: 1.9129, Perplexity: 6.7729
Epoch [3/3], Step [1488/3236], Loss: 2.0215, Perplexity: 7.5498
Epoch [3/3], Step [1489/3236], Loss: 1.9595, Perplexity: 7.0959
Epoch [3/3], Step [1490/3236], Loss: 2.1546, Perplexity: 8.6246
Epoch [3/3], Step [1491/3236], Loss: 1.8006, Perplexity: 6.0533
Epoch [3/3], Step [1492/3236], Loss: 1.8003, Perplexity: 6.0517
Epoch [3/3], Step [1493/3236], Loss: 1.9620, Perplexity: 7.1138
Epoch [3/3], Step [1494/3236], Loss: 1.9257, Perplexity: 6.8600
Epoch [3/3], Step [1495/3236], Loss: 1.9504, Perplexity: 7.0315
Epoch [3/3], Step [1496/3236], Loss: 2.0237, Perplexity: 7.5662
Epoch [3/3], Step [1497/3236], Loss: 1.8549, Perplexity: 6.3909
Epoch [3/3], Step [1498/3236], Loss: 1.8760, Perplexity: 6.5271
Epoch [3/3], Step [1499/3236], Loss: 1.8649, Perplexity: 6.4551
Epoch [3/3], Step [1500/3236], Loss: 1.8418, Perplexity: 6.3076
Epoch [3/3], Step [1501/3236], Loss: 1.9549, Perplexity: 7.0631
Epoch [3/3], Step [1502/3236], Loss: 1.9708, Perplexity: 7.1766
Epoch [3/3], Step [1503/3236], Loss: 2.0415, Perplexity: 7.7021
Epoch [3/3], Step [1504/3236], Loss: 2.2519, Perplexity: 9.5057
Epoch [3/3], Step [1505/3236], Loss: 2.5116, Perplexity: 12.3252
Epoch [3/3], Step [1506/3236], Loss: 1.9884, Perplexity: 7.3042
Epoch [3/3], Step [1507/3236], Loss: 1.9987, Perplexity: 7.3794
Epoch [3/3], Step [1508/3236], Loss: 2.0194, Perplexity: 7.5336
Epoch [3/3], Step [1509/3236], Loss: 2.0234, Perplexity: 7.5643
Epoch [3/3], Step [1510/3236], Loss: 1.8136, Perplexity: 6.1325
Epoch [3/3], Step [1511/3236], Loss: 2.0203, Perplexity: 7.5409
Epoch [3/3], Step [1512/3236], Loss: 2.1182, Perplexity: 8.3160
Epoch [3/3], Step [1513/3236], Loss: 2.5393, Perplexity: 12.6707
Epoch [3/3], Step [1514/3236], Loss: 2.0245, Perplexity: 7.5723
Epoch [3/3], Step [1515/3236], Loss: 1.9082, Perplexity: 6.7412
Epoch [3/3], Step [1516/3236], Loss: 1.9821, Perplexity: 7.2582
Epoch [3/3], Step [1517/3236], Loss: 1.8353, Perplexity: 6.2668
Epoch [3/3], Step [1518/3236], Loss: 1.9717, Perplexity: 7.1831
Epoch [3/3], Step [1519/3236], Loss: 1.9175, Perplexity: 6.8041
Epoch [3/3], Step [1520/3236], Loss: 2.0137, Perplexity: 7.4907
Epoch [3/3], Step [1521/3236], Loss: 2.2809, Perplexity: 9.7856
Epoch [3/3], Step [1522/3236], Loss: 1.8801, Perplexity: 6.5545
Epoch [3/3], Step [1523/3236], Loss: 2.9994, Perplexity: 20.0729
Epoch [3/3], Step [1524/3236], Loss: 1.8824, Perplexity: 6.5696
Epoch [3/3], Step [1525/3236], Loss: 2.1485, Perplexity: 8.5723
Epoch [3/3], Step [1526/3236], Loss: 2.0353, Perplexity: 7.6542
Epoch [3/3], Step [1527/3236], Loss: 2.0018, Perplexity: 7.4025
Epoch [3/3], Step [1528/3236], Loss: 1.9054, Perplexity: 6.7222
Epoch [3/3], Step [1529/3236], Loss: 1.8980, Perplexity: 6.6729
Epoch [3/3], Step [1530/3236], Loss: 1.9238, Perplexity: 6.8472
Epoch [3/3], Step [1531/3236], Loss: 1.9610, Perplexity: 7.1068
Epoch [3/3], Step [1532/3236], Loss: 1.9590, Perplexity: 7.0921
Epoch [3/3], Step [1533/3236], Loss: 1.8246, Perplexity: 6.2006
Epoch [3/3], Step [1534/3236], Loss: 1.7988, Perplexity: 6.0422
Epoch [3/3], Step [1535/3236], Loss: 2.2066, Perplexity: 9.0848
Epoch [3/3], Step [1536/3236], Loss: 1.8791, Perplexity: 6.5474
Epoch [3/3], Step [1537/3236], Loss: 2.1249, Perplexity: 8.3718
Epoch [3/3], Step [1538/3236], Loss: 1.9305, Perplexity: 6.8932
Epoch [3/3], Step [1539/3236], Loss: 1.9231, Perplexity: 6.8421
Epoch [3/3], Step [1540/3236], Loss: 1.9184, Perplexity: 6.8103
Epoch [3/3], Step [1541/3236], Loss: 1.9837, Perplexity: 7.2697
Epoch [3/3], Step [1542/3236], Loss: 2.2736, Perplexity: 9.7144
Epoch [3/3], Step [1543/3236], Loss: 1.8502, Perplexity: 6.3611
Epoch [3/3], Step [1544/3236], Loss: 1.8910, Perplexity: 6.6263
Epoch [3/3], Step [1545/3236], Loss: 2.0180, Perplexity: 7.5234
Epoch [3/3], Step [1546/3236], Loss: 1.9447, Perplexity: 6.9912
Epoch [3/3], Step [1547/3236], Loss: 2.0449, Perplexity: 7.7285
Epoch [3/3], Step [1548/3236], Loss: 2.1239, Perplexity: 8.3637
Epoch [3/3], Step [1549/3236], Loss: 1.9702, Perplexity: 7.1721
Epoch [3/3], Step [1550/3236], Loss: 1.9442, Perplexity: 6.9877
Epoch [3/3], Step [1551/3236], Loss: 1.9265, Perplexity: 6.8654
Epoch [3/3], Step [1552/3236], Loss: 2.0444, Perplexity: 7.7245
Epoch [3/3], Step [1553/3236], Loss: 2.2119, Perplexity: 9.1328
Epoch [3/3], Step [1554/3236], Loss: 1.9909, Perplexity: 7.3225
Epoch [3/3], Step [1555/3236], Loss: 1.8106, Perplexity: 6.1144
Epoch [3/3], Step [1556/3236], Loss: 1.9975, Perplexity: 7.3708
Epoch [3/3], Step [1557/3236], Loss: 2.0569, Perplexity: 7.8214
Epoch [3/3], Step [1558/3236], Loss: 1.9743, Perplexity: 7.2017
Epoch [3/3], Step [1559/3236], Loss: 1.8843, Perplexity: 6.5816
Epoch [3/3], Step [1560/3236], Loss: 2.0386, Perplexity: 7.6799
Epoch [3/3], Step [1561/3236], Loss: 1.8507, Perplexity: 6.3643
Epoch [3/3], Step [1562/3236], Loss: 1.9947, Perplexity: 7.3499
Epoch [3/3], Step [1563/3236], Loss: 1.9110, Perplexity: 6.7597
Epoch [3/3], Step [1564/3236], Loss: 1.9281, Perplexity: 6.8768
Epoch [3/3], Step [1565/3236], Loss: 1.8909, Perplexity: 6.6251
Epoch [3/3], Step [1566/3236], Loss: 1.8585, Perplexity: 6.4140
Epoch [3/3], Step [1567/3236], Loss: 1.9612, Perplexity: 7.1082
Epoch [3/3], Step [1568/3236], Loss: 2.1375, Perplexity: 8.4786
Epoch [3/3], Step [1569/3236], Loss: 2.0010, Perplexity: 7.3967
Epoch [3/3], Step [1570/3236], Loss: 1.9039, Perplexity: 6.7119
Epoch [3/3], Step [1571/3236], Loss: 1.8821, Perplexity: 6.5670
Epoch [3/3], Step [1572/3236], Loss: 1.9974, Perplexity: 7.3701
Epoch [3/3], Step [1573/3236], Loss: 1.8465, Perplexity: 6.3374
Epoch [3/3], Step [1574/3236], Loss: 1.9630, Perplexity: 7.1208
Epoch [3/3], Step [1575/3236], Loss: 2.4272, Perplexity: 11.3276
Epoch [3/3], Step [1576/3236], Loss: 2.0026, Perplexity: 7.4085
Epoch [3/3], Step [1577/3236], Loss: 1.8044, Perplexity: 6.0764
Epoch [3/3], Step [1578/3236], Loss: 2.2458, Perplexity: 9.4478
Epoch [3/3], Step [1579/3236], Loss: 2.8437, Perplexity: 17.1786
Epoch [3/3], Step [1580/3236], Loss: 2.4003, Perplexity: 11.0266
Epoch [3/3], Step [1581/3236], Loss: 1.9228, Perplexity: 6.8402
Epoch [3/3], Step [1582/3236], Loss: 1.8974, Perplexity: 6.6686
Epoch [3/3], Step [1583/3236], Loss: 2.0108, Perplexity: 7.4693
Epoch [3/3], Step [1584/3236], Loss: 1.8555, Perplexity: 6.3947
Epoch [3/3], Step [1585/3236], Loss: 1.8521, Perplexity: 6.3730
Epoch [3/3], Step [1586/3236], Loss: 1.9532, Perplexity: 7.0512
Epoch [3/3], Step [1587/3236], Loss: 1.8972, Perplexity: 6.6671
Epoch [3/3], Step [1588/3236], Loss: 2.0713, Perplexity: 7.9352
Epoch [3/3], Step [1589/3236], Loss: 2.1810, Perplexity: 8.8549
Epoch [3/3], Step [1590/3236], Loss: 1.9230, Perplexity: 6.8415
Epoch [3/3], Step [1591/3236], Loss: 2.0191, Perplexity: 7.5313
Epoch [3/3], Step [1592/3236], Loss: 1.9943, Perplexity: 7.3468
Epoch [3/3], Step [1593/3236], Loss: 1.8357, Perplexity: 6.2696
Epoch [3/3], Step [1594/3236], Loss: 1.8676, Perplexity: 6.4728
Epoch [3/3], Step [1595/3236], Loss: 1.8214, Perplexity: 6.1805
Epoch [3/3], Step [1596/3236], Loss: 1.9515, Perplexity: 7.0392
Epoch [3/3], Step [1597/3236], Loss: 1.9435, Perplexity: 6.9829
Epoch [3/3], Step [1598/3236], Loss: 2.7456, Perplexity: 15.5745
Epoch [3/3], Step [1599/3236], Loss: 2.1958, Perplexity: 8.9868
Epoch [3/3], Step [1600/3236], Loss: 1.9518, Perplexity: 7.0413
Epoch [3/3], Step [1601/3236], Loss: 2.3717, Perplexity: 10.7161
Epoch [3/3], Step [1602/3236], Loss: 1.8772, Perplexity: 6.5353
Epoch [3/3], Step [1603/3236], Loss: 2.0034, Perplexity: 7.4140
Epoch [3/3], Step [1604/3236], Loss: 1.8591, Perplexity: 6.4182
Epoch [3/3], Step [1605/3236], Loss: 1.8906, Perplexity: 6.6235
Epoch [3/3], Step [1606/3236], Loss: 2.0255, Perplexity: 7.5796
Epoch [3/3], Step [1607/3236], Loss: 2.2523, Perplexity: 9.5094
Epoch [3/3], Step [1608/3236], Loss: 1.8789, Perplexity: 6.5462
Epoch [3/3], Step [1609/3236], Loss: 1.9505, Perplexity: 7.0319
Epoch [3/3], Step [1610/3236], Loss: 2.0326, Perplexity: 7.6342
Epoch [3/3], Step [1611/3236], Loss: 1.8803, Perplexity: 6.5557
Epoch [3/3], Step [1612/3236], Loss: 1.9408, Perplexity: 6.9643
Epoch [3/3], Step [1613/3236], Loss: 2.5399, Perplexity: 12.6788
Epoch [3/3], Step [1614/3236], Loss: 1.9158, Perplexity: 6.7924
Epoch [3/3], Step [1615/3236], Loss: 2.0755, Perplexity: 7.9683
Epoch [3/3], Step [1616/3236], Loss: 2.2367, Perplexity: 9.3622
Epoch [3/3], Step [1617/3236], Loss: 2.0760, Perplexity: 7.9723
Epoch [3/3], Step [1618/3236], Loss: 1.8317, Perplexity: 6.2442
Epoch [3/3], Step [1619/3236], Loss: 2.1370, Perplexity: 8.4742
Epoch [3/3], Step [1620/3236], Loss: 1.9271, Perplexity: 6.8697
Epoch [3/3], Step [1621/3236], Loss: 1.8619, Perplexity: 6.4361
Epoch [3/3], Step [1622/3236], Loss: 1.9355, Perplexity: 6.9276
Epoch [3/3], Step [1623/3236], Loss: 1.8774, Perplexity: 6.5363
Epoch [3/3], Step [1624/3236], Loss: 1.9150, Perplexity: 6.7868
Epoch [3/3], Step [1625/3236], Loss: 1.9975, Perplexity: 7.3709
Epoch [3/3], Step [1626/3236], Loss: 1.9814, Perplexity: 7.2530
Epoch [3/3], Step [1627/3236], Loss: 2.3539, Perplexity: 10.5270
Epoch [3/3], Step [1628/3236], Loss: 1.7601, Perplexity: 5.8129
Epoch [3/3], Step [1629/3236], Loss: 2.2285, Perplexity: 9.2863
Epoch [3/3], Step [1630/3236], Loss: 2.0309, Perplexity: 7.6212
Epoch [3/3], Step [1631/3236], Loss: 2.8768, Perplexity: 17.7573
Epoch [3/3], Step [1632/3236], Loss: 2.1056, Perplexity: 8.2119
Epoch [3/3], Step [1633/3236], Loss: 1.9511, Perplexity: 7.0364
Epoch [3/3], Step [1634/3236], Loss: 2.1881, Perplexity: 8.9184
Epoch [3/3], Step [1635/3236], Loss: 1.9282, Perplexity: 6.8771
Epoch [3/3], Step [1636/3236], Loss: 1.9433, Perplexity: 6.9816
Epoch [3/3], Step [1637/3236], Loss: 1.9777, Perplexity: 7.2264
Epoch [3/3], Step [1638/3236], Loss: 1.8601, Perplexity: 6.4242
Epoch [3/3], Step [1639/3236], Loss: 1.9241, Perplexity: 6.8488
Epoch [3/3], Step [1640/3236], Loss: 2.0555, Perplexity: 7.8107
Epoch [3/3], Step [1641/3236], Loss: 1.9920, Perplexity: 7.3304
Epoch [3/3], Step [1642/3236], Loss: 2.0368, Perplexity: 7.6662
Epoch [3/3], Step [1643/3236], Loss: 1.8781, Perplexity: 6.5410
Epoch [3/3], Step [1644/3236], Loss: 2.2211, Perplexity: 9.2179
Epoch [3/3], Step [1645/3236], Loss: 1.9502, Perplexity: 7.0298
Epoch [3/3], Step [1646/3236], Loss: 2.1769, Perplexity: 8.8190
Epoch [3/3], Step [1647/3236], Loss: 2.1159, Perplexity: 8.2971
Epoch [3/3], Step [1648/3236], Loss: 1.9647, Perplexity: 7.1329
Epoch [3/3], Step [1649/3236], Loss: 1.9840, Perplexity: 7.2716
Epoch [3/3], Step [1650/3236], Loss: 1.8441, Perplexity: 6.3227
Epoch [3/3], Step [1651/3236], Loss: 2.1605, Perplexity: 8.6751
Epoch [3/3], Step [1652/3236], Loss: 2.1282, Perplexity: 8.3995
Epoch [3/3], Step [1653/3236], Loss: 2.0346, Perplexity: 7.6489
Epoch [3/3], Step [1654/3236], Loss: 2.0411, Perplexity: 7.6989
Epoch [3/3], Step [1655/3236], Loss: 1.8327, Perplexity: 6.2506
Epoch [3/3], Step [1656/3236], Loss: 2.0677, Perplexity: 7.9064
Epoch [3/3], Step [1657/3236], Loss: 1.9128, Perplexity: 6.7723
Epoch [3/3], Step [1658/3236], Loss: 1.9552, Perplexity: 7.0652
Epoch [3/3], Step [1659/3236], Loss: 2.2153, Perplexity: 9.1639
Epoch [3/3], Step [1660/3236], Loss: 2.0369, Perplexity: 7.6671
Epoch [3/3], Step [1661/3236], Loss: 1.8313, Perplexity: 6.2419
Epoch [3/3], Step [1662/3236], Loss: 1.9200, Perplexity: 6.8212
Epoch [3/3], Step [1663/3236], Loss: 2.0007, Perplexity: 7.3940
Epoch [3/3], Step [1664/3236], Loss: 1.8583, Perplexity: 6.4125
Epoch [3/3], Step [1665/3236], Loss: 1.9669, Perplexity: 7.1487
Epoch [3/3], Step [1666/3236], Loss: 2.4242, Perplexity: 11.2930
Epoch [3/3], Step [1667/3236], Loss: 1.8237, Perplexity: 6.1945
Epoch [3/3], Step [1668/3236], Loss: 1.9114, Perplexity: 6.7622
Epoch [3/3], Step [1669/3236], Loss: 1.8510, Perplexity: 6.3659
Epoch [3/3], Step [1670/3236], Loss: 2.0395, Perplexity: 7.6867
Epoch [3/3], Step [1671/3236], Loss: 2.2400, Perplexity: 9.3935
Epoch [3/3], Step [1672/3236], Loss: 2.1796, Perplexity: 8.8428
Epoch [3/3], Step [1673/3236], Loss: 1.9207, Perplexity: 6.8261
Epoch [3/3], Step [1674/3236], Loss: 1.9376, Perplexity: 6.9421
Epoch [3/3], Step [1675/3236], Loss: 1.9252, Perplexity: 6.8564
Epoch [3/3], Step [1676/3236], Loss: 1.8896, Perplexity: 6.6166
Epoch [3/3], Step [1677/3236], Loss: 1.8366, Perplexity: 6.2755
Epoch [3/3], Step [1678/3236], Loss: 2.2213, Perplexity: 9.2197
Epoch [3/3], Step [1679/3236], Loss: 3.6173, Perplexity: 37.2362
Epoch [3/3], Step [1680/3236], Loss: 1.8962, Perplexity: 6.6609
Epoch [3/3], Step [1681/3236], Loss: 1.9311, Perplexity: 6.8973
Epoch [3/3], Step [1682/3236], Loss: 1.9383, Perplexity: 6.9468
Epoch [3/3], Step [1683/3236], Loss: 2.2329, Perplexity: 9.3270
Epoch [3/3], Step [1684/3236], Loss: 2.0118, Perplexity: 7.4771
Epoch [3/3], Step [1685/3236], Loss: 2.1077, Perplexity: 8.2295
Epoch [3/3], Step [1686/3236], Loss: 1.9184, Perplexity: 6.8099
Epoch [3/3], Step [1687/3236], Loss: 1.8711, Perplexity: 6.4953
Epoch [3/3], Step [1688/3236], Loss: 2.4197, Perplexity: 11.2423
Epoch [3/3], Step [1689/3236], Loss: 2.8253, Perplexity: 16.8662
Epoch [3/3], Step [1690/3236], Loss: 2.0673, Perplexity: 7.9032
Epoch [3/3], Step [1691/3236], Loss: 1.9407, Perplexity: 6.9635
Epoch [3/3], Step [1692/3236], Loss: 2.2134, Perplexity: 9.1471
Epoch [3/3], Step [1693/3236], Loss: 2.0441, Perplexity: 7.7221
Epoch [3/3], Step [1694/3236], Loss: 2.0750, Perplexity: 7.9645
Epoch [3/3], Step [1695/3236], Loss: 1.9045, Perplexity: 6.7161
Epoch [3/3], Step [1696/3236], Loss: 2.0634, Perplexity: 7.8725
Epoch [3/3], Step [1697/3236], Loss: 2.0208, Perplexity: 7.5444
Epoch [3/3], Step [1698/3236], Loss: 2.0359, Perplexity: 7.6591
Epoch [3/3], Step [1699/3236], Loss: 1.9960, Perplexity: 7.3592
Epoch [3/3], Step [1700/3236], Loss: 1.8316, Perplexity: 6.2441
Epoch [3/3], Step [1701/3236], Loss: 1.9414, Perplexity: 6.9687
Epoch [3/3], Step [1702/3236], Loss: 1.9154, Perplexity: 6.7899
Epoch [3/3], Step [1703/3236], Loss: 1.9381, Perplexity: 6.9457
Epoch [3/3], Step [1704/3236], Loss: 1.8913, Perplexity: 6.6280
Epoch [3/3], Step [1705/3236], Loss: 1.9568, Perplexity: 7.0764
