diff --git a/models/intel_optimized_models/int8/inceptionv3_int8.prototxt b/models/intel_optimized_models/int8/inceptionv3_int8.prototxt
index 70a52b698..40a810628 100644
--- a/models/intel_optimized_models/int8/inceptionv3_int8.prototxt
+++ b/models/intel_optimized_models/int8/inceptionv3_int8.prototxt
@@ -1,3 +1,4 @@
+# For INT8 reference
 name: "InceptionV3"
 layer {
   name: "data"
@@ -90,7 +91,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 75.3399963379
-    scale_out: 24.3700008392
+    scale_out: 48.9399986267
     scale_params: 125.419998169
   }
 }
@@ -134,7 +135,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 48.9399986267
-    scale_out: 33.9199981689
+    scale_out: 68.1200027466
     scale_params: 238.300003052
   }
 }
@@ -190,7 +191,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 68.1200027466
-    scale_out: 37.0099983215
+    scale_out: 74.3199996948
     scale_params: 115.459999084
   }
 }
@@ -234,7 +235,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 74.3199996948
-    scale_out: 61.7999992371
+    scale_out: 124.080001831
     scale_params: 490.480010986
   }
 }
@@ -290,7 +291,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 124.080001831
-    scale_out: 62.7400016785
+    scale_out: 125.970001221
     scale_params: 239.179992676
   }
 }
@@ -334,7 +335,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 124.080001831
-    scale_out: 79.2399978638
+    scale_out: 159.11000061
     scale_params: 303.589996338
   }
 }
@@ -378,7 +379,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 159.11000061
-    scale_out: 69.2699966431
+    scale_out: 139.080001831
     scale_params: 495.859985352
   }
 }
@@ -422,7 +423,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 124.080001831
-    scale_out: 72.8399963379
+    scale_out: 146.25
     scale_params: 215.38999939
   }
 }
@@ -466,7 +467,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 146.25
-    scale_out: 85.2699966431
+    scale_out: 171.210006714
     scale_params: 322.309997559
   }
 }
@@ -510,7 +511,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 171.210006714
-    scale_out: 57.8400001526
+    scale_out: 116.150001526
     scale_params: 303.209991455
   }
 }
@@ -566,7 +567,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 164.720001221
-    scale_out: 87.5400009155
+    scale_out: 175.770004272
     scale_params: 160.130004883
   }
 }
@@ -621,8 +622,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 57.8400001526
-    scale_out: 63.8800010681
+    scale_in: 116.150001526
+    scale_out: 128.270004272
     scale_params: 229.300003052
   }
 }
@@ -665,8 +666,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 57.8400001526
-    scale_out: 48.7599983215
+    scale_in: 116.150001526
+    scale_out: 97.9100036621
     scale_params: 171.490005493
   }
 }
@@ -710,7 +711,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 97.9100036621
-    scale_out: 54.2099990845
+    scale_out: 108.839996338
     scale_params: 516.630004883
   }
 }
@@ -753,8 +754,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 57.8400001526
-    scale_out: 77.5999984741
+    scale_in: 116.150001526
+    scale_out: 155.809997559
     scale_params: 194.460006714
   }
 }
@@ -798,7 +799,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 155.809997559
-    scale_out: 100.75
+    scale_out: 202.300003052
     scale_params: 243.229995728
   }
 }
@@ -842,7 +843,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 202.300003052
-    scale_out: 52.8199996948
+    scale_out: 106.050003052
     scale_params: 90.2200012207
   }
 }
@@ -897,8 +898,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 80.2900009155
-    scale_out: 100.050003052
+    scale_in: 161.210006714
+    scale_out: 200.88999939
     scale_params: 146.809997559
   }
 }
@@ -953,8 +954,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 52.8199996948
-    scale_out: 78.9400024414
+    scale_in: 106.050003052
+    scale_out: 158.509994507
     scale_params: 251.970001221
   }
 }
@@ -997,8 +998,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 52.8199996948
-    scale_out: 50.7700004578
+    scale_in: 106.050003052
+    scale_out: 101.940002441
     scale_params: 211.789993286
   }
 }
@@ -1042,7 +1043,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 101.940002441
-    scale_out: 62.0
+    scale_out: 124.5
     scale_params: 537.25
   }
 }
@@ -1085,8 +1086,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 52.8199996948
-    scale_out: 80.1200027466
+    scale_in: 106.050003052
+    scale_out: 160.86000061
     scale_params: 225.440002441
   }
 }
@@ -1130,7 +1131,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 160.86000061
-    scale_out: 62.7799987793
+    scale_out: 126.050003052
     scale_params: 141.570007324
   }
 }