Epoch [3/3], Step [1706/3236], Loss: 2.0585, Perplexity: 7.8342
Epoch [3/3], Step [1707/3236], Loss: 1.9414, Perplexity: 6.9685
Epoch [3/3], Step [1708/3236], Loss: 2.4105, Perplexity: 11.1398
Epoch [3/3], Step [1709/3236], Loss: 2.2645, Perplexity: 9.6259
Epoch [3/3], Step [1710/3236], Loss: 1.8854, Perplexity: 6.5889
Epoch [3/3], Step [1711/3236], Loss: 2.2023, Perplexity: 9.0462
Epoch [3/3], Step [1712/3236], Loss: 1.9748, Perplexity: 7.2049
Epoch [3/3], Step [1713/3236], Loss: 2.2899, Perplexity: 9.8741
Epoch [3/3], Step [1714/3236], Loss: 1.9883, Perplexity: 7.3030
Epoch [3/3], Step [1715/3236], Loss: 2.0302, Perplexity: 7.6159
Epoch [3/3], Step [1716/3236], Loss: 2.1042, Perplexity: 8.2009
Epoch [3/3], Step [1717/3236], Loss: 1.9767, Perplexity: 7.2188
Epoch [3/3], Step [1718/3236], Loss: 1.9189, Perplexity: 6.8135
Epoch [3/3], Step [1719/3236], Loss: 2.0036, Perplexity: 7.4159
Epoch [3/3], Step [1720/3236], Loss: 1.9410, Perplexity: 6.9657
Epoch [3/3], Step [1721/3236], Loss: 2.0822, Perplexity: 8.0218
Epoch [3/3], Step [1722/3236], Loss: 2.0208, Perplexity: 7.5447
Epoch [3/3], Step [1723/3236], Loss: 1.8513, Perplexity: 6.3683
Epoch [3/3], Step [1724/3236], Loss: 2.0090, Perplexity: 7.4556
Epoch [3/3], Step [1725/3236], Loss: 1.8313, Perplexity: 6.2422
Epoch [3/3], Step [1726/3236], Loss: 1.8274, Perplexity: 6.2178
Epoch [3/3], Step [1727/3236], Loss: 1.9122, Perplexity: 6.7682
Epoch [3/3], Step [1728/3236], Loss: 1.8951, Perplexity: 6.6530
Epoch [3/3], Step [1729/3236], Loss: 2.1463, Perplexity: 8.5529
Epoch [3/3], Step [1730/3236], Loss: 2.0190, Perplexity: 7.5310
Epoch [3/3], Step [1731/3236], Loss: 1.8840, Perplexity: 6.5799
Epoch [3/3], Step [1732/3236], Loss: 1.9655, Perplexity: 7.1383
Epoch [3/3], Step [1733/3236], Loss: 1.9381, Perplexity: 6.9453
Epoch [3/3], Step [1734/3236], Loss: 2.0076, Perplexity: 7.4458
Epoch [3/3], Step [1735/3236], Loss: 1.8910, Perplexity: 6.6261
Epoch [3/3], Step [1736/3236], Loss: 1.8008, Perplexity: 6.0548
Epoch [3/3], Step [1737/3236], Loss: 1.8808, Perplexity: 6.5586
Epoch [3/3], Step [1738/3236], Loss: 2.1120, Perplexity: 8.2650
Epoch [3/3], Step [1739/3236], Loss: 2.0812, Perplexity: 8.0140
Epoch [3/3], Step [1740/3236], Loss: 2.0362, Perplexity: 7.6617
Epoch [3/3], Step [1741/3236], Loss: 2.1510, Perplexity: 8.5931
Epoch [3/3], Step [1742/3236], Loss: 1.8894, Perplexity: 6.6156
Epoch [3/3], Step [1743/3236], Loss: 2.2895, Perplexity: 9.8704
Epoch [3/3], Step [1744/3236], Loss: 1.8948, Perplexity: 6.6509
Epoch [3/3], Step [1745/3236], Loss: 1.9309, Perplexity: 6.8958
Epoch [3/3], Step [1746/3236], Loss: 1.7833, Perplexity: 5.9496
Epoch [3/3], Step [1747/3236], Loss: 1.8726, Perplexity: 6.5051
Epoch [3/3], Step [1748/3236], Loss: 1.9186, Perplexity: 6.8113
Epoch [3/3], Step [1749/3236], Loss: 1.9289, Perplexity: 6.8820
Epoch [3/3], Step [1750/3236], Loss: 1.8897, Perplexity: 6.6176
Epoch [3/3], Step [1751/3236], Loss: 1.9369, Perplexity: 6.9374
Epoch [3/3], Step [1752/3236], Loss: 1.8805, Perplexity: 6.5568
Epoch [3/3], Step [1753/3236], Loss: 1.9228, Perplexity: 6.8401
Epoch [3/3], Step [1754/3236], Loss: 2.0492, Perplexity: 7.7618
Epoch [3/3], Step [1755/3236], Loss: 1.9814, Perplexity: 7.2526
Epoch [3/3], Step [1756/3236], Loss: 1.9512, Perplexity: 7.0374
Epoch [3/3], Step [1757/3236], Loss: 2.2468, Perplexity: 9.4576
Epoch [3/3], Step [1758/3236], Loss: 2.0173, Perplexity: 7.5182
Epoch [3/3], Step [1759/3236], Loss: 1.9244, Perplexity: 6.8508
Epoch [3/3], Step [1760/3236], Loss: 1.9717, Perplexity: 7.1832
Epoch [3/3], Step [1761/3236], Loss: 1.8277, Perplexity: 6.2197
Epoch [3/3], Step [1762/3236], Loss: 1.8820, Perplexity: 6.5665
Epoch [3/3], Step [1763/3236], Loss: 2.2030, Perplexity: 9.0523
Epoch [3/3], Step [1764/3236], Loss: 2.6540, Perplexity: 14.2105
Epoch [3/3], Step [1765/3236], Loss: 2.1030, Perplexity: 8.1903
Epoch [3/3], Step [1766/3236], Loss: 2.1861, Perplexity: 8.9001
Epoch [3/3], Step [1767/3236], Loss: 1.9447, Perplexity: 6.9915
Epoch [3/3], Step [1768/3236], Loss: 1.9387, Perplexity: 6.9496
Epoch [3/3], Step [1769/3236], Loss: 2.0082, Perplexity: 7.4500
Epoch [3/3], Step [1770/3236], Loss: 2.0118, Perplexity: 7.4769
Epoch [3/3], Step [1771/3236], Loss: 1.8311, Perplexity: 6.2407
Epoch [3/3], Step [1772/3236], Loss: 2.0718, Perplexity: 7.9392
Epoch [3/3], Step [1773/3236], Loss: 1.8232, Perplexity: 6.1914
Epoch [3/3], Step [1774/3236], Loss: 1.9703, Perplexity: 7.1727
Epoch [3/3], Step [1775/3236], Loss: 2.0048, Perplexity: 7.4246
Epoch [3/3], Step [1776/3236], Loss: 1.9932, Perplexity: 7.3389
Epoch [3/3], Step [1777/3236], Loss: 1.8417, Perplexity: 6.3071
Epoch [3/3], Step [1778/3236], Loss: 1.8062, Perplexity: 6.0871
Epoch [3/3], Step [1779/3236], Loss: 2.2303, Perplexity: 9.3027
Epoch [3/3], Step [1780/3236], Loss: 1.9839, Perplexity: 7.2709
Epoch [3/3], Step [1781/3236], Loss: 2.3505, Perplexity: 10.4910
Epoch [3/3], Step [1782/3236], Loss: 1.9480, Perplexity: 7.0149
Epoch [3/3], Step [1783/3236], Loss: 1.9824, Perplexity: 7.2601
Epoch [3/3], Step [1784/3236], Loss: 2.0488, Perplexity: 7.7586
Epoch [3/3], Step [1785/3236], Loss: 1.9972, Perplexity: 7.3680
Epoch [3/3], Step [1786/3236], Loss: 1.9037, Perplexity: 6.7104
Epoch [3/3], Step [1787/3236], Loss: 2.0318, Perplexity: 7.6279
Epoch [3/3], Step [1788/3236], Loss: 2.5555, Perplexity: 12.8777
Epoch [3/3], Step [1789/3236], Loss: 1.9292, Perplexity: 6.8841
Epoch [3/3], Step [1790/3236], Loss: 2.1842, Perplexity: 8.8837
Epoch [3/3], Step [1791/3236], Loss: 1.8747, Perplexity: 6.5189
Epoch [3/3], Step [1792/3236], Loss: 1.9895, Perplexity: 7.3117
Epoch [3/3], Step [1793/3236], Loss: 2.1418, Perplexity: 8.5145
Epoch [3/3], Step [1794/3236], Loss: 2.0862, Perplexity: 8.0546
Epoch [3/3], Step [1795/3236], Loss: 2.0597, Perplexity: 7.8437
Epoch [3/3], Step [1796/3236], Loss: 2.0622, Perplexity: 7.8633
Epoch [3/3], Step [1797/3236], Loss: 2.1473, Perplexity: 8.5620
Epoch [3/3], Step [1798/3236], Loss: 2.0163, Perplexity: 7.5108
Epoch [3/3], Step [1799/3236], Loss: 2.0900, Perplexity: 8.0846
Epoch [3/3], Step [1800/3236], Loss: 2.0732, Perplexity: 7.9506
Epoch [3/3], Step [1801/3236], Loss: 1.8032, Perplexity: 6.0692
Epoch [3/3], Step [1802/3236], Loss: 1.8561, Perplexity: 6.3989
Epoch [3/3], Step [1803/3236], Loss: 2.1531, Perplexity: 8.6116
Epoch [3/3], Step [1804/3236], Loss: 2.0094, Perplexity: 7.4585
Epoch [3/3], Step [1805/3236], Loss: 1.9008, Perplexity: 6.6916
Epoch [3/3], Step [1806/3236], Loss: 1.9025, Perplexity: 6.7030
Epoch [3/3], Step [1807/3236], Loss: 1.9874, Perplexity: 7.2967
Epoch [3/3], Step [1808/3236], Loss: 1.8333, Perplexity: 6.2547
Epoch [3/3], Step [1809/3236], Loss: 1.8652, Perplexity: 6.4570
Epoch [3/3], Step [1810/3236], Loss: 1.9180, Perplexity: 6.8076
Epoch [3/3], Step [1811/3236], Loss: 2.1277, Perplexity: 8.3956
Epoch [3/3], Step [1812/3236], Loss: 1.9758, Perplexity: 7.2122
Epoch [3/3], Step [1813/3236], Loss: 1.8879, Perplexity: 6.6052
Epoch [3/3], Step [1814/3236], Loss: 2.0098, Perplexity: 7.4621
Epoch [3/3], Step [1815/3236], Loss: 1.8926, Perplexity: 6.6367
Epoch [3/3], Step [1816/3236], Loss: 1.8629, Perplexity: 6.4427
Epoch [3/3], Step [1817/3236], Loss: 1.9556, Perplexity: 7.0680
Epoch [3/3], Step [1818/3236], Loss: 1.8880, Perplexity: 6.6061
Epoch [3/3], Step [1819/3236], Loss: 1.9429, Perplexity: 6.9792
Epoch [3/3], Step [1820/3236], Loss: 2.0749, Perplexity: 7.9639
Epoch [3/3], Step [1821/3236], Loss: 1.9575, Perplexity: 7.0813
Epoch [3/3], Step [1822/3236], Loss: 2.0768, Perplexity: 7.9793
Epoch [3/3], Step [1823/3236], Loss: 1.7842, Perplexity: 5.9545
Epoch [3/3], Step [1824/3236], Loss: 1.8604, Perplexity: 6.4263
Epoch [3/3], Step [1825/3236], Loss: 1.9839, Perplexity: 7.2711
Epoch [3/3], Step [1826/3236], Loss: 1.8411, Perplexity: 6.3038
Epoch [3/3], Step [1827/3236], Loss: 2.1403, Perplexity: 8.5019
Epoch [3/3], Step [1828/3236], Loss: 1.9365, Perplexity: 6.9342
Epoch [3/3], Step [1829/3236], Loss: 1.8466, Perplexity: 6.3382
Epoch [3/3], Step [1830/3236], Loss: 1.8519, Perplexity: 6.3721
Epoch [3/3], Step [1831/3236], Loss: 1.8535, Perplexity: 6.3824
Epoch [3/3], Step [1832/3236], Loss: 2.0777, Perplexity: 7.9862
Epoch [3/3], Step [1833/3236], Loss: 1.9744, Perplexity: 7.2026
Epoch [3/3], Step [1834/3236], Loss: 2.0820, Perplexity: 8.0209
Epoch [3/3], Step [1835/3236], Loss: 1.9083, Perplexity: 6.7413
Epoch [3/3], Step [1836/3236], Loss: 2.0029, Perplexity: 7.4108
Epoch [3/3], Step [1837/3236], Loss: 1.8839, Perplexity: 6.5791
Epoch [3/3], Step [1838/3236], Loss: 1.9239, Perplexity: 6.8473
Epoch [3/3], Step [1839/3236], Loss: 1.9937, Perplexity: 7.3425
Epoch [3/3], Step [1840/3236], Loss: 1.9730, Perplexity: 7.1921
Epoch [3/3], Step [1841/3236], Loss: 1.8796, Perplexity: 6.5510
Epoch [3/3], Step [1842/3236], Loss: 2.3065, Perplexity: 10.0391
Epoch [3/3], Step [1843/3236], Loss: 2.4877, Perplexity: 12.0337
Epoch [3/3], Step [1844/3236], Loss: 2.1977, Perplexity: 9.0046
Epoch [3/3], Step [1845/3236], Loss: 1.8798, Perplexity: 6.5525
Epoch [3/3], Step [1846/3236], Loss: 1.8413, Perplexity: 6.3050
Epoch [3/3], Step [1847/3236], Loss: 1.9372, Perplexity: 6.9393
Epoch [3/3], Step [1848/3236], Loss: 1.9801, Perplexity: 7.2432
Epoch [3/3], Step [1849/3236], Loss: 1.7890, Perplexity: 5.9835
Epoch [3/3], Step [1850/3236], Loss: 2.0980, Perplexity: 8.1497
Epoch [3/3], Step [1851/3236], Loss: 1.8733, Perplexity: 6.5100
Epoch [3/3], Step [1852/3236], Loss: 1.9613, Perplexity: 7.1086
Epoch [3/3], Step [1853/3236], Loss: 1.9095, Perplexity: 6.7499
Epoch [3/3], Step [1854/3236], Loss: 1.9440, Perplexity: 6.9867
Epoch [3/3], Step [1855/3236], Loss: 1.9667, Perplexity: 7.1468
Epoch [3/3], Step [1856/3236], Loss: 1.9353, Perplexity: 6.9261
Epoch [3/3], Step [1857/3236], Loss: 1.9279, Perplexity: 6.8752
Epoch [3/3], Step [1858/3236], Loss: 1.9145, Perplexity: 6.7833
Epoch [3/3], Step [1859/3236], Loss: 2.0106, Perplexity: 7.4676
Epoch [3/3], Step [1860/3236], Loss: 2.0644, Perplexity: 7.8802
Epoch [3/3], Step [1861/3236], Loss: 1.9735, Perplexity: 7.1957
Epoch [3/3], Step [1862/3236], Loss: 1.9116, Perplexity: 6.7640
Epoch [3/3], Step [1863/3236], Loss: 2.3251, Perplexity: 10.2278
Epoch [3/3], Step [1864/3236], Loss: 1.8822, Perplexity: 6.5681
Epoch [3/3], Step [1865/3236], Loss: 2.0856, Perplexity: 8.0493
Epoch [3/3], Step [1866/3236], Loss: 1.8683, Perplexity: 6.4771
Epoch [3/3], Step [1867/3236], Loss: 1.9381, Perplexity: 6.9453
Epoch [3/3], Step [1868/3236], Loss: 1.9218, Perplexity: 6.8335
Epoch [3/3], Step [1869/3236], Loss: 1.8763, Perplexity: 6.5293
Epoch [3/3], Step [1870/3236], Loss: 2.0576, Perplexity: 7.8269
Epoch [3/3], Step [1871/3236], Loss: 1.9898, Perplexity: 7.3140
Epoch [3/3], Step [1872/3236], Loss: 1.8837, Perplexity: 6.5779
Epoch [3/3], Step [1873/3236], Loss: 2.0367, Perplexity: 7.6656
Epoch [3/3], Step [1874/3236], Loss: 1.8073, Perplexity: 6.0938
Epoch [3/3], Step [1875/3236], Loss: 2.0292, Perplexity: 7.6078
Epoch [3/3], Step [1876/3236], Loss: 1.9041, Perplexity: 6.7134
Epoch [3/3], Step [1877/3236], Loss: 1.9625, Perplexity: 7.1172
Epoch [3/3], Step [1878/3236], Loss: 1.8523, Perplexity: 6.3747
Epoch [3/3], Step [1879/3236], Loss: 1.9749, Perplexity: 7.2062
Epoch [3/3], Step [1880/3236], Loss: 1.9215, Perplexity: 6.8311
Epoch [3/3], Step [1881/3236], Loss: 1.8446, Perplexity: 6.3258
Epoch [3/3], Step [1882/3236], Loss: 2.1277, Perplexity: 8.3955
Epoch [3/3], Step [1883/3236], Loss: 1.8187, Perplexity: 6.1637
Epoch [3/3], Step [1884/3236], Loss: 2.1855, Perplexity: 8.8947
Epoch [3/3], Step [1885/3236], Loss: 1.9755, Perplexity: 7.2101
Epoch [3/3], Step [1886/3236], Loss: 2.3116, Perplexity: 10.0904
Epoch [3/3], Step [1887/3236], Loss: 1.8979, Perplexity: 6.6718
Epoch [3/3], Step [1888/3236], Loss: 2.0773, Perplexity: 7.9833
Epoch [3/3], Step [1889/3236], Loss: 1.8932, Perplexity: 6.6405
Epoch [3/3], Step [1890/3236], Loss: 2.1396, Perplexity: 8.4961
Epoch [3/3], Step [1891/3236], Loss: 1.9490, Perplexity: 7.0218
Epoch [3/3], Step [1892/3236], Loss: 1.9084, Perplexity: 6.7422
Epoch [3/3], Step [1893/3236], Loss: 1.7934, Perplexity: 6.0098
Epoch [3/3], Step [1894/3236], Loss: 2.1430, Perplexity: 8.5250
Epoch [3/3], Step [1895/3236], Loss: 1.7474, Perplexity: 5.7399
Epoch [3/3], Step [1896/3236], Loss: 2.0738, Perplexity: 7.9550
Epoch [3/3], Step [1897/3236], Loss: 2.2723, Perplexity: 9.7020
Epoch [3/3], Step [1898/3236], Loss: 2.9706, Perplexity: 19.5038
Epoch [3/3], Step [1899/3236], Loss: 1.9914, Perplexity: 7.3256
Epoch [3/3], Step [1900/3236], Loss: 1.9904, Perplexity: 7.3187
Epoch [3/3], Step [1901/3236], Loss: 2.0119, Perplexity: 7.4772
Epoch [3/3], Step [1902/3236], Loss: 2.3278, Perplexity: 10.2554
Epoch [3/3], Step [1903/3236], Loss: 2.0028, Perplexity: 7.4100
Epoch [3/3], Step [1904/3236], Loss: 2.2067, Perplexity: 9.0853
Epoch [3/3], Step [1905/3236], Loss: 2.0679, Perplexity: 7.9083
Epoch [3/3], Step [1906/3236], Loss: 2.3497, Perplexity: 10.4819
Epoch [3/3], Step [1907/3236], Loss: 2.1875, Perplexity: 8.9133
Epoch [3/3], Step [1908/3236], Loss: 1.8837, Perplexity: 6.5780
Epoch [3/3], Step [1909/3236], Loss: 2.1449, Perplexity: 8.5410
Epoch [3/3], Step [1910/3236], Loss: 2.1126, Perplexity: 8.2694
Epoch [3/3], Step [1911/3236], Loss: 1.9751, Perplexity: 7.2072
Epoch [3/3], Step [1912/3236], Loss: 2.0649, Perplexity: 7.8849
Epoch [3/3], Step [1913/3236], Loss: 1.8378, Perplexity: 6.2830
Epoch [3/3], Step [1914/3236], Loss: 1.8998, Perplexity: 6.6847
Epoch [3/3], Step [1915/3236], Loss: 1.9091, Perplexity: 6.7471
Epoch [3/3], Step [1916/3236], Loss: 1.8196, Perplexity: 6.1691
Epoch [3/3], Step [1917/3236], Loss: 1.9356, Perplexity: 6.9285
Epoch [3/3], Step [1918/3236], Loss: 2.0650, Perplexity: 7.8851
Epoch [3/3], Step [1919/3236], Loss: 1.9216, Perplexity: 6.8317
Epoch [3/3], Step [1920/3236], Loss: 2.0540, Perplexity: 7.7987
Epoch [3/3], Step [1921/3236], Loss: 2.1588, Perplexity: 8.6604
Epoch [3/3], Step [1922/3236], Loss: 1.8829, Perplexity: 6.5723
Epoch [3/3], Step [1923/3236], Loss: 1.9333, Perplexity: 6.9121
Epoch [3/3], Step [1924/3236], Loss: 2.0467, Perplexity: 7.7426
Epoch [3/3], Step [1925/3236], Loss: 1.9264, Perplexity: 6.8651
Epoch [3/3], Step [1926/3236], Loss: 2.4145, Perplexity: 11.1845
Epoch [3/3], Step [1927/3236], Loss: 1.9496, Perplexity: 7.0261
Epoch [3/3], Step [1928/3236], Loss: 1.9826, Perplexity: 7.2614
Epoch [3/3], Step [1929/3236], Loss: 2.0091, Perplexity: 7.4568
Epoch [3/3], Step [1930/3236], Loss: 1.8658, Perplexity: 6.4610
Epoch [3/3], Step [1931/3236], Loss: 1.9117, Perplexity: 6.7643
Epoch [3/3], Step [1932/3236], Loss: 1.9926, Perplexity: 7.3346
Epoch [3/3], Step [1933/3236], Loss: 2.0117, Perplexity: 7.4760
Epoch [3/3], Step [1934/3236], Loss: 1.8356, Perplexity: 6.2690
Epoch [3/3], Step [1935/3236], Loss: 1.8819, Perplexity: 6.5660
Epoch [3/3], Step [1936/3236], Loss: 1.9440, Perplexity: 6.9864
Epoch [3/3], Step [1937/3236], Loss: 1.8986, Perplexity: 6.6767
Epoch [3/3], Step [1938/3236], Loss: 1.9042, Perplexity: 6.7140
Epoch [3/3], Step [1939/3236], Loss: 1.8978, Perplexity: 6.6714
Epoch [3/3], Step [1940/3236], Loss: 2.0813, Perplexity: 8.0149
Epoch [3/3], Step [1941/3236], Loss: 2.0865, Perplexity: 8.0567
Epoch [3/3], Step [1942/3236], Loss: 1.8759, Perplexity: 6.5266
Epoch [3/3], Step [1943/3236], Loss: 1.8564, Perplexity: 6.4007
Epoch [3/3], Step [1944/3236], Loss: 1.9403, Perplexity: 6.9605
Epoch [3/3], Step [1945/3236], Loss: 1.7683, Perplexity: 5.8609
Epoch [3/3], Step [1946/3236], Loss: 2.2085, Perplexity: 9.1017
Epoch [3/3], Step [1947/3236], Loss: 1.9331, Perplexity: 6.9106
Epoch [3/3], Step [1948/3236], Loss: 2.1695, Perplexity: 8.7539
Epoch [3/3], Step [1949/3236], Loss: 1.8220, Perplexity: 6.1842
Epoch [3/3], Step [1950/3236], Loss: 1.9430, Perplexity: 6.9797
Epoch [3/3], Step [1951/3236], Loss: 2.1489, Perplexity: 8.5752
Epoch [3/3], Step [1952/3236], Loss: 2.7401, Perplexity: 15.4884
Epoch [3/3], Step [1953/3236], Loss: 1.8941, Perplexity: 6.6464
Epoch [3/3], Step [1954/3236], Loss: 1.9607, Perplexity: 7.1043
Epoch [3/3], Step [1955/3236], Loss: 1.8919, Perplexity: 6.6318
Epoch [3/3], Step [1956/3236], Loss: 1.9115, Perplexity: 6.7630
Epoch [3/3], Step [1957/3236], Loss: 2.1873, Perplexity: 8.9107
Epoch [3/3], Step [1958/3236], Loss: 1.8730, Perplexity: 6.5080
Epoch [3/3], Step [1959/3236], Loss: 1.9729, Perplexity: 7.1912
Epoch [3/3], Step [1960/3236], Loss: 1.8965, Perplexity: 6.6629
Epoch [3/3], Step [1961/3236], Loss: 1.8246, Perplexity: 6.2004
Epoch [3/3], Step [1962/3236], Loss: 2.0562, Perplexity: 7.8160
Epoch [3/3], Step [1963/3236], Loss: 1.9413, Perplexity: 6.9681
Epoch [3/3], Step [1964/3236], Loss: 2.1859, Perplexity: 8.8985
Epoch [3/3], Step [1965/3236], Loss: 1.8993, Perplexity: 6.6810
Epoch [3/3], Step [1966/3236], Loss: 3.4304, Perplexity: 30.8875
Epoch [3/3], Step [1967/3236], Loss: 2.3046, Perplexity: 10.0197
Epoch [3/3], Step [1968/3236], Loss: 2.0743, Perplexity: 7.9588
Epoch [3/3], Step [1969/3236], Loss: 2.0100, Perplexity: 7.4636
Epoch [3/3], Step [1970/3236], Loss: 2.3864, Perplexity: 10.8741
Epoch [3/3], Step [1971/3236], Loss: 1.8602, Perplexity: 6.4248
Epoch [3/3], Step [1972/3236], Loss: 2.0433, Perplexity: 7.7160
Epoch [3/3], Step [1973/3236], Loss: 1.9579, Perplexity: 7.0846
Epoch [3/3], Step [1974/3236], Loss: 1.8389, Perplexity: 6.2896
Epoch [3/3], Step [1975/3236], Loss: 2.1568, Perplexity: 8.6431
Epoch [3/3], Step [1976/3236], Loss: 1.8773, Perplexity: 6.5356
Epoch [3/3], Step [1977/3236], Loss: 1.9993, Perplexity: 7.3840
Epoch [3/3], Step [1978/3236], Loss: 2.0141, Perplexity: 7.4937
Epoch [3/3], Step [1979/3236], Loss: 1.9699, Perplexity: 7.1701
Epoch [3/3], Step [1980/3236], Loss: 1.9617, Perplexity: 7.1115
Epoch [3/3], Step [1981/3236], Loss: 1.8683, Perplexity: 6.4773
Epoch [3/3], Step [1982/3236], Loss: 2.0737, Perplexity: 7.9543
Epoch [3/3], Step [1983/3236], Loss: 1.9600, Perplexity: 7.0993
Epoch [3/3], Step [1984/3236], Loss: 1.8340, Perplexity: 6.2592
Epoch [3/3], Step [1985/3236], Loss: 2.0126, Perplexity: 7.4829
Epoch [3/3], Step [1986/3236], Loss: 1.8643, Perplexity: 6.4515
Epoch [3/3], Step [1987/3236], Loss: 1.8684, Perplexity: 6.4776
Epoch [3/3], Step [1988/3236], Loss: 2.2803, Perplexity: 9.7791
Epoch [3/3], Step [1989/3236], Loss: 1.9572, Perplexity: 7.0797
Epoch [3/3], Step [1990/3236], Loss: 1.8951, Perplexity: 6.6531
Epoch [3/3], Step [1991/3236], Loss: 1.8984, Perplexity: 6.6750
Epoch [3/3], Step [1992/3236], Loss: 2.0917, Perplexity: 8.0985
Epoch [3/3], Step [1993/3236], Loss: 1.9488, Perplexity: 7.0204
Epoch [3/3], Step [1994/3236], Loss: 2.1147, Perplexity: 8.2868
Epoch [3/3], Step [1995/3236], Loss: 1.9390, Perplexity: 6.9518
Epoch [3/3], Step [1996/3236], Loss: 2.0690, Perplexity: 7.9170
Epoch [3/3], Step [1997/3236], Loss: 1.8723, Perplexity: 6.5030
Epoch [3/3], Step [1998/3236], Loss: 1.9873, Perplexity: 7.2955
Epoch [3/3], Step [1999/3236], Loss: 1.9432, Perplexity: 6.9813
Epoch [3/3], Step [2000/3236], Loss: 2.0326, Perplexity: 7.6341
Epoch [3/3], Step [2001/3236], Loss: 1.9838, Perplexity: 7.2706
Epoch [3/3], Step [2002/3236], Loss: 1.9668, Perplexity: 7.1480
Epoch [3/3], Step [2003/3236], Loss: 1.8449, Perplexity: 6.3272
Epoch [3/3], Step [2004/3236], Loss: 1.8878, Perplexity: 6.6046
Epoch [3/3], Step [2005/3236], Loss: 2.0260, Perplexity: 7.5840
Epoch [3/3], Step [2006/3236], Loss: 2.3194, Perplexity: 10.1699
Epoch [3/3], Step [2007/3236], Loss: 1.9361, Perplexity: 6.9320
Epoch [3/3], Step [2008/3236], Loss: 2.3395, Perplexity: 10.3762
Epoch [3/3], Step [2009/3236], Loss: 1.9502, Perplexity: 7.0299
Epoch [3/3], Step [2010/3236], Loss: 1.9227, Perplexity: 6.8393
Epoch [3/3], Step [2011/3236], Loss: 1.9818, Perplexity: 7.2559
Epoch [3/3], Step [2012/3236], Loss: 1.9734, Perplexity: 7.1951
Epoch [3/3], Step [2013/3236], Loss: 2.0746, Perplexity: 7.9613
Epoch [3/3], Step [2014/3236], Loss: 1.9381, Perplexity: 6.9456
Epoch [3/3], Step [2015/3236], Loss: 1.8534, Perplexity: 6.3814
Epoch [3/3], Step [2016/3236], Loss: 2.0381, Perplexity: 7.6757
Epoch [3/3], Step [2017/3236], Loss: 1.9945, Perplexity: 7.3484
Epoch [3/3], Step [2018/3236], Loss: 1.7404, Perplexity: 5.6996
Epoch [3/3], Step [2019/3236], Loss: 1.9045, Perplexity: 6.7158
Epoch [3/3], Step [2020/3236], Loss: 1.8405, Perplexity: 6.2994
Epoch [3/3], Step [2021/3236], Loss: 2.0237, Perplexity: 7.5661
Epoch [3/3], Step [2022/3236], Loss: 1.9927, Perplexity: 7.3352
Epoch [3/3], Step [2023/3236], Loss: 1.8577, Perplexity: 6.4092
Epoch [3/3], Step [2024/3236], Loss: 1.9147, Perplexity: 6.7846
Epoch [3/3], Step [2025/3236], Loss: 2.1417, Perplexity: 8.5140
Epoch [3/3], Step [2026/3236], Loss: 1.8651, Perplexity: 6.4568
Epoch [3/3], Step [2027/3236], Loss: 1.8921, Perplexity: 6.6330
Epoch [3/3], Step [2028/3236], Loss: 1.8775, Perplexity: 6.5371
Epoch [3/3], Step [2029/3236], Loss: 2.0074, Perplexity: 7.4437
Epoch [3/3], Step [2030/3236], Loss: 1.9510, Perplexity: 7.0354
Epoch [3/3], Step [2031/3236], Loss: 1.9139, Perplexity: 6.7795
Epoch [3/3], Step [2032/3236], Loss: 2.0176, Perplexity: 7.5202
Epoch [3/3], Step [2033/3236], Loss: 2.0298, Perplexity: 7.6129
Epoch [3/3], Step [2034/3236], Loss: 1.9960, Perplexity: 7.3592
Epoch [3/3], Step [2035/3236], Loss: 1.9502, Perplexity: 7.0301
Epoch [3/3], Step [2036/3236], Loss: 1.9118, Perplexity: 6.7653
Epoch [3/3], Step [2037/3236], Loss: 1.8973, Perplexity: 6.6680
Epoch [3/3], Step [2038/3236], Loss: 1.9456, Perplexity: 6.9979
Epoch [3/3], Step [2039/3236], Loss: 2.1969, Perplexity: 8.9975
Epoch [3/3], Step [2040/3236], Loss: 1.9455, Perplexity: 6.9974
Epoch [3/3], Step [2041/3236], Loss: 1.9363, Perplexity: 6.9328
Epoch [3/3], Step [2042/3236], Loss: 2.2173, Perplexity: 9.1824
Epoch [3/3], Step [2043/3236], Loss: 2.1269, Perplexity: 8.3892
Epoch [3/3], Step [2044/3236], Loss: 1.8983, Perplexity: 6.6746
Epoch [3/3], Step [2045/3236], Loss: 1.8163, Perplexity: 6.1492
Epoch [3/3], Step [2046/3236], Loss: 2.0781, Perplexity: 7.9890
Epoch [3/3], Step [2047/3236], Loss: 1.8308, Perplexity: 6.2386
Epoch [3/3], Step [2048/3236], Loss: 1.9757, Perplexity: 7.2118
Epoch [3/3], Step [2049/3236], Loss: 1.8609, Perplexity: 6.4295
Epoch [3/3], Step [2050/3236], Loss: 2.1561, Perplexity: 8.6373
Epoch [3/3], Step [2051/3236], Loss: 1.8543, Perplexity: 6.3871
Epoch [3/3], Step [2052/3236], Loss: 1.9423, Perplexity: 6.9744
Epoch [3/3], Step [2053/3236], Loss: 1.8792, Perplexity: 6.5485
Epoch [3/3], Step [2054/3236], Loss: 1.9591, Perplexity: 7.0926
Epoch [3/3], Step [2055/3236], Loss: 1.7773, Perplexity: 5.9139
Epoch [3/3], Step [2056/3236], Loss: 2.7574, Perplexity: 15.7596
Epoch [3/3], Step [2057/3236], Loss: 2.2120, Perplexity: 9.1342
Epoch [3/3], Step [2058/3236], Loss: 1.8874, Perplexity: 6.6019
Epoch [3/3], Step [2059/3236], Loss: 1.8741, Perplexity: 6.5148
Epoch [3/3], Step [2060/3236], Loss: 1.8473, Perplexity: 6.3428
Epoch [3/3], Step [2061/3236], Loss: 1.9082, Perplexity: 6.7412
Epoch [3/3], Step [2062/3236], Loss: 2.1497, Perplexity: 8.5820
Epoch [3/3], Step [2063/3236], Loss: 1.8931, Perplexity: 6.6402
Epoch [3/3], Step [2064/3236], Loss: 1.8633, Perplexity: 6.4452
Epoch [3/3], Step [2065/3236], Loss: 1.9847, Perplexity: 7.2766
Epoch [3/3], Step [2066/3236], Loss: 2.1799, Perplexity: 8.8458
Epoch [3/3], Step [2067/3236], Loss: 1.9325, Perplexity: 6.9068
Epoch [3/3], Step [2068/3236], Loss: 1.9612, Perplexity: 7.1077
Epoch [3/3], Step [2069/3236], Loss: 1.9692, Perplexity: 7.1653
Epoch [3/3], Step [2070/3236], Loss: 2.0602, Perplexity: 7.8474
Epoch [3/3], Step [2071/3236], Loss: 1.9241, Perplexity: 6.8490
Epoch [3/3], Step [2072/3236], Loss: 1.8993, Perplexity: 6.6814
Epoch [3/3], Step [2073/3236], Loss: 1.9436, Perplexity: 6.9835
Epoch [3/3], Step [2074/3236], Loss: 2.0008, Perplexity: 7.3951
Epoch [3/3], Step [2075/3236], Loss: 1.9226, Perplexity: 6.8387
Epoch [3/3], Step [2076/3236], Loss: 1.8944, Perplexity: 6.6483
Epoch [3/3], Step [2077/3236], Loss: 1.9008, Perplexity: 6.6913
Epoch [3/3], Step [2078/3236], Loss: 1.7664, Perplexity: 5.8500
Epoch [3/3], Step [2079/3236], Loss: 1.8755, Perplexity: 6.5239
Epoch [3/3], Step [2080/3236], Loss: 2.2084, Perplexity: 9.1013
Epoch [3/3], Step [2081/3236], Loss: 2.0550, Perplexity: 7.8068
Epoch [3/3], Step [2082/3236], Loss: 1.8444, Perplexity: 6.3241
Epoch [3/3], Step [2083/3236], Loss: 1.8603, Perplexity: 6.4256
Epoch [3/3], Step [2084/3236], Loss: 1.7317, Perplexity: 5.6504
Epoch [3/3], Step [2085/3236], Loss: 1.8122, Perplexity: 6.1241
Epoch [3/3], Step [2086/3236], Loss: 1.9802, Perplexity: 7.2444
Epoch [3/3], Step [2087/3236], Loss: 1.9117, Perplexity: 6.7647
Epoch [3/3], Step [2088/3236], Loss: 1.8436, Perplexity: 6.3191
Epoch [3/3], Step [2089/3236], Loss: 1.9358, Perplexity: 6.9297
Epoch [3/3], Step [2090/3236], Loss: 1.8322, Perplexity: 6.2478
Epoch [3/3], Step [2091/3236], Loss: 1.8772, Perplexity: 6.5352
Epoch [3/3], Step [2092/3236], Loss: 1.9121, Perplexity: 6.7672
Epoch [3/3], Step [2093/3236], Loss: 2.1461, Perplexity: 8.5515
Epoch [3/3], Step [2094/3236], Loss: 2.0987, Perplexity: 8.1552
Epoch [3/3], Step [2095/3236], Loss: 1.8191, Perplexity: 6.1663
Epoch [3/3], Step [2096/3236], Loss: 1.8447, Perplexity: 6.3261
Epoch [3/3], Step [2097/3236], Loss: 1.9746, Perplexity: 7.2037
Epoch [3/3], Step [2098/3236], Loss: 2.0690, Perplexity: 7.9166
Epoch [3/3], Step [2099/3236], Loss: 1.9469, Perplexity: 7.0066
Epoch [3/3], Step [2100/3236], Loss: 1.7745, Perplexity: 5.8972
Epoch [3/3], Step [2101/3236], Loss: 2.1331, Perplexity: 8.4410
Epoch [3/3], Step [2102/3236], Loss: 2.0506, Perplexity: 7.7728
Epoch [3/3], Step [2103/3236], Loss: 1.9503, Perplexity: 7.0305
Epoch [3/3], Step [2104/3236], Loss: 2.5443, Perplexity: 12.7347
Epoch [3/3], Step [2105/3236], Loss: 1.9807, Perplexity: 7.2475
Epoch [3/3], Step [2106/3236], Loss: 1.9522, Perplexity: 7.0438
Epoch [3/3], Step [2107/3236], Loss: 2.0310, Perplexity: 7.6221
Epoch [3/3], Step [2108/3236], Loss: 2.2038, Perplexity: 9.0590
Epoch [3/3], Step [2109/3236], Loss: 1.9139, Perplexity: 6.7795
Epoch [3/3], Step [2110/3236], Loss: 2.1560, Perplexity: 8.6368
Epoch [3/3], Step [2111/3236], Loss: 1.9408, Perplexity: 6.9642
Epoch [3/3], Step [2112/3236], Loss: 2.1373, Perplexity: 8.4767
Epoch [3/3], Step [2113/3236], Loss: 1.9956, Perplexity: 7.3563
Epoch [3/3], Step [2114/3236], Loss: 1.9959, Perplexity: 7.3589
Epoch [3/3], Step [2115/3236], Loss: 1.9038, Perplexity: 6.7114
Epoch [3/3], Step [2116/3236], Loss: 2.9769, Perplexity: 19.6264
Epoch [3/3], Step [2117/3236], Loss: 1.9149, Perplexity: 6.7862
Epoch [3/3], Step [2118/3236], Loss: 2.0149, Perplexity: 7.4996
Epoch [3/3], Step [2119/3236], Loss: 1.9032, Perplexity: 6.7070
Epoch [3/3], Step [2120/3236], Loss: 1.8486, Perplexity: 6.3511
Epoch [3/3], Step [2121/3236], Loss: 2.1176, Perplexity: 8.3111
Epoch [3/3], Step [2122/3236], Loss: 1.9018, Perplexity: 6.6978
Epoch [3/3], Step [2123/3236], Loss: 1.9470, Perplexity: 7.0077
Epoch [3/3], Step [2124/3236], Loss: 1.8955, Perplexity: 6.6557
Epoch [3/3], Step [2125/3236], Loss: 1.8862, Perplexity: 6.5941
Epoch [3/3], Step [2126/3236], Loss: 1.9687, Perplexity: 7.1615
Epoch [3/3], Step [2127/3236], Loss: 1.8841, Perplexity: 6.5803
Epoch [3/3], Step [2128/3236], Loss: 2.5175, Perplexity: 12.3980
Epoch [3/3], Step [2129/3236], Loss: 1.8864, Perplexity: 6.5955
Epoch [3/3], Step [2130/3236], Loss: 1.8306, Perplexity: 6.2378
Epoch [3/3], Step [2131/3236], Loss: 1.9012, Perplexity: 6.6942
Epoch [3/3], Step [2132/3236], Loss: 2.7764, Perplexity: 16.0610
Epoch [3/3], Step [2133/3236], Loss: 1.8440, Perplexity: 6.3218
Epoch [3/3], Step [2134/3236], Loss: 1.8766, Perplexity: 6.5312
Epoch [3/3], Step [2135/3236], Loss: 1.8668, Perplexity: 6.4674
Epoch [3/3], Step [2136/3236], Loss: 1.8655, Perplexity: 6.4594
Epoch [3/3], Step [2137/3236], Loss: 1.8157, Perplexity: 6.1455
Epoch [3/3], Step [2138/3236], Loss: 2.0245, Perplexity: 7.5727
Epoch [3/3], Step [2139/3236], Loss: 2.1318, Perplexity: 8.4301
Epoch [3/3], Step [2140/3236], Loss: 1.7981, Perplexity: 6.0384
Epoch [3/3], Step [2141/3236], Loss: 2.0069, Perplexity: 7.4399
Epoch [3/3], Step [2142/3236], Loss: 1.9558, Perplexity: 7.0699
Epoch [3/3], Step [2143/3236], Loss: 1.8918, Perplexity: 6.6314