@@ -1174,7 +1175,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 126.050003052
-    scale_out: 65.6800003052
+    scale_out: 131.869995117
     scale_params: 423.859985352
   }
 }
@@ -1229,8 +1230,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 85.0
-    scale_out: 95.9300003052
+    scale_in: 170.679992676
+    scale_out: 192.61000061
     scale_params: 127.099998474
   }
 }
@@ -1285,8 +1286,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 62.0
-    scale_out: 76.8199996948
+    scale_in: 124.5
+    scale_out: 154.25
     scale_params: 251.36000061
   }
 }
@@ -1329,8 +1330,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 62.0
-    scale_out: 96.2900009155
+    scale_in: 124.5
+    scale_out: 193.350006104
     scale_params: 278.760009766
   }
 }
@@ -1374,7 +1375,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 193.350006104
-    scale_out: 132.729995728
+    scale_out: 266.5
     scale_params: 457.459991455
   }
 }
@@ -1418,7 +1419,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 266.5
-    scale_out: 58.7700004578
+    scale_out: 118.0
     scale_params: 270.910003662
   }
 }
@@ -1484,8 +1485,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 58.7700004578
-    scale_out: 75.8399963379
+    scale_in: 118.0
+    scale_out: 152.289993286
     scale_params: 153.630004883
   }
 }
@@ -1528,8 +1529,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 58.7700004578
-    scale_out: 55.1300010681
+    scale_in: 118.0
+    scale_out: 110.699996948
     scale_params: 182.440002441
   }
 }
@@ -1575,7 +1576,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 110.699996948
-    scale_out: 48.0499992371
+    scale_out: 96.4899978638
     scale_params: 157.910003662
   }
 }
@@ -1621,7 +1622,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 96.4899978638
-    scale_out: 62.3899993896
+    scale_out: 125.279998779
     scale_params: 241.160003662
   }
 }
@@ -1664,8 +1665,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 58.7700004578
-    scale_out: 145.690002441
+    scale_in: 118.0
+    scale_out: 292.529998779
     scale_params: 284.149993896
   }
 }
@@ -1711,7 +1712,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 292.529998779
-    scale_out: 125.720001221
+    scale_out: 252.419998169
     scale_params: 290.660003662
   }
 }
@@ -1757,7 +1758,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 252.419998169
-    scale_out: 69.25
+    scale_out: 139.050003052
     scale_params: 173.460006714
   }
 }
@@ -1803,7 +1804,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 139.050003052
-    scale_out: 79.2600021362
+    scale_out: 159.149993896
     scale_params: 185.009994507
   }
 }
@@ -1849,7 +1850,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 159.149993896
-    scale_out: 76.5999984741
+    scale_out: 153.809997559
     scale_params: 206.669998169
   }
 }
@@ -1904,8 +1905,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 100.790000916
-    scale_out: 93.4000015259
+    scale_in: 202.369995117
+    scale_out: 187.529998779
     scale_params: 116.709999084
   }
 }
@@ -1960,8 +1961,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 62.3899993896
-    scale_out: 69.3000030518
+    scale_in: 125.279998779
+    scale_out: 139.149993896
     scale_params: 192.5
   }
 }
@@ -2004,8 +2005,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 62.3899993896
-    scale_out: 54.7299995422
+    scale_in: 125.279998779
+    scale_out: 109.88999939
     scale_params: 252.080001831
   }
 }
@@ -2051,7 +2052,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 109.88999939
-    scale_out: 54.1800003052
+    scale_out: 108.779998779
     scale_params: 196.350006104
   }
 }
@@ -2097,7 +2098,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 108.779998779
-    scale_out: 47.1699981689
+    scale_out: 94.7200012207
     scale_params: 218.880004883
   }
 }
@@ -2140,8 +2141,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 62.3899993896
-    scale_out: 107.379997253
+    scale_in: 125.279998779
+    scale_out: 215.61000061
     scale_params: 294.489990234
   }
 }
@@ -2187,7 +2188,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 215.61000061
-    scale_out: 72.3499984741
+    scale_out: 145.259994507
     scale_params: 186.509994507
   }
 }
@@ -2233,7 +2234,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 145.259994507
-    scale_out: 62.2599983215
+    scale_out: 125.019996643
     scale_params: 243.020004272
   }
 }
@@ -2279,7 +2280,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 125.019996643
-    scale_out: 82.0299987793
+    scale_out: 164.699996948
     scale_params: 320.420013428
   }
 }
@@ -2325,7 +2326,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 164.699996948
-    scale_out: 78.0699996948
+    scale_out: 156.759994507
     scale_params: 216.5
   }
 }
@@ -2380,8 +2381,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 101.949996948
-    scale_out: 64.6999969482
+    scale_in: 204.699996948
+    scale_out: 129.899993896
     scale_params: 149.190002441
   }
 }
@@ -2436,8 +2437,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 47.1699981689
-    scale_out: 67.3399963379
+    scale_in: 94.7200012207
+    scale_out: 135.210006714
     scale_params: 165.38999939
   }
 }
@@ -2480,8 +2481,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 47.1699981689
-    scale_out: 72.6299972534
+    scale_in: 94.7200012207
+    scale_out: 145.830001831
     scale_params: 194.589996338
   }
 }
@@ -2527,7 +2528,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 145.830001831
-    scale_out: 71.2099990845
+    scale_out: 142.979995728
     scale_params: 200.419998169
   }
 }
@@ -2573,7 +2574,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 142.979995728
-    scale_out: 71.7600021362
+    scale_out: 144.089996338
     scale_params: 248.820007324
   }
 }
@@ -2616,8 +2617,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 47.1699981689
-    scale_out: 89.7200012207
+    scale_in: 94.7200012207
+    scale_out: 180.13999939
     scale_params: 276.739990234
   }
 }
@@ -2663,7 +2664,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 180.13999939
-    scale_out: 51.8300018311
+    scale_out: 104.069999695
     scale_params: 231.600006104
   }
 }
@@ -2709,7 +2710,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 104.069999695
-    scale_out: 58.4500007629
+    scale_out: 117.370002747
     scale_params: 192.809997559
   }
 }
@@ -2755,7 +2756,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 117.370002747
-    scale_out: 54.4000015259
+    scale_out: 109.230003357
     scale_params: 210.199996948
   }
 }
@@ -2801,7 +2802,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 109.230003357
-    scale_out: 86.8899993896
+    scale_out: 174.460006714
     scale_params: 312.559997559
   }
 }
@@ -2856,8 +2857,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 72.2799987793
-    scale_out: 104.019996643
+    scale_in: 145.119995117
+    scale_out: 208.850006104
     scale_params: 119.339996338
   }
 }
@@ -2912,8 +2913,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 67.3399963379
-    scale_out: 67.4499969482
+    scale_in: 135.210006714
+    scale_out: 135.440002441
     scale_params: 252.38999939
   }
 }
@@ -2956,8 +2957,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 67.3399963379
-    scale_out: 64.8700027466
+    scale_in: 135.210006714
+    scale_out: 130.25
     scale_params: 369.850006104
   }
 }
@@ -3003,7 +3004,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 130.25
-    scale_out: 52.3899993896
+    scale_out: 105.180000305
     scale_params: 260.700012207
   }
 }
@@ -3049,7 +3050,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 105.180000305
-    scale_out: 76.4199981689
+    scale_out: 153.440002441
     scale_params: 316.489990234
   }
 }
@@ -3092,8 +3093,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 67.3399963379
-    scale_out: 80.0599975586
+    scale_in: 135.210006714
+    scale_out: 160.740005493
     scale_params: 264.869995117
   }
 }
@@ -3139,7 +3140,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 160.740005493
-    scale_out: 72.5599975586
+    scale_out: 145.690002441
     scale_params: 273.700012207
   }
 }
@@ -3185,7 +3186,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 145.690002441
-    scale_out: 67.7699966431
+    scale_out: 136.080001831
     scale_params: 273.950012207
   }
 }
@@ -3231,7 +3232,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 136.080001831
-    scale_out: 68.7900009155
+    scale_out: 138.130004883
     scale_params: 271.679992676
   }
 }
@@ -3277,7 +3278,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 138.130004883
-    scale_out: 83.0400009155
+    scale_out: 166.729995728
     scale_params: 331.309997559
   }
 }
@@ -3332,8 +3333,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 116.220001221
-    scale_out: 93.6600036621
+    scale_in: 233.350006104
+    scale_out: 188.050003052
     scale_params: 129.759994507
   }
 }
@@ -3388,8 +3389,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 67.4499969482
-    scale_out: 75.5800018311
+    scale_in: 135.440002441
+    scale_out: 151.75
     scale_params: 245.309997559
   }
 }
@@ -3433,7 +3434,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 151.75
-    scale_out: 40.4199981689
+    scale_out: 81.1500015259
     scale_params: 171.190002441
   }
 }
@@ -3476,8 +3477,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 67.4499969482
-    scale_out: 73.7399978638
+    scale_in: 135.440002441
+    scale_out: 148.070007324
     scale_params: 135.559997559