Epoch [3/3], Step [2144/3236], Loss: 1.8520, Perplexity: 6.3728
Epoch [3/3], Step [2145/3236], Loss: 1.9064, Perplexity: 6.7290
Epoch [3/3], Step [2146/3236], Loss: 2.4728, Perplexity: 11.8550
Epoch [3/3], Step [2147/3236], Loss: 2.0580, Perplexity: 7.8307
Epoch [3/3], Step [2148/3236], Loss: 2.0202, Perplexity: 7.5398
Epoch [3/3], Step [2149/3236], Loss: 1.8423, Perplexity: 6.3108
Epoch [3/3], Step [2150/3236], Loss: 2.1107, Perplexity: 8.2537
Epoch [3/3], Step [2151/3236], Loss: 2.0363, Perplexity: 7.6623
Epoch [3/3], Step [2152/3236], Loss: 1.8962, Perplexity: 6.6606
Epoch [3/3], Step [2153/3236], Loss: 1.8667, Perplexity: 6.4666
Epoch [3/3], Step [2154/3236], Loss: 2.7542, Perplexity: 15.7080
Epoch [3/3], Step [2155/3236], Loss: 1.9099, Perplexity: 6.7524
Epoch [3/3], Step [2156/3236], Loss: 2.0810, Perplexity: 8.0128
Epoch [3/3], Step [2157/3236], Loss: 2.2737, Perplexity: 9.7150
Epoch [3/3], Step [2158/3236], Loss: 1.8951, Perplexity: 6.6535
Epoch [3/3], Step [2159/3236], Loss: 1.9071, Perplexity: 6.7334
Epoch [3/3], Step [2160/3236], Loss: 2.0109, Perplexity: 7.4703
Epoch [3/3], Step [2161/3236], Loss: 1.9546, Perplexity: 7.0610
Epoch [3/3], Step [2162/3236], Loss: 1.9618, Perplexity: 7.1120
Epoch [3/3], Step [2163/3236], Loss: 1.8977, Perplexity: 6.6704
Epoch [3/3], Step [2164/3236], Loss: 2.3518, Perplexity: 10.5042
Epoch [3/3], Step [2165/3236], Loss: 1.9039, Perplexity: 6.7121
Epoch [3/3], Step [2166/3236], Loss: 2.2343, Perplexity: 9.3400
Epoch [3/3], Step [2167/3236], Loss: 1.9740, Perplexity: 7.1997
Epoch [3/3], Step [2168/3236], Loss: 2.0195, Perplexity: 7.5342
Epoch [3/3], Step [2169/3236], Loss: 1.8684, Perplexity: 6.4776
Epoch [3/3], Step [2170/3236], Loss: 1.8629, Perplexity: 6.4421
Epoch [3/3], Step [2171/3236], Loss: 1.8651, Perplexity: 6.4564
Epoch [3/3], Step [2172/3236], Loss: 2.0514, Perplexity: 7.7789
Epoch [3/3], Step [2173/3236], Loss: 2.7493, Perplexity: 15.6322
Epoch [3/3], Step [2174/3236], Loss: 1.9109, Perplexity: 6.7591
Epoch [3/3], Step [2175/3236], Loss: 1.8748, Perplexity: 6.5193
Epoch [3/3], Step [2176/3236], Loss: 1.9774, Perplexity: 7.2238
Epoch [3/3], Step [2177/3236], Loss: 1.8194, Perplexity: 6.1683
Epoch [3/3], Step [2178/3236], Loss: 2.0537, Perplexity: 7.7964
Epoch [3/3], Step [2179/3236], Loss: 2.0374, Perplexity: 7.6703
Epoch [3/3], Step [2180/3236], Loss: 1.8970, Perplexity: 6.6661
Epoch [3/3], Step [2181/3236], Loss: 1.8982, Perplexity: 6.6737
Epoch [3/3], Step [2182/3236], Loss: 1.9178, Perplexity: 6.8061
Epoch [3/3], Step [2183/3236], Loss: 1.9866, Perplexity: 7.2908
Epoch [3/3], Step [2184/3236], Loss: 2.0365, Perplexity: 7.6638
Epoch [3/3], Step [2185/3236], Loss: 2.0294, Perplexity: 7.6095
Epoch [3/3], Step [2186/3236], Loss: 1.8652, Perplexity: 6.4571
Epoch [3/3], Step [2187/3236], Loss: 1.9619, Perplexity: 7.1131
Epoch [3/3], Step [2188/3236], Loss: 1.9035, Perplexity: 6.7094
Epoch [3/3], Step [2189/3236], Loss: 1.9173, Perplexity: 6.8023
Epoch [3/3], Step [2190/3236], Loss: 1.8456, Perplexity: 6.3322
Epoch [3/3], Step [2191/3236], Loss: 1.9897, Perplexity: 7.3130
Epoch [3/3], Step [2192/3236], Loss: 1.9281, Perplexity: 6.8763
Epoch [3/3], Step [2193/3236], Loss: 1.8410, Perplexity: 6.3027
Epoch [3/3], Step [2194/3236], Loss: 1.9625, Perplexity: 7.1173
Epoch [3/3], Step [2195/3236], Loss: 1.8312, Perplexity: 6.2414
Epoch [3/3], Step [2196/3236], Loss: 2.1298, Perplexity: 8.4134
Epoch [3/3], Step [2197/3236], Loss: 1.8108, Perplexity: 6.1154
Epoch [3/3], Step [2198/3236], Loss: 3.6152, Perplexity: 37.1604
Epoch [3/3], Step [2199/3236], Loss: 2.0578, Perplexity: 7.8286
Epoch [3/3], Step [2200/3236], Loss: 2.2513, Perplexity: 9.4999
Epoch [3/3], Step [2201/3236], Loss: 2.4186, Perplexity: 11.2297
Epoch [3/3], Step [2202/3236], Loss: 1.9973, Perplexity: 7.3693
Epoch [3/3], Step [2203/3236], Loss: 1.9484, Perplexity: 7.0172
Epoch [3/3], Step [2204/3236], Loss: 2.0351, Perplexity: 7.6530
Epoch [3/3], Step [2205/3236], Loss: 2.4716, Perplexity: 11.8412
Epoch [3/3], Step [2206/3236], Loss: 1.9783, Perplexity: 7.2307
Epoch [3/3], Step [2207/3236], Loss: 2.1994, Perplexity: 9.0196
Epoch [3/3], Step [2208/3236], Loss: 1.9249, Perplexity: 6.8545
Epoch [3/3], Step [2209/3236], Loss: 1.9422, Perplexity: 6.9741
Epoch [3/3], Step [2210/3236], Loss: 1.8657, Perplexity: 6.4608
Epoch [3/3], Step [2211/3236], Loss: 2.2853, Perplexity: 9.8285
Epoch [3/3], Step [2212/3236], Loss: 2.0967, Perplexity: 8.1389
Epoch [3/3], Step [2213/3236], Loss: 1.8759, Perplexity: 6.5269
Epoch [3/3], Step [2214/3236], Loss: 2.0015, Perplexity: 7.4004
Epoch [3/3], Step [2215/3236], Loss: 1.9317, Perplexity: 6.9010
Epoch [3/3], Step [2216/3236], Loss: 1.8899, Perplexity: 6.6190
Epoch [3/3], Step [2217/3236], Loss: 1.9648, Perplexity: 7.1333
Epoch [3/3], Step [2218/3236], Loss: 1.9461, Perplexity: 7.0015
Epoch [3/3], Step [2219/3236], Loss: 1.9127, Perplexity: 6.7713
Epoch [3/3], Step [2220/3236], Loss: 1.9272, Perplexity: 6.8704
Epoch [3/3], Step [2221/3236], Loss: 1.9993, Perplexity: 7.3838
Epoch [3/3], Step [2222/3236], Loss: 1.8115, Perplexity: 6.1196
Epoch [3/3], Step [2223/3236], Loss: 2.5426, Perplexity: 12.7128
Epoch [3/3], Step [2224/3236], Loss: 1.8438, Perplexity: 6.3204
Epoch [3/3], Step [2225/3236], Loss: 2.2549, Perplexity: 9.5342
Epoch [3/3], Step [2226/3236], Loss: 1.9897, Perplexity: 7.3135
Epoch [3/3], Step [2227/3236], Loss: 1.9304, Perplexity: 6.8924
Epoch [3/3], Step [2228/3236], Loss: 1.8775, Perplexity: 6.5369
Epoch [3/3], Step [2229/3236], Loss: 2.0855, Perplexity: 8.0482
Epoch [3/3], Step [2230/3236], Loss: 2.1183, Perplexity: 8.3166
Epoch [3/3], Step [2231/3236], Loss: 1.9558, Perplexity: 7.0699
Epoch [3/3], Step [2232/3236], Loss: 1.8507, Perplexity: 6.3644
Epoch [3/3], Step [2233/3236], Loss: 1.8392, Perplexity: 6.2917
Epoch [3/3], Step [2234/3236], Loss: 1.9797, Perplexity: 7.2405
Epoch [3/3], Step [2235/3236], Loss: 1.9654, Perplexity: 7.1379
Epoch [3/3], Step [2236/3236], Loss: 1.9018, Perplexity: 6.6983
Epoch [3/3], Step [2237/3236], Loss: 1.8722, Perplexity: 6.5028
Epoch [3/3], Step [2238/3236], Loss: 1.8667, Perplexity: 6.4668
Epoch [3/3], Step [2239/3236], Loss: 1.9107, Perplexity: 6.7580
Epoch [3/3], Step [2240/3236], Loss: 4.5235, Perplexity: 92.1582
Epoch [3/3], Step [2241/3236], Loss: 1.8778, Perplexity: 6.5394
Epoch [3/3], Step [2242/3236], Loss: 1.7953, Perplexity: 6.0214
Epoch [3/3], Step [2243/3236], Loss: 2.0359, Perplexity: 7.6593
Epoch [3/3], Step [2244/3236], Loss: 1.9048, Perplexity: 6.7182
Epoch [3/3], Step [2245/3236], Loss: 2.0508, Perplexity: 7.7737
Epoch [3/3], Step [2246/3236], Loss: 2.6198, Perplexity: 13.7323
Epoch [3/3], Step [2247/3236], Loss: 2.1363, Perplexity: 8.4682
Epoch [3/3], Step [2248/3236], Loss: 1.9435, Perplexity: 6.9831
Epoch [3/3], Step [2249/3236], Loss: 1.8641, Perplexity: 6.4502
Epoch [3/3], Step [2250/3236], Loss: 1.9252, Perplexity: 6.8567
Epoch [3/3], Step [2251/3236], Loss: 1.9610, Perplexity: 7.1061
Epoch [3/3], Step [2252/3236], Loss: 2.2829, Perplexity: 9.8051
Epoch [3/3], Step [2253/3236], Loss: 1.7682, Perplexity: 5.8604
Epoch [3/3], Step [2254/3236], Loss: 2.2333, Perplexity: 9.3302
Epoch [3/3], Step [2255/3236], Loss: 2.1705, Perplexity: 8.7624
Epoch [3/3], Step [2256/3236], Loss: 1.8785, Perplexity: 6.5437
Epoch [3/3], Step [2257/3236], Loss: 1.8628, Perplexity: 6.4417
Epoch [3/3], Step [2258/3236], Loss: 1.9578, Perplexity: 7.0836
Epoch [3/3], Step [2259/3236], Loss: 1.9577, Perplexity: 7.0828
Epoch [3/3], Step [2260/3236], Loss: 2.0520, Perplexity: 7.7831
Epoch [3/3], Step [2261/3236], Loss: 2.3056, Perplexity: 10.0302
Epoch [3/3], Step [2262/3236], Loss: 2.1780, Perplexity: 8.8283
Epoch [3/3], Step [2263/3236], Loss: 2.0382, Perplexity: 7.6768
Epoch [3/3], Step [2264/3236], Loss: 1.8389, Perplexity: 6.2894
Epoch [3/3], Step [2265/3236], Loss: 1.9765, Perplexity: 7.2174
Epoch [3/3], Step [2266/3236], Loss: 1.8900, Perplexity: 6.6196
Epoch [3/3], Step [2267/3236], Loss: 2.1443, Perplexity: 8.5360
Epoch [3/3], Step [2268/3236], Loss: 1.8782, Perplexity: 6.5416
Epoch [3/3], Step [2269/3236], Loss: 1.8857, Perplexity: 6.5910
Epoch [3/3], Step [2270/3236], Loss: 2.0409, Perplexity: 7.6973
Epoch [3/3], Step [2271/3236], Loss: 1.8918, Perplexity: 6.6313
Epoch [3/3], Step [2272/3236], Loss: 1.9092, Perplexity: 6.7475
Epoch [3/3], Step [2273/3236], Loss: 2.2792, Perplexity: 9.7684
Epoch [3/3], Step [2274/3236], Loss: 1.9704, Perplexity: 7.1739
Epoch [3/3], Step [2275/3236], Loss: 2.7348, Perplexity: 15.4074
Epoch [3/3], Step [2276/3236], Loss: 2.1141, Perplexity: 8.2819
Epoch [3/3], Step [2277/3236], Loss: 1.9885, Perplexity: 7.3047
Epoch [3/3], Step [2278/3236], Loss: 2.0006, Perplexity: 7.3936
Epoch [3/3], Step [2279/3236], Loss: 2.3915, Perplexity: 10.9298
Epoch [3/3], Step [2280/3236], Loss: 1.8660, Perplexity: 6.4622
Epoch [3/3], Step [2281/3236], Loss: 2.0481, Perplexity: 7.7534
Epoch [3/3], Step [2282/3236], Loss: 2.2379, Perplexity: 9.3738
Epoch [3/3], Step [2283/3236], Loss: 1.9830, Perplexity: 7.2648
Epoch [3/3], Step [2284/3236], Loss: 1.9628, Perplexity: 7.1189
Epoch [3/3], Step [2285/3236], Loss: 1.9146, Perplexity: 6.7842
Epoch [3/3], Step [2286/3236], Loss: 1.9360, Perplexity: 6.9312
Epoch [3/3], Step [2287/3236], Loss: 1.9811, Perplexity: 7.2507
Epoch [3/3], Step [2288/3236], Loss: 2.0951, Perplexity: 8.1265
Epoch [3/3], Step [2289/3236], Loss: 1.8923, Perplexity: 6.6344
Epoch [3/3], Step [2290/3236], Loss: 2.0179, Perplexity: 7.5223
Epoch [3/3], Step [2291/3236], Loss: 2.1635, Perplexity: 8.7016
Epoch [3/3], Step [2292/3236], Loss: 1.8557, Perplexity: 6.3963
Epoch [3/3], Step [2293/3236], Loss: 2.2996, Perplexity: 9.9699
Epoch [3/3], Step [2294/3236], Loss: 1.9144, Perplexity: 6.7831
Epoch [3/3], Step [2295/3236], Loss: 2.0555, Perplexity: 7.8107
Epoch [3/3], Step [2296/3236], Loss: 2.0230, Perplexity: 7.5613
Epoch [3/3], Step [2297/3236], Loss: 2.0229, Perplexity: 7.5602
Epoch [3/3], Step [2298/3236], Loss: 2.0500, Perplexity: 7.7676
Epoch [3/3], Step [2299/3236], Loss: 1.9842, Perplexity: 7.2731
Epoch [3/3], Step [2300/3236], Loss: 1.9013, Perplexity: 6.6945
Epoch [3/3], Step [2301/3236], Loss: 2.7249, Perplexity: 15.2541
Epoch [3/3], Step [2302/3236], Loss: 1.9193, Perplexity: 6.8165
Epoch [3/3], Step [2303/3236], Loss: 2.0722, Perplexity: 7.9426
Epoch [3/3], Step [2304/3236], Loss: 1.8785, Perplexity: 6.5439
Epoch [3/3], Step [2305/3236], Loss: 1.8383, Perplexity: 6.2860
Epoch [3/3], Step [2306/3236], Loss: 1.7994, Perplexity: 6.0461
Epoch [3/3], Step [2307/3236], Loss: 2.0536, Perplexity: 7.7959
Epoch [3/3], Step [2308/3236], Loss: 2.0536, Perplexity: 7.7958
Epoch [3/3], Step [2309/3236], Loss: 1.9159, Perplexity: 6.7931
Epoch [3/3], Step [2310/3236], Loss: 1.9901, Perplexity: 7.3163
Epoch [3/3], Step [2311/3236], Loss: 1.9058, Perplexity: 6.7246
Epoch [3/3], Step [2312/3236], Loss: 1.8689, Perplexity: 6.4809
Epoch [3/3], Step [2313/3236], Loss: 1.9082, Perplexity: 6.7410
Epoch [3/3], Step [2314/3236], Loss: 1.8704, Perplexity: 6.4911
Epoch [3/3], Step [2315/3236], Loss: 1.8802, Perplexity: 6.5547
Epoch [3/3], Step [2316/3236], Loss: 1.9101, Perplexity: 6.7536
Epoch [3/3], Step [2317/3236], Loss: 1.8921, Perplexity: 6.6336
Epoch [3/3], Step [2318/3236], Loss: 2.3564, Perplexity: 10.5533
Epoch [3/3], Step [2319/3236], Loss: 1.8673, Perplexity: 6.4706
Epoch [3/3], Step [2320/3236], Loss: 2.3075, Perplexity: 10.0488
Epoch [3/3], Step [2321/3236], Loss: 1.9969, Perplexity: 7.3662
Epoch [3/3], Step [2322/3236], Loss: 1.8811, Perplexity: 6.5606
Epoch [3/3], Step [2323/3236], Loss: 1.9690, Perplexity: 7.1633
Epoch [3/3], Step [2324/3236], Loss: 2.1547, Perplexity: 8.6252
Epoch [3/3], Step [2325/3236], Loss: 2.2422, Perplexity: 9.4136
Epoch [3/3], Step [2326/3236], Loss: 2.0640, Perplexity: 7.8774
Epoch [3/3], Step [2327/3236], Loss: 1.8719, Perplexity: 6.5006
Epoch [3/3], Step [2328/3236], Loss: 2.0287, Perplexity: 7.6042
Epoch [3/3], Step [2329/3236], Loss: 1.9359, Perplexity: 6.9301
Epoch [3/3], Step [2330/3236], Loss: 1.9595, Perplexity: 7.0961
Epoch [3/3], Step [2331/3236], Loss: 2.1576, Perplexity: 8.6502
Epoch [3/3], Step [2332/3236], Loss: 1.8006, Perplexity: 6.0533
Epoch [3/3], Step [2333/3236], Loss: 1.9314, Perplexity: 6.8993
Epoch [3/3], Step [2334/3236], Loss: 1.8966, Perplexity: 6.6635
Epoch [3/3], Step [2335/3236], Loss: 1.9199, Perplexity: 6.8203
Epoch [3/3], Step [2336/3236], Loss: 1.9725, Perplexity: 7.1884
Epoch [3/3], Step [2337/3236], Loss: 1.8945, Perplexity: 6.6491
Epoch [3/3], Step [2338/3236], Loss: 1.9471, Perplexity: 7.0083
Epoch [3/3], Step [2339/3236], Loss: 1.9354, Perplexity: 6.9267
Epoch [3/3], Step [2340/3236], Loss: 1.9358, Perplexity: 6.9298
Epoch [3/3], Step [2341/3236], Loss: 1.9423, Perplexity: 6.9750
Epoch [3/3], Step [2342/3236], Loss: 1.9184, Perplexity: 6.8102
Epoch [3/3], Step [2343/3236], Loss: 1.9541, Perplexity: 7.0575
Epoch [3/3], Step [2344/3236], Loss: 2.4166, Perplexity: 11.2076
Epoch [3/3], Step [2345/3236], Loss: 1.9266, Perplexity: 6.8660
Epoch [3/3], Step [2346/3236], Loss: 1.8740, Perplexity: 6.5142
Epoch [3/3], Step [2347/3236], Loss: 2.0606, Perplexity: 7.8508
Epoch [3/3], Step [2348/3236], Loss: 1.8874, Perplexity: 6.6024
Epoch [3/3], Step [2349/3236], Loss: 2.0261, Perplexity: 7.5848
Epoch [3/3], Step [2350/3236], Loss: 1.9375, Perplexity: 6.9416
Epoch [3/3], Step [2351/3236], Loss: 1.9428, Perplexity: 6.9785
Epoch [3/3], Step [2352/3236], Loss: 1.9711, Perplexity: 7.1786
Epoch [3/3], Step [2353/3236], Loss: 1.7893, Perplexity: 5.9855
Epoch [3/3], Step [2354/3236], Loss: 1.8251, Perplexity: 6.2035
Epoch [3/3], Step [2355/3236], Loss: 1.8576, Perplexity: 6.4081
Epoch [3/3], Step [2356/3236], Loss: 2.3680, Perplexity: 10.6758
Epoch [3/3], Step [2357/3236], Loss: 1.9915, Perplexity: 7.3267
Epoch [3/3], Step [2358/3236], Loss: 1.9482, Perplexity: 7.0159
Epoch [3/3], Step [2359/3236], Loss: 1.9452, Perplexity: 6.9951
Epoch [3/3], Step [2360/3236], Loss: 1.8639, Perplexity: 6.4489
Epoch [3/3], Step [2361/3236], Loss: 2.0150, Perplexity: 7.5008