   }
 }
@@ -3523,7 +3524,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 148.070007324
-    scale_out: 57.2099990845
+    scale_out: 114.86000061
     scale_params: 123.540000916
   }
 }
@@ -3569,7 +3570,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 114.86000061
-    scale_out: 96.2099990845
+    scale_out: 193.179992676
     scale_params: 310.769989014
   }
 }
@@ -3613,7 +3614,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 193.179992676
-    scale_out: 67.4700012207
+    scale_out: 135.479995728
     scale_params: 313.369995117
   }
 }
@@ -3680,8 +3681,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 40.4199981689
-    scale_out: 156.029998779
+    scale_in: 81.1500015259
+    scale_out: 313.279998779
     scale_params: 400.739990234
   }
 }
@@ -3724,8 +3725,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 40.4199981689
-    scale_out: 87.5299987793
+    scale_in: 81.1500015259
+    scale_out: 175.759994507
     scale_params: 289.25
   }
 }
@@ -3771,7 +3772,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 175.759994507
-    scale_out: 103.349998474
+    scale_out: 207.509994507
     scale_params: 346.429992676
   }
 }
@@ -3817,7 +3818,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 175.759994507
-    scale_out: 93.3399963379
+    scale_out: 187.419998169
     scale_params: 263.369995117
   }
 }
@@ -3860,8 +3861,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 40.4199981689
-    scale_out: 70.1999969482
+    scale_in: 81.1500015259
+    scale_out: 140.960006714
     scale_params: 241.759994507
   }
 }
@@ -3905,7 +3906,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 140.960006714
-    scale_out: 63.2599983215
+    scale_out: 127.019996643
     scale_params: 514.510009766
   }
 }
@@ -3951,7 +3952,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 127.019996643
-    scale_out: 98.9100036621
+    scale_out: 198.600006104
     scale_params: 570.369995117
   }
 }
@@ -3997,7 +3998,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 127.019996643
-    scale_out: 99.2799987793
+    scale_out: 199.339996338
     scale_params: 365.929992676
   }
 }
@@ -4052,8 +4053,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 75.75
-    scale_out: 153.800003052
+    scale_in: 152.100006104
+    scale_out: 308.799987793
     scale_params: 119.88999939
   }
 }
@@ -4111,8 +4112,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 93.3399963379
-    scale_out: 16.8299999237
+    scale_in: 187.419998169
+    scale_out: 33.7900009155
     scale_params: 19.4200000763
   }
 }
@@ -4155,8 +4156,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 93.3399963379
-    scale_out: 61.1500015259
+    scale_in: 187.419998169
+    scale_out: 122.779998779
     scale_params: 105.769996643
   }
 }
@@ -4202,7 +4203,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 122.779998779
-    scale_out: 9.39999961853
+    scale_out: 18.8600006104
     scale_params: 35.6800003052
   }
 }
@@ -4248,7 +4249,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 122.779998779
-    scale_out: 9.22000026703
+    scale_out: 18.5200004578
     scale_params: 37.3199996948
   }
 }
@@ -4291,8 +4292,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 93.3399963379
-    scale_out: 60.1800003052
+    scale_in: 187.419998169
+    scale_out: 120.839996338
     scale_params: 92.4100036621
   }
 }
@@ -4336,7 +4337,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 120.839996338
-    scale_out: 94.8700027466
+    scale_out: 190.479995728
     scale_params: 345.420013428
   }
 }
@@ -4382,7 +4383,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 190.479995728
-    scale_out: 8.35000038147
+    scale_out: 16.7700004578
     scale_params: 44.3199996948
   }
 }
@@ -4428,7 +4429,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 190.479995728
-    scale_out: 10.9600000381
+    scale_out: 22.0100002289
     scale_params: 44.2700004578
   }
 }
@@ -4483,8 +4484,8 @@ layer {
     bw_layer_in: 8
     bw_layer_out: 8
     bw_params: 8
-    scale_in: 93.3399963379
-    scale_out: 19.9799995422
+    scale_in: 187.419998169
+    scale_out: 40.1100006104
     scale_params: 81.5699996948
   }
 }
@@ -4567,33 +4568,4 @@ layer {
     }
   }
 }
-layer {
-  name: "loss"
-  type: "SoftmaxWithLoss"
-  bottom: "fc1"
-  bottom: "label"
-  top: "loss"
-}
-layer {
-  name: "acc/top-1"
-  type: "Accuracy"
-  bottom: "fc1"
-  bottom: "label"
-  top: "acc/top-1"
-  include {
-    phase: TEST
-  }
-}
-layer {
-  name: "acc/top-5"
-  type: "Accuracy"
-  bottom: "fc1"
-  bottom: "label"
-  top: "acc/top-5"
-  include {
-    phase: TEST
-  }
-  accuracy_param {
-    top_k: 5
-  }
-}
+
diff --git a/models/intel_optimized_models/int8/resnet50_int8.prototxt b/models/intel_optimized_models/int8/resnet50_int8.prototxt
index c48020181..8d6aacd66 100644
--- a/models/intel_optimized_models/int8/resnet50_int8.prototxt
+++ b/models/intel_optimized_models/int8/resnet50_int8.prototxt
@@ -1,3 +1,4 @@
+# For INT8 reference
 name: "ResNet-50"
 layer {
   name: "data"
@@ -6,10 +7,10 @@ layer {
   dummy_data_param {
     data_filler {
       type: "constant"
-      value: 0.0099999997764825821
+      value: 0.01
     }    
     shape {
-      dim: 50
+      dim: 64
       dim: 3
       dim: 224
       dim: 224
@@ -25,7 +26,7 @@ layer {
       type: "constant"
     }    
     shape {
-      dim: 50
+      dim: 64
     }    
   }
 }
@@ -263,7 +264,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 19.9799995422
-    scale_out: 5.73999977112
+    scale_out: 11.5200004578
     scale_params: 72.1699981689
   }
 }
@@ -434,7 +435,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 19.5799999237
-    scale_out: 5.73999977112
+    scale_out: 11.5200004578
     scale_params: 88.4599990845
   }
 }
@@ -605,7 +606,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 13.5799999237
-    scale_out: 5.73999977112
+    scale_out: 11.5200004578
     scale_params: 103.910003662
   }
 }
@@ -822,7 +823,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 25.5400009155
-    scale_out: 4.78000020981
+    scale_out: 9.59000015259
     scale_params: 65.7200012207
   }
 }
@@ -993,7 +994,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 20.4899997711
-    scale_out: 4.78000020981
+    scale_out: 9.59000015259
     scale_params: 55.6899986267
   }
 }
@@ -1164,7 +1165,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 3.74000000954
-    scale_out: 4.78000020981
+    scale_out: 9.59000015259
     scale_params: 66.2300033569
   }
 }
@@ -1335,7 +1336,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 15.0600004196
-    scale_out: 4.78000020981
+    scale_out: 9.59000015259
     scale_params: 113.980003357
   }
 }
@@ -1552,7 +1553,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 15.3699998856
-    scale_out: 3.70000004768
+    scale_out: 7.42999982834
     scale_params: 75.6200027466
   }
 }
@@ -1723,7 +1724,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 6.44000005722
-    scale_out: 3.70000004768
+    scale_out: 7.42999982834
     scale_params: 67.4400024414
   }
 }
@@ -1894,7 +1895,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 6.94999980927
-    scale_out: 3.70000004768
+    scale_out: 7.42999982834
     scale_params: 76.9700012207
   }
 }
@@ -2065,7 +2066,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 5.40000009537
-    scale_out: 3.70000004768
+    scale_out: 7.42999982834
     scale_params: 77.1500015259
   }
 }
@@ -2236,7 +2237,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 6.51999998093
-    scale_out: 3.70000004768
+    scale_out: 7.42999982834
     scale_params: 98.8199996948
   }
 }
@@ -2407,7 +2408,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 5.01999998093
-    scale_out: 3.70000004768
+    scale_out: 7.42999982834
     scale_params: 93.3000030518
   }
 }
@@ -2624,7 +2625,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 25.2000007629
-    scale_out: 0.939999997616
+    scale_out: 1.87999999523
     scale_params: 27.1299991608
   }
 }
@@ -2795,7 +2796,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 29.2800006866
-    scale_out: 0.939999997616
+    scale_out: 1.87999999523
     scale_params: 35.3400001526
   }
 }
@@ -2966,7 +2967,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 14.7100000381
-    scale_out: 0.939999997616
+    scale_out: 1.87999999523
     scale_params: 23.2399997711
   }
 }
@@ -3031,26 +3032,4 @@ layer {
     }
   }
 }
-layer {
-  name: "accuracy/top-1"
-  type: "Accuracy"
-  bottom: "fc1000"
-  bottom: "label"
-  top: "accuracy-top1"
-  include {
-    phase: TEST
-  }
-}
-layer {
-  name: "accuracy/top-5"
-  type: "Accuracy"
-  bottom: "fc1000"
-  bottom: "label"
-  top: "accuracy-top5"
-  include {
-    phase: TEST
-  }
-  accuracy_param {
-    top_k: 5
-  }
-}
+
diff --git a/models/intel_optimized_models/int8/resnet50_sparse_int8.prototxt b/models/intel_optimized_models/int8/resnet50_sparse_int8.prototxt
index 402b24a63..22eb796ff 100644
--- a/models/intel_optimized_models/int8/resnet50_sparse_int8.prototxt
+++ b/models/intel_optimized_models/int8/resnet50_sparse_int8.prototxt
@@ -1,3 +1,4 @@
+# For INT8 reference
 name: "ResNet-50"
 layer {
   name: "data"
@@ -263,7 +264,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 19.9799995422
-    scale_out: 6.03000020981
+    scale_out: 12.1099996567
     scale_params: 72.1699981689
   }
 }
@@ -434,7 +435,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 19.5799999237
-    scale_out: 6.03000020981
+    scale_out: 12.1099996567
     scale_params: 88.4599990845
   }
 }