Epoch [3/3], Step [2362/3236], Loss: 1.8548, Perplexity: 6.3906
Epoch [3/3], Step [2363/3236], Loss: 1.8363, Perplexity: 6.2730
Epoch [3/3], Step [2364/3236], Loss: 1.8626, Perplexity: 6.4407
Epoch [3/3], Step [2365/3236], Loss: 1.8903, Perplexity: 6.6212
Epoch [3/3], Step [2366/3236], Loss: 2.1098, Perplexity: 8.2464
Epoch [3/3], Step [2367/3236], Loss: 1.8829, Perplexity: 6.5725
Epoch [3/3], Step [2368/3236], Loss: 2.1569, Perplexity: 8.6439
Epoch [3/3], Step [2369/3236], Loss: 1.9056, Perplexity: 6.7235
Epoch [3/3], Step [2370/3236], Loss: 1.8302, Perplexity: 6.2352
Epoch [3/3], Step [2371/3236], Loss: 1.9481, Perplexity: 7.0155
Epoch [3/3], Step [2372/3236], Loss: 1.9430, Perplexity: 6.9797
Epoch [3/3], Step [2373/3236], Loss: 1.9735, Perplexity: 7.1957
Epoch [3/3], Step [2374/3236], Loss: 1.7791, Perplexity: 5.9246
Epoch [3/3], Step [2375/3236], Loss: 2.0903, Perplexity: 8.0871
Epoch [3/3], Step [2376/3236], Loss: 1.7959, Perplexity: 6.0249
Epoch [3/3], Step [2377/3236], Loss: 2.0305, Perplexity: 7.6178
Epoch [3/3], Step [2378/3236], Loss: 1.9726, Perplexity: 7.1897
Epoch [3/3], Step [2379/3236], Loss: 1.8765, Perplexity: 6.5309
Epoch [3/3], Step [2380/3236], Loss: 2.0627, Perplexity: 7.8673
Epoch [3/3], Step [2381/3236], Loss: 1.8954, Perplexity: 6.6554
Epoch [3/3], Step [2382/3236], Loss: 1.9929, Perplexity: 7.3368
Epoch [3/3], Step [2383/3236], Loss: 1.8633, Perplexity: 6.4449
Epoch [3/3], Step [2384/3236], Loss: 1.9503, Perplexity: 7.0307
Epoch [3/3], Step [2385/3236], Loss: 2.0110, Perplexity: 7.4706
Epoch [3/3], Step [2386/3236], Loss: 2.7945, Perplexity: 16.3541
Epoch [3/3], Step [2387/3236], Loss: 1.9327, Perplexity: 6.9084
Epoch [3/3], Step [2388/3236], Loss: 2.1042, Perplexity: 8.2004
Epoch [3/3], Step [2389/3236], Loss: 1.9773, Perplexity: 7.2234
Epoch [3/3], Step [2390/3236], Loss: 2.2339, Perplexity: 9.3365
Epoch [3/3], Step [2391/3236], Loss: 1.8657, Perplexity: 6.4606
Epoch [3/3], Step [2392/3236], Loss: 1.8627, Perplexity: 6.4414
Epoch [3/3], Step [2393/3236], Loss: 1.9432, Perplexity: 6.9809
Epoch [3/3], Step [2394/3236], Loss: 1.9865, Perplexity: 7.2903
Epoch [3/3], Step [2395/3236], Loss: 1.9653, Perplexity: 7.1369
Epoch [3/3], Step [2396/3236], Loss: 2.1941, Perplexity: 8.9722
Epoch [3/3], Step [2397/3236], Loss: 2.7791, Perplexity: 16.1041
Epoch [3/3], Step [2398/3236], Loss: 1.8709, Perplexity: 6.4942
Epoch [3/3], Step [2399/3236], Loss: 1.8517, Perplexity: 6.3706
Epoch [3/3], Step [2400/3236], Loss: 1.8579, Perplexity: 6.4104
Epoch [3/3], Step [2401/3236], Loss: 2.1289, Perplexity: 8.4057
Epoch [3/3], Step [2402/3236], Loss: 1.8814, Perplexity: 6.5629
Epoch [3/3], Step [2403/3236], Loss: 1.9860, Perplexity: 7.2863
Epoch [3/3], Step [2404/3236], Loss: 2.0996, Perplexity: 8.1629
Epoch [3/3], Step [2405/3236], Loss: 1.9735, Perplexity: 7.1957
Epoch [3/3], Step [2406/3236], Loss: 1.8408, Perplexity: 6.3014
Epoch [3/3], Step [2407/3236], Loss: 2.0904, Perplexity: 8.0885
Epoch [3/3], Step [2408/3236], Loss: 1.9069, Perplexity: 6.7319
Epoch [3/3], Step [2409/3236], Loss: 1.8767, Perplexity: 6.5318
Epoch [3/3], Step [2410/3236], Loss: 1.8665, Perplexity: 6.4659
Epoch [3/3], Step [2411/3236], Loss: 1.9285, Perplexity: 6.8790
Epoch [3/3], Step [2412/3236], Loss: 1.8879, Perplexity: 6.6055
Epoch [3/3], Step [2413/3236], Loss: 1.9363, Perplexity: 6.9329
Epoch [3/3], Step [2414/3236], Loss: 1.8696, Perplexity: 6.4860
Epoch [3/3], Step [2415/3236], Loss: 1.9605, Perplexity: 7.1032
Epoch [3/3], Step [2416/3236], Loss: 2.0592, Perplexity: 7.8399
Epoch [3/3], Step [2417/3236], Loss: 2.0048, Perplexity: 7.4249
Epoch [3/3], Step [2418/3236], Loss: 1.9120, Perplexity: 6.7668
Epoch [3/3], Step [2419/3236], Loss: 2.1535, Perplexity: 8.6149
Epoch [3/3], Step [2420/3236], Loss: 1.9880, Perplexity: 7.3011
Epoch [3/3], Step [2421/3236], Loss: 1.8108, Perplexity: 6.1154
Epoch [3/3], Step [2422/3236], Loss: 2.2612, Perplexity: 9.5943
Epoch [3/3], Step [2423/3236], Loss: 1.8197, Perplexity: 6.1701
Epoch [3/3], Step [2424/3236], Loss: 1.9267, Perplexity: 6.8668
Epoch [3/3], Step [2425/3236], Loss: 1.9350, Perplexity: 6.9238
Epoch [3/3], Step [2426/3236], Loss: 2.0866, Perplexity: 8.0573
Epoch [3/3], Step [2427/3236], Loss: 1.8874, Perplexity: 6.6024
Epoch [3/3], Step [2428/3236], Loss: 1.9638, Perplexity: 7.1264
Epoch [3/3], Step [2429/3236], Loss: 2.3143, Perplexity: 10.1174
Epoch [3/3], Step [2430/3236], Loss: 1.8567, Perplexity: 6.4024
Epoch [3/3], Step [2431/3236], Loss: 1.8374, Perplexity: 6.2800
Epoch [3/3], Step [2432/3236], Loss: 1.9033, Perplexity: 6.7081
Epoch [3/3], Step [2433/3236], Loss: 2.1032, Perplexity: 8.1922
Epoch [3/3], Step [2434/3236], Loss: 2.1460, Perplexity: 8.5503
Epoch [3/3], Step [2435/3236], Loss: 1.8603, Perplexity: 6.4255
Epoch [3/3], Step [2436/3236], Loss: 1.8716, Perplexity: 6.4990
Epoch [3/3], Step [2437/3236], Loss: 1.8802, Perplexity: 6.5550
Epoch [3/3], Step [2438/3236], Loss: 2.1772, Perplexity: 8.8214
Epoch [3/3], Step [2439/3236], Loss: 1.9774, Perplexity: 7.2237
Epoch [3/3], Step [2440/3236], Loss: 1.8587, Perplexity: 6.4153
Epoch [3/3], Step [2441/3236], Loss: 1.9963, Perplexity: 7.3615
Epoch [3/3], Step [2442/3236], Loss: 1.9079, Perplexity: 6.7387
Epoch [3/3], Step [2443/3236], Loss: 1.9073, Perplexity: 6.7347
Epoch [3/3], Step [2444/3236], Loss: 1.7886, Perplexity: 5.9812
Epoch [3/3], Step [2445/3236], Loss: 1.8310, Perplexity: 6.2402
Epoch [3/3], Step [2446/3236], Loss: 1.9399, Perplexity: 6.9583
Epoch [3/3], Step [2447/3236], Loss: 1.8633, Perplexity: 6.4448
Epoch [3/3], Step [2448/3236], Loss: 1.9894, Perplexity: 7.3113
Epoch [3/3], Step [2449/3236], Loss: 2.0052, Perplexity: 7.4277
Epoch [3/3], Step [2450/3236], Loss: 1.8492, Perplexity: 6.3547
Epoch [3/3], Step [2451/3236], Loss: 2.0539, Perplexity: 7.7980
Epoch [3/3], Step [2452/3236], Loss: 1.9007, Perplexity: 6.6905
Epoch [3/3], Step [2453/3236], Loss: 1.9085, Perplexity: 6.7429
Epoch [3/3], Step [2454/3236], Loss: 1.9719, Perplexity: 7.1842
Epoch [3/3], Step [2455/3236], Loss: 1.8709, Perplexity: 6.4941
Epoch [3/3], Step [2456/3236], Loss: 2.2591, Perplexity: 9.5747
Epoch [3/3], Step [2457/3236], Loss: 2.0305, Perplexity: 7.6182
Epoch [3/3], Step [2458/3236], Loss: 2.7383, Perplexity: 15.4599
Epoch [3/3], Step [2459/3236], Loss: 1.8740, Perplexity: 6.5143
Epoch [3/3], Step [2460/3236], Loss: 2.3842, Perplexity: 10.8506
Epoch [3/3], Step [2461/3236], Loss: 2.1803, Perplexity: 8.8494
Epoch [3/3], Step [2462/3236], Loss: 2.0219, Perplexity: 7.5529
Epoch [3/3], Step [2463/3236], Loss: 2.0537, Perplexity: 7.7967
Epoch [3/3], Step [2464/3236], Loss: 1.8324, Perplexity: 6.2490
Epoch [3/3], Step [2465/3236], Loss: 2.1705, Perplexity: 8.7623
Epoch [3/3], Step [2466/3236], Loss: 1.9074, Perplexity: 6.7358
Epoch [3/3], Step [2467/3236], Loss: 2.0561, Perplexity: 7.8158
Epoch [3/3], Step [2468/3236], Loss: 1.9418, Perplexity: 6.9712
Epoch [3/3], Step [2469/3236], Loss: 2.0941, Perplexity: 8.1183
Epoch [3/3], Step [2470/3236], Loss: 2.2054, Perplexity: 9.0737
Epoch [3/3], Step [2471/3236], Loss: 1.9546, Perplexity: 7.0614
Epoch [3/3], Step [2472/3236], Loss: 1.8849, Perplexity: 6.5858
Epoch [3/3], Step [2473/3236], Loss: 1.9378, Perplexity: 6.9436
Epoch [3/3], Step [2474/3236], Loss: 2.5791, Perplexity: 13.1849
Epoch [3/3], Step [2475/3236], Loss: 1.8900, Perplexity: 6.6192
Epoch [3/3], Step [2476/3236], Loss: 1.9089, Perplexity: 6.7456
Epoch [3/3], Step [2477/3236], Loss: 1.8825, Perplexity: 6.5696
Epoch [3/3], Step [2478/3236], Loss: 1.9519, Perplexity: 7.0423
Epoch [3/3], Step [2479/3236], Loss: 2.1398, Perplexity: 8.4975
Epoch [3/3], Step [2480/3236], Loss: 1.9920, Perplexity: 7.3303
Epoch [3/3], Step [2481/3236], Loss: 1.9253, Perplexity: 6.8571
Epoch [3/3], Step [2482/3236], Loss: 2.0382, Perplexity: 7.6766
Epoch [3/3], Step [2483/3236], Loss: 1.9033, Perplexity: 6.7077
Epoch [3/3], Step [2484/3236], Loss: 2.0608, Perplexity: 7.8520
Epoch [3/3], Step [2485/3236], Loss: 2.0240, Perplexity: 7.5689
Epoch [3/3], Step [2486/3236], Loss: 2.2029, Perplexity: 9.0513
Epoch [3/3], Step [2487/3236], Loss: 1.8951, Perplexity: 6.6531
Epoch [3/3], Step [2488/3236], Loss: 1.9538, Perplexity: 7.0553
Epoch [3/3], Step [2489/3236], Loss: 1.9392, Perplexity: 6.9532
Epoch [3/3], Step [2490/3236], Loss: 2.4904, Perplexity: 12.0658
Epoch [3/3], Step [2491/3236], Loss: 2.0166, Perplexity: 7.5129
Epoch [3/3], Step [2492/3236], Loss: 2.1720, Perplexity: 8.7759
Epoch [3/3], Step [2493/3236], Loss: 1.9190, Perplexity: 6.8139
Epoch [3/3], Step [2494/3236], Loss: 2.1730, Perplexity: 8.7845
Epoch [3/3], Step [2495/3236], Loss: 2.0651, Perplexity: 7.8858
Epoch [3/3], Step [2496/3236], Loss: 1.9246, Perplexity: 6.8522
Epoch [3/3], Step [2497/3236], Loss: 1.8722, Perplexity: 6.5023
Epoch [3/3], Step [2498/3236], Loss: 1.9928, Perplexity: 7.3362
Epoch [3/3], Step [2499/3236], Loss: 2.0226, Perplexity: 7.5581
Epoch [3/3], Step [2500/3236], Loss: 1.9863, Perplexity: 7.2889
Epoch [3/3], Step [2501/3236], Loss: 2.4889, Perplexity: 12.0480
Epoch [3/3], Step [2502/3236], Loss: 1.8801, Perplexity: 6.5544
Epoch [3/3], Step [2503/3236], Loss: 1.9170, Perplexity: 6.8005
Epoch [3/3], Step [2504/3236], Loss: 2.0291, Perplexity: 7.6074
Epoch [3/3], Step [2505/3236], Loss: 1.9834, Perplexity: 7.2672
Epoch [3/3], Step [2506/3236], Loss: 1.8173, Perplexity: 6.1553
Epoch [3/3], Step [2507/3236], Loss: 1.9616, Perplexity: 7.1107
Epoch [3/3], Step [2508/3236], Loss: 1.7796, Perplexity: 5.9273
Epoch [3/3], Step [2509/3236], Loss: 2.1327, Perplexity: 8.4376
Epoch [3/3], Step [2510/3236], Loss: 1.9119, Perplexity: 6.7662
Epoch [3/3], Step [2511/3236], Loss: 1.8378, Perplexity: 6.2830
Epoch [3/3], Step [2512/3236], Loss: 1.8925, Perplexity: 6.6359
Epoch [3/3], Step [2513/3236], Loss: 1.7849, Perplexity: 5.9589
Epoch [3/3], Step [2514/3236], Loss: 1.8163, Perplexity: 6.1494
Epoch [3/3], Step [2515/3236], Loss: 2.0375, Perplexity: 7.6711
Epoch [3/3], Step [2516/3236], Loss: 1.8969, Perplexity: 6.6655
Epoch [3/3], Step [2517/3236], Loss: 2.4638, Perplexity: 11.7491
Epoch [3/3], Step [2518/3236], Loss: 1.9942, Perplexity: 7.3463
Epoch [3/3], Step [2519/3236], Loss: 2.0678, Perplexity: 7.9071
Epoch [3/3], Step [2520/3236], Loss: 1.8921, Perplexity: 6.6332
Epoch [3/3], Step [2521/3236], Loss: 1.8125, Perplexity: 6.1259
Epoch [3/3], Step [2522/3236], Loss: 1.9192, Perplexity: 6.8153
Epoch [3/3], Step [2523/3236], Loss: 1.8146, Perplexity: 6.1385
Epoch [3/3], Step [2524/3236], Loss: 2.0832, Perplexity: 8.0301
Epoch [3/3], Step [2525/3236], Loss: 2.2275, Perplexity: 9.2762
Epoch [3/3], Step [2526/3236], Loss: 1.7298, Perplexity: 5.6397
Epoch [3/3], Step [2527/3236], Loss: 1.8861, Perplexity: 6.5933
Epoch [3/3], Step [2528/3236], Loss: 2.1074, Perplexity: 8.2265
Epoch [3/3], Step [2529/3236], Loss: 1.9431, Perplexity: 6.9804
Epoch [3/3], Step [2530/3236], Loss: 2.0762, Perplexity: 7.9744
Epoch [3/3], Step [2531/3236], Loss: 1.8894, Perplexity: 6.6156
Epoch [3/3], Step [2532/3236], Loss: 1.7460, Perplexity: 5.7318
Epoch [3/3], Step [2533/3236], Loss: 1.9732, Perplexity: 7.1934
Epoch [3/3], Step [2534/3236], Loss: 1.8657, Perplexity: 6.4604
Epoch [3/3], Step [2535/3236], Loss: 1.8241, Perplexity: 6.1974
Epoch [3/3], Step [2536/3236], Loss: 2.0434, Perplexity: 7.7165
Epoch [3/3], Step [2537/3236], Loss: 1.9066, Perplexity: 6.7303
Epoch [3/3], Step [2538/3236], Loss: 3.1966, Perplexity: 24.4493
Epoch [3/3], Step [2539/3236], Loss: 1.7721, Perplexity: 5.8834
Epoch [3/3], Step [2540/3236], Loss: 1.9423, Perplexity: 6.9745
Epoch [3/3], Step [2541/3236], Loss: 1.9124, Perplexity: 6.7693
Epoch [3/3], Step [2542/3236], Loss: 1.9021, Perplexity: 6.7003
Epoch [3/3], Step [2543/3236], Loss: 1.8544, Perplexity: 6.3879
Epoch [3/3], Step [2544/3236], Loss: 2.0098, Perplexity: 7.4615
Epoch [3/3], Step [2545/3236], Loss: 1.8475, Perplexity: 6.3442
Epoch [3/3], Step [2546/3236], Loss: 1.9221, Perplexity: 6.8351
Epoch [3/3], Step [2547/3236], Loss: 2.3718, Perplexity: 10.7163
Epoch [3/3], Step [2548/3236], Loss: 1.8301, Perplexity: 6.2344
Epoch [3/3], Step [2549/3236], Loss: 1.8199, Perplexity: 6.1710
Epoch [3/3], Step [2550/3236], Loss: 2.1995, Perplexity: 9.0206
Epoch [3/3], Step [2551/3236], Loss: 1.9535, Perplexity: 7.0533
Epoch [3/3], Step [2552/3236], Loss: 1.9062, Perplexity: 6.7277
Epoch [3/3], Step [2553/3236], Loss: 1.8354, Perplexity: 6.2677
Epoch [3/3], Step [2554/3236], Loss: 1.9863, Perplexity: 7.2886
Epoch [3/3], Step [2555/3236], Loss: 1.9404, Perplexity: 6.9614
Epoch [3/3], Step [2556/3236], Loss: 2.0590, Perplexity: 7.8380
Epoch [3/3], Step [2557/3236], Loss: 1.8699, Perplexity: 6.4873
Epoch [3/3], Step [2558/3236], Loss: 2.0716, Perplexity: 7.9375
Epoch [3/3], Step [2559/3236], Loss: 2.0319, Perplexity: 7.6282
Epoch [3/3], Step [2560/3236], Loss: 1.8891, Perplexity: 6.6136
Epoch [3/3], Step [2561/3236], Loss: 2.1273, Perplexity: 8.3918
Epoch [3/3], Step [2562/3236], Loss: 2.1034, Perplexity: 8.1936
Epoch [3/3], Step [2563/3236], Loss: 2.2079, Perplexity: 9.0966
Epoch [3/3], Step [2564/3236], Loss: 1.8202, Perplexity: 6.1729
Epoch [3/3], Step [2565/3236], Loss: 2.0698, Perplexity: 7.9233
Epoch [3/3], Step [2566/3236], Loss: 1.8958, Perplexity: 6.6578
Epoch [3/3], Step [2567/3236], Loss: 1.9828, Perplexity: 7.2631
Epoch [3/3], Step [2568/3236], Loss: 1.9134, Perplexity: 6.7763
Epoch [3/3], Step [2569/3236], Loss: 1.8857, Perplexity: 6.5910
Epoch [3/3], Step [2570/3236], Loss: 1.9280, Perplexity: 6.8759
Epoch [3/3], Step [2571/3236], Loss: 2.6620, Perplexity: 14.3243
Epoch [3/3], Step [2572/3236], Loss: 1.8726, Perplexity: 6.5050
Epoch [3/3], Step [2573/3236], Loss: 1.7574, Perplexity: 5.7976
Epoch [3/3], Step [2574/3236], Loss: 2.2778, Perplexity: 9.7553
Epoch [3/3], Step [2575/3236], Loss: 2.7613, Perplexity: 15.8204
Epoch [3/3], Step [2576/3236], Loss: 1.8771, Perplexity: 6.5347
Epoch [3/3], Step [2577/3236], Loss: 2.5564, Perplexity: 12.8890
Epoch [3/3], Step [2578/3236], Loss: 2.0408, Perplexity: 7.6971
Epoch [3/3], Step [2579/3236], Loss: 1.9012, Perplexity: 6.6938
Epoch [3/3], Step [2580/3236], Loss: 1.9289, Perplexity: 6.8818