@@ -616,7 +617,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 16.9200000763
-    scale_out: 5.73999977112
+    scale_out: 11.5200004578
     scale_params: 103.910003662
   }
 }
@@ -833,7 +834,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 25.5400009155
-    scale_out: 4.65000009537
+    scale_out: 9.32999992371
     scale_params: 65.7200012207
   }
 }
@@ -1004,7 +1005,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 20.4899997711
-    scale_out: 4.65000009537
+    scale_out: 9.32999992371
     scale_params: 55.6899986267
   }
 }
@@ -1175,7 +1176,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 3.70000004768
-    scale_out: 4.65000009537
+    scale_out: 9.32999992371
     scale_params: 66.2300033569
   }
 }
@@ -1357,7 +1358,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 14.6999998093
-    scale_out: 4.88999986649
+    scale_out: 9.82999992371
     scale_params: 113.980003357
   }
 }
@@ -1574,7 +1575,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 14.8900003433
-    scale_out: 4.28000020981
+    scale_out: 8.60000038147
     scale_params: 75.6200027466
   }
 }
@@ -1745,7 +1746,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 6.30999994278
-    scale_out: 4.28000020981
+    scale_out: 8.60000038147
     scale_params: 67.4400024414
   }
 }
@@ -1916,7 +1917,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 6.90999984741
-    scale_out: 4.28000020981
+    scale_out: 8.60000038147
     scale_params: 76.9700012207
   }
 }
@@ -2087,7 +2088,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 5.40000009537
-    scale_out: 4.28000020981
+    scale_out: 8.60000038147
     scale_params: 77.1500015259
   }
 }
@@ -2258,7 +2259,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 6.51999998093
-    scale_out: 4.28000020981
+    scale_out: 8.60000038147
     scale_params: 98.8199996948
   }
 }
@@ -2440,7 +2441,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 16.6200008392
-    scale_out: 4.23000001907
+    scale_out: 8.5
     scale_params: 93.3000030518
   }
 }
@@ -2657,7 +2658,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 25.2000007629
-    scale_out: 0.939999997616
+    scale_out: 1.87999999523
     scale_params: 27.1299991608
   }
 }
@@ -2828,7 +2829,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 26.6200008392
-    scale_out: 0.939999997616
+    scale_out: 1.87999999523
     scale_params: 35.3400001526
   }
 }
@@ -2999,7 +3000,7 @@ layer {
     bw_layer_out: 8
     bw_params: 8
     scale_in: 14.7100000381
-    scale_out: 0.939999997616
+    scale_out: 1.87999999523
     scale_params: 23.2399997711
   }
 }
@@ -3064,36 +3065,4 @@ layer {
     }
   }
 }
-layer {
-  name: "prob"
-  type: "SoftmaxWithLoss"
-  bottom: "fc1000"
-  bottom: "label"
-  top: "prob"
-  include {
-    phase: TRAIN
-  }
-}
-layer {
-  name: "accuracy/top-1"
-  type: "Accuracy"
-  bottom: "fc1000"
-  bottom: "label"
-  top: "accuracy-top1"
-  include {
-    phase: TEST
-  }
-}
-layer {
-  name: "accuracy/top-5"
-  type: "Accuracy"
-  bottom: "fc1000"
-  bottom: "label"
-  top: "accuracy-top5"
-  include {
-    phase: TEST
-  }
-  accuracy_param {
-    top_k: 5
-  }
-}
+
diff --git a/models/intel_optimized_models/int8/ssd_int8.prototxt b/models/intel_optimized_models/int8/ssd_int8.prototxt
index 26e1a28dc..4a0ed0332 100644
--- a/models/intel_optimized_models/int8/ssd_int8.prototxt
+++ b/models/intel_optimized_models/int8/ssd_int8.prototxt
@@ -1,3 +1,4 @@
+# For INT8 reference
 name: "VGG_VOC0712_SSD_300x300_test"
 layer {
   name: "data"
@@ -1946,33 +1947,9 @@ layer {
       nms_threshold: 0.449999988079
       top_k: 400
     }
-    save_output_param {
-      output_directory: "data/ssd_out/VOC2007/SSD_300x300"
-      output_name_prefix: "comp4_det_test_"
-      output_format: "VOC"
-      label_map_file: "data/VOC0712/labelmap_voc.prototxt"
-      name_size_file: "data/VOC0712/test_name_size.txt"
-      num_test_image: 4952
-    }
     code_type: CENTER_SIZE
     keep_top_k: 200
     confidence_threshold: 0.00999999977648
   }
 }
-layer {
-  name: "detection_eval"
-  type: "DetectionEvaluate"
-  bottom: "detection_out"
-  bottom: "label"
-  top: "detection_eval"
-  include {
-    phase: TEST
-  }
-  detection_evaluate_param {
-    num_classes: 21
-    background_label_id: 0
-    overlap_threshold: 0.5
-    evaluate_difficult_gt: false
-    name_size_file: "data/VOC0712/test_name_size.txt"
-  }
-}
+
diff --git a/scripts/calibrator.py b/scripts/calibrator.py
index 7dcb7c54a..4bda3bb8d 100644
--- a/scripts/calibrator.py
+++ b/scripts/calibrator.py
@@ -64,23 +64,21 @@ def read_prototxt(prototxt):
     try:
         if not check_existence(prototxt):
             return None
-
         net = caffe_pb2.NetParameter()
-
         with open(prototxt) as f:
             txtf.Merge(f.read(), net)
 
         return net
 
     except Exception as e:
-        raise ("Failed to read {} due to {}".format(prototxt, str(e)))
+        print ("Failed to read {} due to {}".format(prototxt, e))
 
 
-def get_bottom_layers(layer_name, net, start):
+def get_bottom_layers(top_name, net, start):
     bottom_layers = []
     for index, value in enumerate(net.layer[start:]):
         for sub_index, sub_value in enumerate(value.bottom):
-            if sub_value == layer_name:
+            if sub_value == top_name:
                 bottom_layers.append((index, value.name, value.type))
 
     return bottom_layers
@@ -123,9 +121,9 @@ def get_all_top_layers(l, net, end, skip_layers, interesting_layers):
     return all_top_layers
 
 
-def get_all_bottom_layers(layer_name, net, start, skip_layers, interesting_layers):
+def get_all_bottom_layers(top_name, net, start, skip_layers, interesting_layers):
     all_bottom_layers = []
-    bottom_layers = get_bottom_layers(layer_name, net, start)
+    bottom_layers = get_bottom_layers(top_name, net, start)
     while True:
         if len(bottom_layers) == 0:
             break
@@ -149,12 +147,48 @@ def get_all_bottom_layers(layer_name, net, start, skip_layers, interesting_layer
     return all_bottom_layers
 
 
-def transform_convolutions(model_path):
+def get_fusion_conv_names(compiled_model):
+    compiled_net = caffe_pb2.NetParameter()
+    with open(compiled_model) as f:
+        s = f.read()
+        txtf.Merge(s, compiled_net)
+    return [(layer.name, layer.bottom[1]) for _, layer in enumerate(compiled_net.layer)
+            if layer.type == 'Convolution' and len(layer.bottom) > 1]
+
+
+def filter_fusion_layers(net, fusion_layer, conv_layer):
+    if not fusion_layer or not conv_layer:
+        return []
+    interesting_layers = ['ReLU']
+    skip_layers = ['Convolution', 'Eltwise', 'Concat']
+    output_with_relu_layer = [(l.name, net.layer[index].top[0]) for l, index in conv_layer
+                              if len(get_all_bottom_layers(net.layer[index].top[0], net, index + 1,
+                                                           skip_layers, interesting_layers)) == 0]
+    output_without_dict = {v: k for (k, v) in output_with_relu_layer}
+    for layer_name, top_name in fusion_layer:
+        if top_name in output_without_dict.keys():
+            del output_without_dict[top_name]
+
+    return output_without_dict.values()
+
+
+def check_relu_existence(net, start, end, exclude_layer):
+    if net.layer[start].type == 'Convolution' and net.layer[start].name in exclude_layer:
+        return False
+
+    for i in net.layer[start + 1: end]:
+        if i.type == 'ReLU':
+            return True
+    return False
+
+
+def transform_convolutions(model_path, compiled_model_path):
     net = caffe_pb2.NetParameter()
     with open(model_path) as f:
         s = f.read()
         txtf.Merge(s, net)
 
+    fusion_layer = get_fusion_conv_names(compiled_model_path)
     new_net = copy.deepcopy(net)
 
     convolution_layers = [(value, index) for index, value in enumerate(net.layer) if value.type == 'Convolution']
@@ -164,17 +198,18 @@ def transform_convolutions(model_path):
 
     u8_max = 255
     s8_max = 127
-
+    u8_layers = filter_fusion_layers(net, fusion_layer, convolution_layers)
     for (l, index) in convolution_layers:
-        outputwith_relu = get_all_bottom_layers(l.name, net, index + 1, skip_layers, interesting_layers)
+        outputwith_relu = get_all_bottom_layers(net.layer[index].top[0], net, index + 1, skip_layers,
+                                                interesting_layers)
+        conv_relu_flag = check_relu_existence(net, index,
+                                              convolution_layers[convolution_layers.index((l, index)) + 1][1]
+                                              if (l, index) != convolution_layers[-1]
+                                              else len(net.layer), [i[0] for i in fusion_layer])
         inputwith_relu = get_all_top_layers(l, net, index, skip_layers, interesting_layers)
-        # print "Processing", l.type, l.name
-
-        # output_type = 'u8' if outputwith_relu else 's8'
-        # input_type = 'u8' if inputwith_relu else 's8'
 
         for si in range(0, len(new_net.layer[index].quantization_param.scale_out)):
-            if len(outputwith_relu) > 0:  # u8
+            if len(outputwith_relu) > 0 or l.name in u8_layers or conv_relu_flag:  # u8
                 new_net.layer[index].quantization_param.scale_out[si] = round(u8_max / new_net.layer[index].
                                                                               quantization_param.scale_out[si], 2)
             else:  # s8
@@ -182,12 +217,12 @@ def transform_convolutions(model_path):
                                                                               quantization_param.scale_out[si], 2)
 