Epoch [3/3], Step [2581/3236], Loss: 1.9320, Perplexity: 6.9031
Epoch [3/3], Step [2582/3236], Loss: 1.8597, Perplexity: 6.4217
Epoch [3/3], Step [2583/3236], Loss: 2.0077, Perplexity: 7.4462
Epoch [3/3], Step [2584/3236], Loss: 1.9532, Perplexity: 7.0513
Epoch [3/3], Step [2585/3236], Loss: 2.0237, Perplexity: 7.5662
Epoch [3/3], Step [2586/3236], Loss: 1.8451, Perplexity: 6.3286
Epoch [3/3], Step [2587/3236], Loss: 2.0067, Perplexity: 7.4387
Epoch [3/3], Step [2588/3236], Loss: 1.9704, Perplexity: 7.1739
Epoch [3/3], Step [2589/3236], Loss: 2.0152, Perplexity: 7.5019
Epoch [3/3], Step [2590/3236], Loss: 1.9698, Perplexity: 7.1693
Epoch [3/3], Step [2591/3236], Loss: 1.8880, Perplexity: 6.6063
Epoch [3/3], Step [2592/3236], Loss: 1.9426, Perplexity: 6.9771
Epoch [3/3], Step [2593/3236], Loss: 1.9721, Perplexity: 7.1854
Epoch [3/3], Step [2594/3236], Loss: 1.8211, Perplexity: 6.1784
Epoch [3/3], Step [2595/3236], Loss: 1.8481, Perplexity: 6.3477
Epoch [3/3], Step [2596/3236], Loss: 1.8562, Perplexity: 6.3992
Epoch [3/3], Step [2597/3236], Loss: 1.8825, Perplexity: 6.5700
Epoch [3/3], Step [2598/3236], Loss: 1.9032, Perplexity: 6.7071
Epoch [3/3], Step [2599/3236], Loss: 1.9293, Perplexity: 6.8846
Epoch [3/3], Step [2600/3236], Loss: 2.0237, Perplexity: 7.5663
Epoch [3/3], Step [2601/3236], Loss: 1.9184, Perplexity: 6.8103
Epoch [3/3], Step [2602/3236], Loss: 2.6296, Perplexity: 13.8685
Epoch [3/3], Step [2603/3236], Loss: 1.8884, Perplexity: 6.6090
Epoch [3/3], Step [2604/3236], Loss: 2.2368, Perplexity: 9.3634
Epoch [3/3], Step [2605/3236], Loss: 2.0038, Perplexity: 7.4175
Epoch [3/3], Step [2606/3236], Loss: 1.8986, Perplexity: 6.6766
Epoch [3/3], Step [2607/3236], Loss: 1.9210, Perplexity: 6.8280
Epoch [3/3], Step [2608/3236], Loss: 2.1672, Perplexity: 8.7338
Epoch [3/3], Step [2609/3236], Loss: 1.9702, Perplexity: 7.1722
Epoch [3/3], Step [2610/3236], Loss: 1.7915, Perplexity: 5.9985
Epoch [3/3], Step [2611/3236], Loss: 1.9697, Perplexity: 7.1687
Epoch [3/3], Step [2612/3236], Loss: 1.7925, Perplexity: 6.0042
Epoch [3/3], Step [2613/3236], Loss: 1.7626, Perplexity: 5.8275
Epoch [3/3], Step [2614/3236], Loss: 1.9565, Perplexity: 7.0744
Epoch [3/3], Step [2615/3236], Loss: 1.8419, Perplexity: 6.3087
Epoch [3/3], Step [2616/3236], Loss: 2.9269, Perplexity: 18.6706
Epoch [3/3], Step [2617/3236], Loss: 1.9550, Perplexity: 7.0638
Epoch [3/3], Step [2618/3236], Loss: 2.0970, Perplexity: 8.1420
Epoch [3/3], Step [2619/3236], Loss: 1.7988, Perplexity: 6.0422
Epoch [3/3], Step [2620/3236], Loss: 1.9710, Perplexity: 7.1776
Epoch [3/3], Step [2621/3236], Loss: 1.9870, Perplexity: 7.2936
Epoch [3/3], Step [2622/3236], Loss: 2.2374, Perplexity: 9.3687
Epoch [3/3], Step [2623/3236], Loss: 1.9336, Perplexity: 6.9141
Epoch [3/3], Step [2624/3236], Loss: 2.0793, Perplexity: 7.9992
Epoch [3/3], Step [2625/3236], Loss: 2.1570, Perplexity: 8.6448
Epoch [3/3], Step [2626/3236], Loss: 1.9067, Perplexity: 6.7308
Epoch [3/3], Step [2627/3236], Loss: 1.8216, Perplexity: 6.1818
Epoch [3/3], Step [2628/3236], Loss: 2.2789, Perplexity: 9.7664
Epoch [3/3], Step [2629/3236], Loss: 1.8415, Perplexity: 6.3057
Epoch [3/3], Step [2630/3236], Loss: 1.9153, Perplexity: 6.7892
Epoch [3/3], Step [2631/3236], Loss: 1.8580, Perplexity: 6.4107
Epoch [3/3], Step [2632/3236], Loss: 1.9531, Perplexity: 7.0507
Epoch [3/3], Step [2633/3236], Loss: 2.0390, Perplexity: 7.6833
Epoch [3/3], Step [2634/3236], Loss: 1.9998, Perplexity: 7.3878
Epoch [3/3], Step [2635/3236], Loss: 2.0278, Perplexity: 7.5976
Epoch [3/3], Step [2636/3236], Loss: 2.0763, Perplexity: 7.9752
Epoch [3/3], Step [2637/3236], Loss: 1.9115, Perplexity: 6.7630
Epoch [3/3], Step [2638/3236], Loss: 2.0228, Perplexity: 7.5593
Epoch [3/3], Step [2639/3236], Loss: 1.9539, Perplexity: 7.0564
Epoch [3/3], Step [2640/3236], Loss: 1.9118, Perplexity: 6.7650
Epoch [3/3], Step [2641/3236], Loss: 1.8270, Perplexity: 6.2152
Epoch [3/3], Step [2642/3236], Loss: 2.0544, Perplexity: 7.8022
Epoch [3/3], Step [2643/3236], Loss: 1.8890, Perplexity: 6.6128
Epoch [3/3], Step [2644/3236], Loss: 1.8172, Perplexity: 6.1545
Epoch [3/3], Step [2645/3236], Loss: 1.8042, Perplexity: 6.0752
Epoch [3/3], Step [2646/3236], Loss: 2.0059, Perplexity: 7.4330
Epoch [3/3], Step [2647/3236], Loss: 1.8493, Perplexity: 6.3555
Epoch [3/3], Step [2648/3236], Loss: 1.9220, Perplexity: 6.8344
Epoch [3/3], Step [2649/3236], Loss: 1.9911, Perplexity: 7.3235
Epoch [3/3], Step [2650/3236], Loss: 2.2348, Perplexity: 9.3450
Epoch [3/3], Step [2651/3236], Loss: 1.8328, Perplexity: 6.2516
Epoch [3/3], Step [2652/3236], Loss: 2.0343, Perplexity: 7.6471
Epoch [3/3], Step [2653/3236], Loss: 2.6684, Perplexity: 14.4170
Epoch [3/3], Step [2654/3236], Loss: 1.8820, Perplexity: 6.5663
Epoch [3/3], Step [2655/3236], Loss: 1.9216, Perplexity: 6.8316
Epoch [3/3], Step [2656/3236], Loss: 2.3516, Perplexity: 10.5028
Epoch [3/3], Step [2657/3236], Loss: 1.7657, Perplexity: 5.8456
Epoch [3/3], Step [2658/3236], Loss: 1.9538, Perplexity: 7.0557
Epoch [3/3], Step [2659/3236], Loss: 1.9218, Perplexity: 6.8335
Epoch [3/3], Step [2660/3236], Loss: 1.9313, Perplexity: 6.8987
Epoch [3/3], Step [2661/3236], Loss: 1.7884, Perplexity: 5.9801
Epoch [3/3], Step [2662/3236], Loss: 1.9654, Perplexity: 7.1376
Epoch [3/3], Step [2663/3236], Loss: 1.9125, Perplexity: 6.7697
Epoch [3/3], Step [2664/3236], Loss: 2.1141, Perplexity: 8.2823
Epoch [3/3], Step [2665/3236], Loss: 1.9120, Perplexity: 6.7664
Epoch [3/3], Step [2666/3236], Loss: 1.9242, Perplexity: 6.8496
Epoch [3/3], Step [2667/3236], Loss: 1.7926, Perplexity: 6.0053
Epoch [3/3], Step [2668/3236], Loss: 2.1184, Perplexity: 8.3180
Epoch [3/3], Step [2669/3236], Loss: 2.2298, Perplexity: 9.2976
Epoch [3/3], Step [2670/3236], Loss: 1.9600, Perplexity: 7.0997
Epoch [3/3], Step [2671/3236], Loss: 2.2547, Perplexity: 9.5320
Epoch [3/3], Step [2672/3236], Loss: 2.0970, Perplexity: 8.1416
Epoch [3/3], Step [2673/3236], Loss: 2.1895, Perplexity: 8.9308
Epoch [3/3], Step [2674/3236], Loss: 1.9338, Perplexity: 6.9159
Epoch [3/3], Step [2675/3236], Loss: 1.8404, Perplexity: 6.2990
Epoch [3/3], Step [2676/3236], Loss: 1.9309, Perplexity: 6.8957
Epoch [3/3], Step [2677/3236], Loss: 1.9829, Perplexity: 7.2640
Epoch [3/3], Step [2678/3236], Loss: 2.3497, Perplexity: 10.4825
Epoch [3/3], Step [2679/3236], Loss: 2.0186, Perplexity: 7.5278
Epoch [3/3], Step [2680/3236], Loss: 2.1029, Perplexity: 8.1901
Epoch [3/3], Step [2681/3236], Loss: 1.8551, Perplexity: 6.3924
Epoch [3/3], Step [2682/3236], Loss: 1.9857, Perplexity: 7.2842
Epoch [3/3], Step [2683/3236], Loss: 1.9393, Perplexity: 6.9539
Epoch [3/3], Step [2684/3236], Loss: 1.7729, Perplexity: 5.8880
Epoch [3/3], Step [2685/3236], Loss: 2.0163, Perplexity: 7.5104
Epoch [3/3], Step [2686/3236], Loss: 1.6412, Perplexity: 5.1616
Epoch [3/3], Step [2687/3236], Loss: 2.2176, Perplexity: 9.1853
Epoch [3/3], Step [2688/3236], Loss: 1.8385, Perplexity: 6.2869
Epoch [3/3], Step [2689/3236], Loss: 1.9640, Perplexity: 7.1279
Epoch [3/3], Step [2690/3236], Loss: 1.9079, Perplexity: 6.7389
Epoch [3/3], Step [2691/3236], Loss: 2.4240, Perplexity: 11.2912
Epoch [3/3], Step [2692/3236], Loss: 2.5540, Perplexity: 12.8587
Epoch [3/3], Step [2693/3236], Loss: 1.8860, Perplexity: 6.5932
Epoch [3/3], Step [2694/3236], Loss: 1.8964, Perplexity: 6.6620
Epoch [3/3], Step [2695/3236], Loss: 1.8685, Perplexity: 6.4784
Epoch [3/3], Step [2696/3236], Loss: 1.8092, Perplexity: 6.1055
Epoch [3/3], Step [2697/3236], Loss: 1.9108, Perplexity: 6.7585
Epoch [3/3], Step [2698/3236], Loss: 1.8643, Perplexity: 6.4513
Epoch [3/3], Step [2699/3236], Loss: 1.9552, Perplexity: 7.0653
Epoch [3/3], Step [2700/3236], Loss: 2.1733, Perplexity: 8.7875
Epoch [3/3], Step [2701/3236], Loss: 2.2688, Perplexity: 9.6681
Epoch [3/3], Step [2702/3236], Loss: 1.9667, Perplexity: 7.1473
Epoch [3/3], Step [2703/3236], Loss: 1.8865, Perplexity: 6.5962
Epoch [3/3], Step [2704/3236], Loss: 1.8457, Perplexity: 6.3327
Epoch [3/3], Step [2705/3236], Loss: 2.6006, Perplexity: 13.4717
Epoch [3/3], Step [2706/3236], Loss: 1.8321, Perplexity: 6.2470
Epoch [3/3], Step [2707/3236], Loss: 1.9544, Perplexity: 7.0599
Epoch [3/3], Step [2708/3236], Loss: 2.0676, Perplexity: 7.9062
Epoch [3/3], Step [2709/3236], Loss: 2.1261, Perplexity: 8.3823
Epoch [3/3], Step [2710/3236], Loss: 2.0820, Perplexity: 8.0202
Epoch [3/3], Step [2711/3236], Loss: 1.9297, Perplexity: 6.8876
Epoch [3/3], Step [2712/3236], Loss: 1.9669, Perplexity: 7.1484
Epoch [3/3], Step [2713/3236], Loss: 1.9948, Perplexity: 7.3510
Epoch [3/3], Step [2714/3236], Loss: 2.0701, Perplexity: 7.9259
Epoch [3/3], Step [2715/3236], Loss: 1.8383, Perplexity: 6.2861
Epoch [3/3], Step [2716/3236], Loss: 1.9182, Perplexity: 6.8087
Epoch [3/3], Step [2717/3236], Loss: 2.2217, Perplexity: 9.2229
Epoch [3/3], Step [2718/3236], Loss: 1.8478, Perplexity: 6.3459
Epoch [3/3], Step [2719/3236], Loss: 1.9252, Perplexity: 6.8564
Epoch [3/3], Step [2720/3236], Loss: 1.9650, Perplexity: 7.1352
Epoch [3/3], Step [2721/3236], Loss: 1.8442, Perplexity: 6.3228
Epoch [3/3], Step [2722/3236], Loss: 1.9350, Perplexity: 6.9243
Epoch [3/3], Step [2723/3236], Loss: 1.7753, Perplexity: 5.9023
Epoch [3/3], Step [2724/3236], Loss: 1.9925, Perplexity: 7.3336
Epoch [3/3], Step [2725/3236], Loss: 1.9575, Perplexity: 7.0815
Epoch [3/3], Step [2726/3236], Loss: 1.8480, Perplexity: 6.3469
Epoch [3/3], Step [2727/3236], Loss: 1.9483, Perplexity: 7.0168
Epoch [3/3], Step [2728/3236], Loss: 1.9788, Perplexity: 7.2342
Epoch [3/3], Step [2729/3236], Loss: 1.8792, Perplexity: 6.5483
Epoch [3/3], Step [2730/3236], Loss: 1.8368, Perplexity: 6.2763
Epoch [3/3], Step [2731/3236], Loss: 1.8913, Perplexity: 6.6280
Epoch [3/3], Step [2732/3236], Loss: 1.9462, Perplexity: 7.0023
Epoch [3/3], Step [2733/3236], Loss: 2.1154, Perplexity: 8.2927
Epoch [3/3], Step [2734/3236], Loss: 1.9748, Perplexity: 7.2050
Epoch [3/3], Step [2735/3236], Loss: 1.9164, Perplexity: 6.7967
Epoch [3/3], Step [2736/3236], Loss: 1.9396, Perplexity: 6.9560
Epoch [3/3], Step [2737/3236], Loss: 1.9417, Perplexity: 6.9703
Epoch [3/3], Step [2738/3236], Loss: 1.9006, Perplexity: 6.6896
Epoch [3/3], Step [2739/3236], Loss: 2.0369, Perplexity: 7.6671
Epoch [3/3], Step [2740/3236], Loss: 1.9000, Perplexity: 6.6861
Epoch [3/3], Step [2741/3236], Loss: 1.9375, Perplexity: 6.9413
Epoch [3/3], Step [2742/3236], Loss: 1.9217, Perplexity: 6.8323
Epoch [3/3], Step [2743/3236], Loss: 1.9317, Perplexity: 6.9012
Epoch [3/3], Step [2744/3236], Loss: 1.9688, Perplexity: 7.1624
Epoch [3/3], Step [2745/3236], Loss: 1.8971, Perplexity: 6.6663
Epoch [3/3], Step [2746/3236], Loss: 1.8703, Perplexity: 6.4904
Epoch [3/3], Step [2747/3236], Loss: 1.9551, Perplexity: 7.0644
Epoch [3/3], Step [2748/3236], Loss: 1.7991, Perplexity: 6.0443
Epoch [3/3], Step [2749/3236], Loss: 1.9871, Perplexity: 7.2941
Epoch [3/3], Step [2750/3236], Loss: 2.4973, Perplexity: 12.1494
Epoch [3/3], Step [2751/3236], Loss: 1.9419, Perplexity: 6.9716
Epoch [3/3], Step [2752/3236], Loss: 1.9154, Perplexity: 6.7895
Epoch [3/3], Step [2753/3236], Loss: 1.9040, Perplexity: 6.7129
Epoch [3/3], Step [2754/3236], Loss: 1.9284, Perplexity: 6.8783
Epoch [3/3], Step [2755/3236], Loss: 1.8722, Perplexity: 6.5026
Epoch [3/3], Step [2756/3236], Loss: 1.8686, Perplexity: 6.4795
Epoch [3/3], Step [2757/3236], Loss: 1.9549, Perplexity: 7.0633
Epoch [3/3], Step [2758/3236], Loss: 1.9598, Perplexity: 7.0977
Epoch [3/3], Step [2759/3236], Loss: 1.9740, Perplexity: 7.1992
Epoch [3/3], Step [2760/3236], Loss: 1.8661, Perplexity: 6.4631
Epoch [3/3], Step [2761/3236], Loss: 2.0095, Perplexity: 7.4593
Epoch [3/3], Step [2762/3236], Loss: 1.9609, Perplexity: 7.1055
Epoch [3/3], Step [2763/3236], Loss: 2.1958, Perplexity: 8.9869
Epoch [3/3], Step [2764/3236], Loss: 1.8802, Perplexity: 6.5545
Epoch [3/3], Step [2765/3236], Loss: 1.7784, Perplexity: 5.9203
Epoch [3/3], Step [2766/3236], Loss: 1.9545, Perplexity: 7.0603
Epoch [3/3], Step [2767/3236], Loss: 2.0543, Perplexity: 7.8017
Epoch [3/3], Step [2768/3236], Loss: 1.8106, Perplexity: 6.1138
Epoch [3/3], Step [2769/3236], Loss: 1.9913, Perplexity: 7.3248
Epoch [3/3], Step [2770/3236], Loss: 2.2861, Perplexity: 9.8364
Epoch [3/3], Step [2771/3236], Loss: 1.9648, Perplexity: 7.1337
Epoch [3/3], Step [2772/3236], Loss: 1.8597, Perplexity: 6.4218
Epoch [3/3], Step [2773/3236], Loss: 1.9348, Perplexity: 6.9228
Epoch [3/3], Step [2774/3236], Loss: 1.8450, Perplexity: 6.3278
Epoch [3/3], Step [2775/3236], Loss: 1.8212, Perplexity: 6.1795
Epoch [3/3], Step [2776/3236], Loss: 1.8762, Perplexity: 6.5285
Epoch [3/3], Step [2777/3236], Loss: 1.9194, Perplexity: 6.8170
Epoch [3/3], Step [2778/3236], Loss: 1.8317, Perplexity: 6.2443
Epoch [3/3], Step [2779/3236], Loss: 1.8741, Perplexity: 6.5148
Epoch [3/3], Step [2780/3236], Loss: 2.1750, Perplexity: 8.8022
Epoch [3/3], Step [2781/3236], Loss: 2.0643, Perplexity: 7.8800
Epoch [3/3], Step [2782/3236], Loss: 2.1181, Perplexity: 8.3151
Epoch [3/3], Step [2783/3236], Loss: 2.1147, Perplexity: 8.2874
Epoch [3/3], Step [2784/3236], Loss: 1.9748, Perplexity: 7.2052
Epoch [3/3], Step [2785/3236], Loss: 2.6551, Perplexity: 14.2260
Epoch [3/3], Step [2786/3236], Loss: 2.8277, Perplexity: 16.9066
Epoch [3/3], Step [2787/3236], Loss: 1.8978, Perplexity: 6.6714
Epoch [3/3], Step [2788/3236], Loss: 2.1876, Perplexity: 8.9140
Epoch [3/3], Step [2789/3236], Loss: 2.0978, Perplexity: 8.1482
Epoch [3/3], Step [2790/3236], Loss: 2.0528, Perplexity: 7.7897
Epoch [3/3], Step [2791/3236], Loss: 2.0718, Perplexity: 7.9391
Epoch [3/3], Step [2792/3236], Loss: 1.8376, Perplexity: 6.2814
Epoch [3/3], Step [2793/3236], Loss: 1.9799, Perplexity: 7.2419
Epoch [3/3], Step [2794/3236], Loss: 2.0687, Perplexity: 7.9148
Epoch [3/3], Step [2795/3236], Loss: 1.8469, Perplexity: 6.3399
Epoch [3/3], Step [2796/3236], Loss: 2.0714, Perplexity: 7.9356
Epoch [3/3], Step [2797/3236], Loss: 1.9493, Perplexity: 7.0240
Epoch [3/3], Step [2798/3236], Loss: 1.9466, Perplexity: 7.0050
Epoch [3/3], Step [2799/3236], Loss: 1.8668, Perplexity: 6.4676
Epoch [3/3], Step [2800/3236], Loss: 1.8988, Perplexity: 6.6781
Epoch [3/3], Step [2801/3236], Loss: 2.0063, Perplexity: 7.4360
Epoch [3/3], Step [2802/3236], Loss: 1.9928, Perplexity: 7.3361
Epoch [3/3], Step [2803/3236], Loss: 2.0115, Perplexity: 7.4747
Epoch [3/3], Step [2804/3236], Loss: 2.0663, Perplexity: 7.8954
Epoch [3/3], Step [2805/3236], Loss: 2.0881, Perplexity: 8.0699
Epoch [3/3], Step [2806/3236], Loss: 1.9145, Perplexity: 6.7839
Epoch [3/3], Step [2807/3236], Loss: 1.9359, Perplexity: 6.9303
Epoch [3/3], Step [2808/3236], Loss: 2.2388, Perplexity: 9.3817
Epoch [3/3], Step [2809/3236], Loss: 1.9227, Perplexity: 6.8395
Epoch [3/3], Step [2810/3236], Loss: 2.1344, Perplexity: 8.4521
Epoch [3/3], Step [2811/3236], Loss: 1.8122, Perplexity: 6.1236
Epoch [3/3], Step [2812/3236], Loss: 1.8659, Perplexity: 6.4617
Epoch [3/3], Step [2813/3236], Loss: 1.7976, Perplexity: 6.0349
Epoch [3/3], Step [2814/3236], Loss: 1.8943, Perplexity: 6.6478
Epoch [3/3], Step [2815/3236], Loss: 2.2345, Perplexity: 9.3414
Epoch [3/3], Step [2816/3236], Loss: 1.9695, Perplexity: 7.1671
Epoch [3/3], Step [2817/3236], Loss: 1.7565, Perplexity: 5.7923
Epoch [3/3], Step [2818/3236], Loss: 1.8100, Perplexity: 6.1103
Epoch [3/3], Step [2819/3236], Loss: 1.8241, Perplexity: 6.1970
Epoch [3/3], Step [2820/3236], Loss: 1.8894, Perplexity: 6.6151
Epoch [3/3], Step [2821/3236], Loss: 1.7357, Perplexity: 5.6731
Epoch [3/3], Step [2822/3236], Loss: 1.8933, Perplexity: 6.6410
Epoch [3/3], Step [2823/3236], Loss: 1.9918, Perplexity: 7.3287
Epoch [3/3], Step [2824/3236], Loss: 1.8592, Perplexity: 6.4184
Epoch [3/3], Step [2825/3236], Loss: 1.9389, Perplexity: 6.9510
Epoch [3/3], Step [2826/3236], Loss: 2.5217, Perplexity: 12.4503
Epoch [3/3], Step [2827/3236], Loss: 1.8891, Perplexity: 6.6135
Epoch [3/3], Step [2828/3236], Loss: 1.8337, Perplexity: 6.2570
Epoch [3/3], Step [2829/3236], Loss: 1.9230, Perplexity: 6.8416
Epoch [3/3], Step [2830/3236], Loss: 2.0832, Perplexity: 8.0303
Epoch [3/3], Step [2831/3236], Loss: 1.9556, Perplexity: 7.0680
Epoch [3/3], Step [2832/3236], Loss: 1.9592, Perplexity: 7.0937
Epoch [3/3], Step [2833/3236], Loss: 1.9158, Perplexity: 6.7926
Epoch [3/3], Step [2834/3236], Loss: 2.0859, Perplexity: 8.0519
Epoch [3/3], Step [2835/3236], Loss: 2.0935, Perplexity: 8.1131
Epoch [3/3], Step [2836/3236], Loss: 1.8604, Perplexity: 6.4264
Epoch [3/3], Step [2837/3236], Loss: 1.8963, Perplexity: 6.6609
Epoch [3/3], Step [2838/3236], Loss: 1.8113, Perplexity: 6.1183
Epoch [3/3], Step [2839/3236], Loss: 1.9107, Perplexity: 6.7580
Epoch [3/3], Step [2840/3236], Loss: 1.9718, Perplexity: 7.1833
Epoch [3/3], Step [2841/3236], Loss: 2.0129, Perplexity: 7.4849
Epoch [3/3], Step [2842/3236], Loss: 2.7658, Perplexity: 15.8923
Epoch [3/3], Step [2843/3236], Loss: 1.9714, Perplexity: 7.1811
Epoch [3/3], Step [2844/3236], Loss: 1.7778, Perplexity: 5.9166
Epoch [3/3], Step [2845/3236], Loss: 1.7754, Perplexity: 5.9025
Epoch [3/3], Step [2846/3236], Loss: 2.1014, Perplexity: 8.1776
Epoch [3/3], Step [2847/3236], Loss: 2.0243, Perplexity: 7.5709
Epoch [3/3], Step [2848/3236], Loss: 2.4508, Perplexity: 11.5977
Epoch [3/3], Step [2849/3236], Loss: 2.0836, Perplexity: 8.0335
Epoch [3/3], Step [2850/3236], Loss: 1.9931, Perplexity: 7.3384
Epoch [3/3], Step [2851/3236], Loss: 1.8872, Perplexity: 6.6005
Epoch [3/3], Step [2852/3236], Loss: 1.8313, Perplexity: 6.2421
Epoch [3/3], Step [2853/3236], Loss: 2.1214, Perplexity: 8.3431
Epoch [3/3], Step [2854/3236], Loss: 1.8570, Perplexity: 6.4042
Epoch [3/3], Step [2855/3236], Loss: 1.9056, Perplexity: 6.7234
Epoch [3/3], Step [2856/3236], Loss: 2.2695, Perplexity: 9.6743
Epoch [3/3], Step [2857/3236], Loss: 1.8913, Perplexity: 6.6278
Epoch [3/3], Step [2858/3236], Loss: 2.1910, Perplexity: 8.9439
Epoch [3/3], Step [2859/3236], Loss: 1.8909, Perplexity: 6.6254
Epoch [3/3], Step [2860/3236], Loss: 2.0937, Perplexity: 8.1148
Epoch [3/3], Step [2861/3236], Loss: 1.8304, Perplexity: 6.2361
Epoch [3/3], Step [2862/3236], Loss: 1.8437, Perplexity: 6.3197
Epoch [3/3], Step [2863/3236], Loss: 2.1214, Perplexity: 8.3430
Epoch [3/3], Step [2864/3236], Loss: 1.9114, Perplexity: 6.7626
Epoch [3/3], Step [2865/3236], Loss: 1.9317, Perplexity: 6.9009
Epoch [3/3], Step [2866/3236], Loss: 2.0323, Perplexity: 7.6314
Epoch [3/3], Step [2867/3236], Loss: 1.8832, Perplexity: 6.5745
Epoch [3/3], Step [2868/3236], Loss: 2.2575, Perplexity: 9.5589
Epoch [3/3], Step [2869/3236], Loss: 1.9780, Perplexity: 7.2281
Epoch [3/3], Step [2870/3236], Loss: 2.5449, Perplexity: 12.7421
Epoch [3/3], Step [2871/3236], Loss: 1.9416, Perplexity: 6.9700
Epoch [3/3], Step [2872/3236], Loss: 1.8262, Perplexity: 6.2103
Epoch [3/3], Step [2873/3236], Loss: 1.9489, Perplexity: 7.0212
Epoch [3/3], Step [2874/3236], Loss: 1.9842, Perplexity: 7.2729
Epoch [3/3], Step [2875/3236], Loss: 1.7925, Perplexity: 6.0042
Epoch [3/3], Step [2876/3236], Loss: 1.9031, Perplexity: 6.7063
Epoch [3/3], Step [2877/3236], Loss: 1.8993, Perplexity: 6.6809
Epoch [3/3], Step [2878/3236], Loss: 1.9611, Perplexity: 7.1072
Epoch [3/3], Step [2879/3236], Loss: 3.1879, Perplexity: 24.2379
Epoch [3/3], Step [2880/3236], Loss: 1.8480, Perplexity: 6.3469
Epoch [3/3], Step [2881/3236], Loss: 2.1884, Perplexity: 8.9212
Epoch [3/3], Step [2882/3236], Loss: 1.9520, Perplexity: 7.0426
Epoch [3/3], Step [2883/3236], Loss: 1.9772, Perplexity: 7.2228
Epoch [3/3], Step [2884/3236], Loss: 1.8908, Perplexity: 6.6249
Epoch [3/3], Step [2885/3236], Loss: 1.9292, Perplexity: 6.8843
Epoch [3/3], Step [2886/3236], Loss: 2.2421, Perplexity: 9.4133
Epoch [3/3], Step [2887/3236], Loss: 1.8823, Perplexity: 6.5683
Epoch [3/3], Step [2888/3236], Loss: 1.9415, Perplexity: 6.9690
Epoch [3/3], Step [2889/3236], Loss: 1.8703, Perplexity: 6.4903
Epoch [3/3], Step [2890/3236], Loss: 2.7067, Perplexity: 14.9799
Epoch [3/3], Step [2891/3236], Loss: 1.9474, Perplexity: 7.0103
Epoch [3/3], Step [2892/3236], Loss: 1.8625, Perplexity: 6.4400
Epoch [3/3], Step [2893/3236], Loss: 2.2426, Perplexity: 9.4175
Epoch [3/3], Step [2894/3236], Loss: 1.8291, Perplexity: 6.2280
Epoch [3/3], Step [2895/3236], Loss: 1.9179, Perplexity: 6.8065
Epoch [3/3], Step [2896/3236], Loss: 2.0423, Perplexity: 7.7081
Epoch [3/3], Step [2897/3236], Loss: 1.9222, Perplexity: 6.8358
Epoch [3/3], Step [2898/3236], Loss: 2.1172, Perplexity: 8.3080
Epoch [3/3], Step [2899/3236], Loss: 2.1373, Perplexity: 8.4768
Epoch [3/3], Step [2900/3236], Loss: 1.9361, Perplexity: 6.9315
Epoch [3/3], Step [2901/3236], Loss: 1.9128, Perplexity: 6.7719
Epoch [3/3], Step [2902/3236], Loss: 1.8554, Perplexity: 6.3944
Epoch [3/3], Step [2903/3236], Loss: 1.9418, Perplexity: 6.9715
Epoch [3/3], Step [2904/3236], Loss: 2.1899, Perplexity: 8.9341
Epoch [3/3], Step [2905/3236], Loss: 1.9409, Perplexity: 6.9653
Epoch [3/3], Step [2906/3236], Loss: 1.9433, Perplexity: 6.9820
Epoch [3/3], Step [2907/3236], Loss: 2.1271, Perplexity: 8.3904
Epoch [3/3], Step [2908/3236], Loss: 1.8656, Perplexity: 6.4596
Epoch [3/3], Step [2909/3236], Loss: 2.1662, Perplexity: 8.7251
Epoch [3/3], Step [2910/3236], Loss: 1.9181, Perplexity: 6.8078
Epoch [3/3], Step [2911/3236], Loss: 1.9187, Perplexity: 6.8124
Epoch [3/3], Step [2912/3236], Loss: 1.8935, Perplexity: 6.6424
Epoch [3/3], Step [2913/3236], Loss: 1.7808, Perplexity: 5.9346
Epoch [3/3], Step [2914/3236], Loss: 2.0611, Perplexity: 7.8549
Epoch [3/3], Step [2915/3236], Loss: 1.9252, Perplexity: 6.8564
Epoch [3/3], Step [2916/3236], Loss: 1.9852, Perplexity: 7.2806
Epoch [3/3], Step [2917/3236], Loss: 1.8607, Perplexity: 6.4284
Epoch [3/3], Step [2918/3236], Loss: 1.8438, Perplexity: 6.3208
Epoch [3/3], Step [2919/3236], Loss: 1.8693, Perplexity: 6.4840
Epoch [3/3], Step [2920/3236], Loss: 1.9266, Perplexity: 6.8658
Epoch [3/3], Step [2921/3236], Loss: 2.2552, Perplexity: 9.5369
Epoch [3/3], Step [2922/3236], Loss: 2.1368, Perplexity: 8.4723
Epoch [3/3], Step [2923/3236], Loss: 1.8375, Perplexity: 6.2806
Epoch [3/3], Step [2924/3236], Loss: 1.8818, Perplexity: 6.5650
Epoch [3/3], Step [2925/3236], Loss: 1.8247, Perplexity: 6.2010
Epoch [3/3], Step [2926/3236], Loss: 1.8684, Perplexity: 6.4782
Epoch [3/3], Step [2927/3236], Loss: 1.7051, Perplexity: 5.5019
Epoch [3/3], Step [2928/3236], Loss: 1.8299, Perplexity: 6.2334
Epoch [3/3], Step [2929/3236], Loss: 2.0213, Perplexity: 7.5481
Epoch [3/3], Step [2930/3236], Loss: 1.8673, Perplexity: 6.4711
Epoch [3/3], Step [2931/3236], Loss: 1.9702, Perplexity: 7.1721
Epoch [3/3], Step [2932/3236], Loss: 1.9626, Perplexity: 7.1178
Epoch [3/3], Step [2933/3236], Loss: 1.8602, Perplexity: 6.4253
Epoch [3/3], Step [2934/3236], Loss: 1.7977, Perplexity: 6.0357
Epoch [3/3], Step [2935/3236], Loss: 2.2850, Perplexity: 9.8255
Epoch [3/3], Step [2936/3236], Loss: 1.9038, Perplexity: 6.7112
Epoch [3/3], Step [2937/3236], Loss: 1.8783, Perplexity: 6.5421
Epoch [3/3], Step [2938/3236], Loss: 1.8504, Perplexity: 6.3625
Epoch [3/3], Step [2939/3236], Loss: 1.9018, Perplexity: 6.6980
Epoch [3/3], Step [2940/3236], Loss: 1.8003, Perplexity: 6.0515
Epoch [3/3], Step [2941/3236], Loss: 1.8805, Perplexity: 6.5568
Epoch [3/3], Step [2942/3236], Loss: 1.8324, Perplexity: 6.2491
Epoch [3/3], Step [2943/3236], Loss: 1.9809, Perplexity: 7.2493
Epoch [3/3], Step [2944/3236], Loss: 1.9820, Perplexity: 7.2575
Epoch [3/3], Step [2945/3236], Loss: 5.0774, Perplexity: 160.3580
Epoch [3/3], Step [2946/3236], Loss: 1.9786, Perplexity: 7.2330
Epoch [3/3], Step [2947/3236], Loss: 2.0018, Perplexity: 7.4021
Epoch [3/3], Step [2948/3236], Loss: 1.8891, Perplexity: 6.6137
Epoch [3/3], Step [2949/3236], Loss: 1.8589, Perplexity: 6.4169
Epoch [3/3], Step [2950/3236], Loss: 1.9958, Perplexity: 7.3577
Epoch [3/3], Step [2951/3236], Loss: 2.1476, Perplexity: 8.5641
Epoch [3/3], Step [2952/3236], Loss: 2.1480, Perplexity: 8.5676
Epoch [3/3], Step [2953/3236], Loss: 2.0337, Perplexity: 7.6422
Epoch [3/3], Step [2954/3236], Loss: 2.3751, Perplexity: 10.7516
Epoch [3/3], Step [2955/3236], Loss: 1.9952, Perplexity: 7.3538
Epoch [3/3], Step [2956/3236], Loss: 1.8150, Perplexity: 6.1414
Epoch [3/3], Step [2957/3236], Loss: 1.9352, Perplexity: 6.9254
Epoch [3/3], Step [2958/3236], Loss: 1.9717, Perplexity: 7.1829
Epoch [3/3], Step [2959/3236], Loss: 1.9258, Perplexity: 6.8604
Epoch [3/3], Step [2960/3236], Loss: 2.0329, Perplexity: 7.6364
Epoch [3/3], Step [2961/3236], Loss: 1.8659, Perplexity: 6.4620
Epoch [3/3], Step [2962/3236], Loss: 2.0524, Perplexity: 7.7867
Epoch [3/3], Step [2963/3236], Loss: 1.9390, Perplexity: 6.9516
Epoch [3/3], Step [2964/3236], Loss: 1.7902, Perplexity: 5.9906
Epoch [3/3], Step [2965/3236], Loss: 1.8770, Perplexity: 6.5341
Epoch [3/3], Step [2966/3236], Loss: 2.2421, Perplexity: 9.4126
Epoch [3/3], Step [2967/3236], Loss: 1.9141, Perplexity: 6.7808
Epoch [3/3], Step [2968/3236], Loss: 2.2882, Perplexity: 9.8575
Epoch [3/3], Step [2969/3236], Loss: 1.8945, Perplexity: 6.6490
Epoch [3/3], Step [2970/3236], Loss: 1.9566, Perplexity: 7.0755
Epoch [3/3], Step [2971/3236], Loss: 1.8861, Perplexity: 6.5936
Epoch [3/3], Step [2972/3236], Loss: 2.0502, Perplexity: 7.7695
Epoch [3/3], Step [2973/3236], Loss: 1.9417, Perplexity: 6.9708
Epoch [3/3], Step [2974/3236], Loss: 2.0005, Perplexity: 7.3931
Epoch [3/3], Step [2975/3236], Loss: 1.8957, Perplexity: 6.6571
Epoch [3/3], Step [2976/3236], Loss: 2.0245, Perplexity: 7.5725
Epoch [3/3], Step [2977/3236], Loss: 1.9824, Perplexity: 7.2599
Epoch [3/3], Step [2978/3236], Loss: 2.2343, Perplexity: 9.3399
Epoch [3/3], Step [2979/3236], Loss: 2.0714, Perplexity: 7.9361
Epoch [3/3], Step [2980/3236], Loss: 1.8881, Perplexity: 6.6065
Epoch [3/3], Step [2981/3236], Loss: 2.2486, Perplexity: 9.4745
Epoch [3/3], Step [2982/3236], Loss: 1.9164, Perplexity: 6.7965
Epoch [3/3], Step [2983/3236], Loss: 1.8527, Perplexity: 6.3771
Epoch [3/3], Step [2984/3236], Loss: 1.8842, Perplexity: 6.5810
Epoch [3/3], Step [2985/3236], Loss: 2.2006, Perplexity: 9.0303
Epoch [3/3], Step [2986/3236], Loss: 2.1114, Perplexity: 8.2597
Epoch [3/3], Step [2987/3236], Loss: 1.8885, Perplexity: 6.6094
Epoch [3/3], Step [2988/3236], Loss: 2.0940, Perplexity: 8.1170
Epoch [3/3], Step [2989/3236], Loss: 1.9887, Perplexity: 7.3061
Epoch [3/3], Step [2990/3236], Loss: 1.8876, Perplexity: 6.6037
Epoch [3/3], Step [2991/3236], Loss: 1.8087, Perplexity: 6.1028
Epoch [3/3], Step [2992/3236], Loss: 2.2233, Perplexity: 9.2375
Epoch [3/3], Step [2993/3236], Loss: 2.1174, Perplexity: 8.3097
Epoch [3/3], Step [2994/3236], Loss: 1.9312, Perplexity: 6.8980
Epoch [3/3], Step [2995/3236], Loss: 2.1769, Perplexity: 8.8191
Epoch [3/3], Step [2996/3236], Loss: 1.8881, Perplexity: 6.6070
Epoch [3/3], Step [2997/3236], Loss: 2.0157, Perplexity: 7.5062
Epoch [3/3], Step [2998/3236], Loss: 2.1313, Perplexity: 8.4258
Epoch [3/3], Step [2999/3236], Loss: 1.9444, Perplexity: 6.9897
Epoch [3/3], Step [3000/3236], Loss: 1.8977, Perplexity: 6.6704
Epoch [3/3], Step [3001/3236], Loss: 2.0382, Perplexity: 7.6768
Epoch [3/3], Step [3002/3236], Loss: 1.7907, Perplexity: 5.9937
Epoch [3/3], Step [3003/3236], Loss: 1.8063, Perplexity: 6.0878
Epoch [3/3], Step [3004/3236], Loss: 2.2964, Perplexity: 9.9380
Epoch [3/3], Step [3005/3236], Loss: 1.9740, Perplexity: 7.1996
Epoch [3/3], Step [3006/3236], Loss: 1.9133, Perplexity: 6.7757
Epoch [3/3], Step [3007/3236], Loss: 1.9027, Perplexity: 6.7037
Epoch [3/3], Step [3008/3236], Loss: 1.8583, Perplexity: 6.4131
Epoch [3/3], Step [3009/3236], Loss: 1.7765, Perplexity: 5.9090
Epoch [3/3], Step [3010/3236], Loss: 1.9241, Perplexity: 6.8488
Epoch [3/3], Step [3011/3236], Loss: 2.2598, Perplexity: 9.5811
Epoch [3/3], Step [3012/3236], Loss: 1.9943, Perplexity: 7.3472
Epoch [3/3], Step [3013/3236], Loss: 1.8919, Perplexity: 6.6319
Epoch [3/3], Step [3014/3236], Loss: 2.1902, Perplexity: 8.9371
Epoch [3/3], Step [3015/3236], Loss: 1.9646, Perplexity: 7.1321
Epoch [3/3], Step [3016/3236], Loss: 1.9145, Perplexity: 6.7837
Epoch [3/3], Step [3017/3236], Loss: 2.0097, Perplexity: 7.4611
Epoch [3/3], Step [3018/3236], Loss: 2.1043, Perplexity: 8.2010
Epoch [3/3], Step [3019/3236], Loss: 2.1700, Perplexity: 8.7587
Epoch [3/3], Step [3020/3236], Loss: 1.8886, Perplexity: 6.6102
Epoch [3/3], Step [3021/3236], Loss: 1.8217, Perplexity: 6.1822
Epoch [3/3], Step [3022/3236], Loss: 2.0360, Perplexity: 7.6599
Epoch [3/3], Step [3023/3236], Loss: 1.8530, Perplexity: 6.3788
Epoch [3/3], Step [3024/3236], Loss: 1.8848, Perplexity: 6.5849
Epoch [3/3], Step [3025/3236], Loss: 1.9083, Perplexity: 6.7418
Epoch [3/3], Step [3026/3236], Loss: 2.6970, Perplexity: 14.8348
Epoch [3/3], Step [3027/3236], Loss: 2.0143, Perplexity: 7.4958
Epoch [3/3], Step [3028/3236], Loss: 2.0122, Perplexity: 7.4797
Epoch [3/3], Step [3029/3236], Loss: 1.8328, Perplexity: 6.2513
Epoch [3/3], Step [3030/3236], Loss: 1.9154, Perplexity: 6.7894
Epoch [3/3], Step [3031/3236], Loss: 1.8404, Perplexity: 6.2989
Epoch [3/3], Step [3032/3236], Loss: 1.8882, Perplexity: 6.6073
Epoch [3/3], Step [3033/3236], Loss: 1.7561, Perplexity: 5.7895
Epoch [3/3], Step [3034/3236], Loss: 1.8436, Perplexity: 6.3191
Epoch [3/3], Step [3035/3236], Loss: 1.9492, Perplexity: 7.0232
Epoch [3/3], Step [3036/3236], Loss: 1.8635, Perplexity: 6.4465
Epoch [3/3], Step [3037/3236], Loss: 1.9983, Perplexity: 7.3768
Epoch [3/3], Step [3038/3236], Loss: 1.7309, Perplexity: 5.6456
Epoch [3/3], Step [3039/3236], Loss: 1.9489, Perplexity: 7.0210
Epoch [3/3], Step [3040/3236], Loss: 1.9986, Perplexity: 7.3786
Epoch [3/3], Step [3041/3236], Loss: 1.9320, Perplexity: 6.9030
Epoch [3/3], Step [3042/3236], Loss: 1.9434, Perplexity: 6.9824
Epoch [3/3], Step [3043/3236], Loss: 2.1875, Perplexity: 8.9130
Epoch [3/3], Step [3044/3236], Loss: 1.8924, Perplexity: 6.6356
Epoch [3/3], Step [3045/3236], Loss: 2.4050, Perplexity: 11.0784
Epoch [3/3], Step [3046/3236], Loss: 1.9094, Perplexity: 6.7492
Epoch [3/3], Step [3047/3236], Loss: 1.9401, Perplexity: 6.9596
Epoch [3/3], Step [3048/3236], Loss: 1.8059, Perplexity: 6.0856
Epoch [3/3], Step [3049/3236], Loss: 1.8533, Perplexity: 6.3808
Epoch [3/3], Step [3050/3236], Loss: 1.9009, Perplexity: 6.6919
Epoch [3/3], Step [3051/3236], Loss: 1.7351, Perplexity: 5.6696
Epoch [3/3], Step [3052/3236], Loss: 1.9800, Perplexity: 7.2424
Epoch [3/3], Step [3053/3236], Loss: 2.0181, Perplexity: 7.5238
Epoch [3/3], Step [3054/3236], Loss: 2.0656, Perplexity: 7.8899
Epoch [3/3], Step [3055/3236], Loss: 1.9997, Perplexity: 7.3865
Epoch [3/3], Step [3056/3236], Loss: 1.9832, Perplexity: 7.2659
Epoch [3/3], Step [3057/3236], Loss: 1.8078, Perplexity: 6.0970
Epoch [3/3], Step [3058/3236], Loss: 1.8543, Perplexity: 6.3875
Epoch [3/3], Step [3059/3236], Loss: 1.8323, Perplexity: 6.2485
Epoch [3/3], Step [3060/3236], Loss: 1.8991, Perplexity: 6.6800
Epoch [3/3], Step [3061/3236], Loss: 1.7752, Perplexity: 5.9016
Epoch [3/3], Step [3062/3236], Loss: 1.9111, Perplexity: 6.7604
Epoch [3/3], Step [3063/3236], Loss: 2.1245, Perplexity: 8.3689
Epoch [3/3], Step [3064/3236], Loss: 1.9425, Perplexity: 6.9760
Epoch [3/3], Step [3065/3236], Loss: 1.9386, Perplexity: 6.9488
Epoch [3/3], Step [3066/3236], Loss: 2.1595, Perplexity: 8.6671
Epoch [3/3], Step [3067/3236], Loss: 2.1732, Perplexity: 8.7867
Epoch [3/3], Step [3068/3236], Loss: 1.8482, Perplexity: 6.3482
Epoch [3/3], Step [3069/3236], Loss: 1.9323, Perplexity: 6.9051
Epoch [3/3], Step [3070/3236], Loss: 1.8827, Perplexity: 6.5715
Epoch [3/3], Step [3071/3236], Loss: 1.9310, Perplexity: 6.8964
Epoch [3/3], Step [3072/3236], Loss: 1.9949, Perplexity: 7.3513
Epoch [3/3], Step [3073/3236], Loss: 1.7758, Perplexity: 5.9048
Epoch [3/3], Step [3074/3236], Loss: 1.8791, Perplexity: 6.5479
Epoch [3/3], Step [3075/3236], Loss: 1.9273, Perplexity: 6.8706
Epoch [3/3], Step [3076/3236], Loss: 1.9725, Perplexity: 7.1886
Epoch [3/3], Step [3077/3236], Loss: 1.7964, Perplexity: 6.0277
Epoch [3/3], Step [3078/3236], Loss: 1.9078, Perplexity: 6.7379
Epoch [3/3], Step [3079/3236], Loss: 1.9696, Perplexity: 7.1680
Epoch [3/3], Step [3080/3236], Loss: 2.1982, Perplexity: 9.0087
Epoch [3/3], Step [3081/3236], Loss: 1.9745, Perplexity: 7.2031
Epoch [3/3], Step [3082/3236], Loss: 2.0052, Perplexity: 7.4277
Epoch [3/3], Step [3083/3236], Loss: 1.8840, Perplexity: 6.5799
Epoch [3/3], Step [3084/3236], Loss: 2.0505, Perplexity: 7.7714
Epoch [3/3], Step [3085/3236], Loss: 1.9275, Perplexity: 6.8722
Epoch [3/3], Step [3086/3236], Loss: 1.7977, Perplexity: 6.0361
Epoch [3/3], Step [3087/3236], Loss: 2.0052, Perplexity: 7.4277
Epoch [3/3], Step [3088/3236], Loss: 1.9062, Perplexity: 6.7277
Epoch [3/3], Step [3089/3236], Loss: 1.8137, Perplexity: 6.1329
Epoch [3/3], Step [3090/3236], Loss: 2.2140, Perplexity: 9.1524
Epoch [3/3], Step [3091/3236], Loss: 1.8742, Perplexity: 6.5153
Epoch [3/3], Step [3092/3236], Loss: 1.9759, Perplexity: 7.2132
Epoch [3/3], Step [3093/3236], Loss: 1.9323, Perplexity: 6.9053
Epoch [3/3], Step [3094/3236], Loss: 2.2116, Perplexity: 9.1305
Epoch [3/3], Step [3095/3236], Loss: 1.8310, Perplexity: 6.2400
Epoch [3/3], Step [3096/3236], Loss: 1.9106, Perplexity: 6.7569
Epoch [3/3], Step [3097/3236], Loss: 1.8988, Perplexity: 6.6779
Epoch [3/3], Step [3098/3236], Loss: 2.1998, Perplexity: 9.0230
Epoch [3/3], Step [3099/3236], Loss: 1.8719, Perplexity: 6.5008
Epoch [3/3], Step [3100/3236], Loss: 1.7935, Perplexity: 6.0106
Epoch [3/3], Step [3101/3236], Loss: 1.8296, Perplexity: 6.2313
Epoch [3/3], Step [3102/3236], Loss: 1.9624, Perplexity: 7.1164
Epoch [3/3], Step [3103/3236], Loss: 2.2004, Perplexity: 9.0289
Epoch [3/3], Step [3104/3236], Loss: 2.2564, Perplexity: 9.5489
Epoch [3/3], Step [3105/3236], Loss: 2.6344, Perplexity: 13.9348
Epoch [3/3], Step [3106/3236], Loss: 1.7521, Perplexity: 5.7669
Epoch [3/3], Step [3107/3236], Loss: 1.8882, Perplexity: 6.6076
Epoch [3/3], Step [3108/3236], Loss: 2.2670, Perplexity: 9.6506
Epoch [3/3], Step [3109/3236], Loss: 1.7997, Perplexity: 6.0480
Epoch [3/3], Step [3110/3236], Loss: 1.8747, Perplexity: 6.5191
Epoch [3/3], Step [3111/3236], Loss: 1.9501, Perplexity: 7.0297
Epoch [3/3], Step [3112/3236], Loss: 1.8467, Perplexity: 6.3389
Epoch [3/3], Step [3113/3236], Loss: 1.8593, Perplexity: 6.4193
Epoch [3/3], Step [3114/3236], Loss: 1.9477, Perplexity: 7.0125
Epoch [3/3], Step [3115/3236], Loss: 1.9310, Perplexity: 6.8963
Epoch [3/3], Step [3116/3236], Loss: 1.7809, Perplexity: 5.9349
Epoch [3/3], Step [3117/3236], Loss: 1.8779, Perplexity: 6.5396
Epoch [3/3], Step [3118/3236], Loss: 1.9243, Perplexity: 6.8503
Epoch [3/3], Step [3119/3236], Loss: 2.5525, Perplexity: 12.8390
Epoch [3/3], Step [3120/3236], Loss: 1.7511, Perplexity: 5.7607
Epoch [3/3], Step [3121/3236], Loss: 2.4025, Perplexity: 11.0511
Epoch [3/3], Step [3122/3236], Loss: 1.8827, Perplexity: 6.5712
Epoch [3/3], Step [3123/3236], Loss: 2.1141, Perplexity: 8.2819
Epoch [3/3], Step [3124/3236], Loss: 2.1724, Perplexity: 8.7795
Epoch [3/3], Step [3125/3236], Loss: 1.8514, Perplexity: 6.3688
Epoch [3/3], Step [3126/3236], Loss: 1.8943, Perplexity: 6.6478
Epoch [3/3], Step [3127/3236], Loss: 1.9572, Perplexity: 7.0797
Epoch [3/3], Step [3128/3236], Loss: 2.0261, Perplexity: 7.5847
Epoch [3/3], Step [3129/3236], Loss: 1.9719, Perplexity: 7.1844
Epoch [3/3], Step [3130/3236], Loss: 1.8307, Perplexity: 6.2382
Epoch [3/3], Step [3131/3236], Loss: 1.9974, Perplexity: 7.3698
Epoch [3/3], Step [3132/3236], Loss: 2.1274, Perplexity: 8.3927
Epoch [3/3], Step [3133/3236], Loss: 1.9741, Perplexity: 7.2004
Epoch [3/3], Step [3134/3236], Loss: 2.0116, Perplexity: 7.4756
Epoch [3/3], Step [3135/3236], Loss: 2.1372, Perplexity: 8.4758
Epoch [3/3], Step [3136/3236], Loss: 1.8881, Perplexity: 6.6068
Epoch [3/3], Step [3137/3236], Loss: 1.8592, Perplexity: 6.4186
Epoch [3/3], Step [3138/3236], Loss: 1.9323, Perplexity: 6.9056
Epoch [3/3], Step [3139/3236], Loss: 2.1246, Perplexity: 8.3698
Epoch [3/3], Step [3140/3236], Loss: 1.7948, Perplexity: 6.0180
Epoch [3/3], Step [3141/3236], Loss: 2.0895, Perplexity: 8.0809
Epoch [3/3], Step [3142/3236], Loss: 1.8902, Perplexity: 6.6208
Epoch [3/3], Step [3143/3236], Loss: 1.8849, Perplexity: 6.5857
Epoch [3/3], Step [3144/3236], Loss: 1.8591, Perplexity: 6.4181
Epoch [3/3], Step [3145/3236], Loss: 1.8447, Perplexity: 6.3264
Epoch [3/3], Step [3146/3236], Loss: 1.8341, Perplexity: 6.2596
Epoch [3/3], Step [3147/3236], Loss: 1.9116, Perplexity: 6.7639
Epoch [3/3], Step [3148/3236], Loss: 1.9124, Perplexity: 6.7694
Epoch [3/3], Step [3149/3236], Loss: 2.0071, Perplexity: 7.4419
Epoch [3/3], Step [3150/3236], Loss: 2.1717, Perplexity: 8.7732
Epoch [3/3], Step [3151/3236], Loss: 1.9700, Perplexity: 7.1709
Epoch [3/3], Step [3152/3236], Loss: 1.7982, Perplexity: 6.0388
Epoch [3/3], Step [3153/3236], Loss: 1.8756, Perplexity: 6.5250
Epoch [3/3], Step [3154/3236], Loss: 2.1118, Perplexity: 8.2634
Epoch [3/3], Step [3155/3236], Loss: 2.1491, Perplexity: 8.5775
Epoch [3/3], Step [3156/3236], Loss: 1.9991, Perplexity: 7.3825
Epoch [3/3], Step [3157/3236], Loss: 1.9335, Perplexity: 6.9136
Epoch [3/3], Step [3158/3236], Loss: 1.8031, Perplexity: 6.0686
Epoch [3/3], Step [3159/3236], Loss: 1.8804, Perplexity: 6.5562
Epoch [3/3], Step [3160/3236], Loss: 1.8667, Perplexity: 6.4668
Epoch [3/3], Step [3161/3236], Loss: 1.9605, Perplexity: 7.1031
Epoch [3/3], Step [3162/3236], Loss: 1.8748, Perplexity: 6.5192
Epoch [3/3], Step [3163/3236], Loss: 1.9848, Perplexity: 7.2773
Epoch [3/3], Step [3164/3236], Loss: 1.9282, Perplexity: 6.8775
Epoch [3/3], Step [3165/3236], Loss: 2.0296, Perplexity: 7.6109
Epoch [3/3], Step [3166/3236], Loss: 2.1304, Perplexity: 8.4183
Epoch [3/3], Step [3167/3236], Loss: 2.1356, Perplexity: 8.4621
Epoch [3/3], Step [3168/3236], Loss: 1.9276, Perplexity: 6.8729
Epoch [3/3], Step [3169/3236], Loss: 1.8524, Perplexity: 6.3752
Epoch [3/3], Step [3170/3236], Loss: 2.0378, Perplexity: 7.6735
Epoch [3/3], Step [3171/3236], Loss: 1.8144, Perplexity: 6.1374
Epoch [3/3], Step [3172/3236], Loss: 1.9723, Perplexity: 7.1873
Epoch [3/3], Step [3173/3236], Loss: 2.0049, Perplexity: 7.4255
Epoch [3/3], Step [3174/3236], Loss: 1.9718, Perplexity: 7.1834
Epoch [3/3], Step [3175/3236], Loss: 2.0931, Perplexity: 8.1103
Epoch [3/3], Step [3176/3236], Loss: 1.8823, Perplexity: 6.5689
Epoch [3/3], Step [3177/3236], Loss: 1.9456, Perplexity: 6.9977
Epoch [3/3], Step [3178/3236], Loss: 1.8813, Perplexity: 6.5622
Epoch [3/3], Step [3179/3236], Loss: 1.8949, Perplexity: 6.6519
Epoch [3/3], Step [3180/3236], Loss: 2.0559, Perplexity: 7.8139
Epoch [3/3], Step [3181/3236], Loss: 1.7937, Perplexity: 6.0117
Epoch [3/3], Step [3182/3236], Loss: 2.0097, Perplexity: 7.4613
Epoch [3/3], Step [3183/3236], Loss: 2.0618, Perplexity: 7.8600
Epoch [3/3], Step [3184/3236], Loss: 1.7948, Perplexity: 6.0183
Epoch [3/3], Step [3185/3236], Loss: 2.0056, Perplexity: 7.4309
Epoch [3/3], Step [3186/3236], Loss: 1.8952, Perplexity: 6.6539
Epoch [3/3], Step [3187/3236], Loss: 2.1473, Perplexity: 8.5615
Epoch [3/3], Step [3188/3236], Loss: 1.8619, Perplexity: 6.4361
Epoch [3/3], Step [3189/3236], Loss: 1.8968, Perplexity: 6.6646
Epoch [3/3], Step [3190/3236], Loss: 1.9376, Perplexity: 6.9418
Epoch [3/3], Step [3191/3236], Loss: 1.9815, Perplexity: 7.2537
Epoch [3/3], Step [3192/3236], Loss: 1.8959, Perplexity: 6.6587
Epoch [3/3], Step [3193/3236], Loss: 1.8178, Perplexity: 6.1583
Epoch [3/3], Step [3194/3236], Loss: 1.9555, Perplexity: 7.0676
Epoch [3/3], Step [3195/3236], Loss: 2.3156, Perplexity: 10.1307
Epoch [3/3], Step [3196/3236], Loss: 1.9671, Perplexity: 7.1501
Epoch [3/3], Step [3197/3236], Loss: 1.9609, Perplexity: 7.1059
Epoch [3/3], Step [3198/3236], Loss: 2.9940, Perplexity: 19.9659
Epoch [3/3], Step [3199/3236], Loss: 1.8589, Perplexity: 6.4168
Epoch [3/3], Step [3200/3236], Loss: 2.1442, Perplexity: 8.5355
Epoch [3/3], Step [3201/3236], Loss: 1.8204, Perplexity: 6.1744
Epoch [3/3], Step [3202/3236], Loss: 2.0193, Perplexity: 7.5329
Epoch [3/3], Step [3203/3236], Loss: 1.9664, Perplexity: 7.1447
Epoch [3/3], Step [3204/3236], Loss: 1.7913, Perplexity: 5.9974
Epoch [3/3], Step [3205/3236], Loss: 2.0703, Perplexity: 7.9272
Epoch [3/3], Step [3206/3236], Loss: 1.8049, Perplexity: 6.0793
Epoch [3/3], Step [3207/3236], Loss: 1.9584, Perplexity: 7.0883
Epoch [3/3], Step [3208/3236], Loss: 1.8885, Perplexity: 6.6092
Epoch [3/3], Step [3209/3236], Loss: 1.9766, Perplexity: 7.2183
Epoch [3/3], Step [3210/3236], Loss: 1.8376, Perplexity: 6.2816
Epoch [3/3], Step [3211/3236], Loss: 2.0411, Perplexity: 7.6988
Epoch [3/3], Step [3212/3236], Loss: 1.9300, Perplexity: 6.8895
Epoch [3/3], Step [3213/3236], Loss: 2.1648, Perplexity: 8.7128
Epoch [3/3], Step [3214/3236], Loss: 2.8042, Perplexity: 16.5143
Epoch [3/3], Step [3215/3236], Loss: 1.7502, Perplexity: 5.7556
Epoch [3/3], Step [3216/3236], Loss: 2.1537, Perplexity: 8.6166
Epoch [3/3], Step [3217/3236], Loss: 1.9580, Perplexity: 7.0853
Epoch [3/3], Step [3218/3236], Loss: 1.7530, Perplexity: 5.7718
Epoch [3/3], Step [3219/3236], Loss: 2.3051, Perplexity: 10.0252
Epoch [3/3], Step [3220/3236], Loss: 1.8628, Perplexity: 6.4419
Epoch [3/3], Step [3221/3236], Loss: 1.8817, Perplexity: 6.5645
Epoch [3/3], Step [3222/3236], Loss: 1.8279, Perplexity: 6.2209
Epoch [3/3], Step [3223/3236], Loss: 1.8162, Perplexity: 6.1487
Epoch [3/3], Step [3224/3236], Loss: 2.0158, Perplexity: 7.5068
Epoch [3/3], Step [3225/3236], Loss: 1.9392, Perplexity: 6.9529
Epoch [3/3], Step [3226/3236], Loss: 2.2174, Perplexity: 9.1837
Epoch [3/3], Step [3227/3236], Loss: 1.9108, Perplexity: 6.7588
Epoch [3/3], Step [3228/3236], Loss: 2.0178, Perplexity: 7.5218
Epoch [3/3], Step [3229/3236], Loss: 1.8793, Perplexity: 6.5487
Epoch [3/3], Step [3230/3236], Loss: 1.9189, Perplexity: 6.8135
Epoch [3/3], Step [3231/3236], Loss: 1.8200, Perplexity: 6.1718
Epoch [3/3], Step [3232/3236], Loss: 1.9759, Perplexity: 7.2129
Epoch [3/3], Step [3233/3236], Loss: 2.0416, Perplexity: 7.7032
Epoch [3/3], Step [3234/3236], Loss: 2.0092, Perplexity: 7.4572
Epoch [3/3], Step [3235/3236], Loss: 1.9189, Perplexity: 6.8132
Epoch [3/3], Step [3236/3236], Loss: 1.9575, Perplexity: 7.0816