         for si in range(0, len(new_net.layer[index].quantization_param.scale_in)):
-            if len(inputwith_relu) > 0:  # u8
+            if len(inputwith_relu) > 0 or l.type == 'Convolution':  # u8
                 new_net.layer[index].quantization_param.scale_in[si] = round(u8_max / new_net.layer[index].
                                                                              quantization_param.scale_in[si], 2)
-            else:  # s8
-                new_net.layer[index].quantization_param.scale_in[si] = round(s8_max / new_net.layer[index].
-                                                                             quantization_param.scale_in[si], 2)
+            else:
+                new_net.layer[index].ClearField('quantization_param')
+                continue
 
         for si in range(0, len(new_net.layer[index].quantization_param.scale_params)):
             new_net.layer[index].quantization_param.scale_params[si] = round(s8_max / new_net.layer[index].
@@ -198,39 +233,51 @@ def transform_convolutions(model_path):
 
 
 def generate_sample(sample_path, input_model, weights,
-                    quantized_model, model_type, iterations=1, error_margin=1, power=0):
+                    quantized_model, detection, iterations=1, error_margin=1, power=0):
     cmd = '{0} quantize -model {1} -weights {2} -model_quantized {3} -iterations {4} ' \
           '-trimming_mode dynamic_fixed_point -error_margin {5} -power {6}'.format(sample_path, input_model, weights,
                                                                                    quantized_model, iterations,
                                                                                    error_margin, power)
-    if model_type == 3:
+    if detection:
         cmd += ' --detection=1'
 
     os.system(cmd)
 
 
-def get_the_accuracy(caffe_bin, model_def, model_weights, iterations, model_type):
+def get_compiled_net(caffe_bin, model_def, model_weights, detection):
+    output_log_name = '.compiled_net.txt'
+
+    cmd = '{} test -model {} -weights {} -iterations 1'.format(caffe_bin, model_def, model_weights)
+    if detection:
+        cmd += ' -detection'
+    cmd += ' 2>&1 > {}'.format(output_log_name)
+
+    os.system(cmd)
+    return os.path.abspath(output_log_name)
+
+
+def get_the_accuracy(caffe_bin, model_def, model_weights, iterations, detection, blob_name):
     output_log_name = 'calibrator_log.txt'
     cmd = '{} test -model {} -weights {} -iterations {}'.format(caffe_bin, model_def, model_weights, iterations)
-    if model_type == 3:
+    if detection:
         cmd += ' -detection'
     cmd += ' 2>&1|tee {}'.format(output_log_name)
+
     os.system(cmd)
+
     with open(output_log_name) as f:
         data = f.readlines()
-    try:
-        if model_type == 1:
-            top_1 = data[-2].strip()
-            return float(top_1.split('=')[-1].strip())
-        elif model_type == 2:
-            top_1 = data[-3].strip()
-            return float(top_1.split('=')[-1].strip())
-        elif model_type == 3:
-            top_1 = data[-1].strip()
-            return float(top_1.split('=')[-1].strip())
-    except Exception as e:
-        print 'Failed to generate accuracy due to {}'.format(str(e))
-        sys.exit(-1)
+
+    for i in data[::-1]:
+        if i.find('{} = '.format(blob_name)) != -1:
+            try:
+                return float(i.split('=')[-1].strip())
+            except Exception as e:
+                print 'Failed to generate accuracy due to {}'.format(str(e))
+                sys.exit(-1)
+
+    print 'Failed to get accuracy, please check the parameters and rerun the scripts.'
+    sys.exit(-1)
 
 
 def remove_top_quantized_parameter(current_quantized_file):
@@ -244,17 +291,33 @@ def remove_top_quantized_parameter(current_quantized_file):
         f.write(str(net))
 
 
-def tuning_quantized_topology(base_top1_accuracy, quantized_file, caffe_bin, model_weights, iterations,
-                              enable_floating_point, toleration, model_type):
-    if enable_floating_point == 0:
+def tuning_quantized_topology(base_top1_accuracy, prototxt, caffe_bin, model_weights, iterations,
+                              is_floating_point, accuracy_loss, detection, blob_name):
+    if is_floating_point == 0:
         print 'Updating quantization parameter...'
-        transform_convolutions(quantized_file)
-    current_top1_accuracy = get_the_accuracy(caffe_bin, quantized_file, model_weights, iterations, model_type)
-    while abs(current_top1_accuracy - base_top1_accuracy) >= toleration:
+
+        transform_convolutions(prototxt, get_compiled_net(caffe_bin, prototxt, model_weights, detection))
+
+    current_top1_accuracy = get_the_accuracy(caffe_bin, prototxt, model_weights, iterations, detection, blob_name)
+
+    while abs(current_top1_accuracy - base_top1_accuracy) >= accuracy_loss:
         print 'Tuning... '
         print abs(current_top1_accuracy - base_top1_accuracy)
-        remove_top_quantized_parameter(quantized_file)
-        current_top1_accuracy = get_the_accuracy(caffe_bin, quantized_prototxt, model_weights, iterations, model_type)
+        remove_top_quantized_parameter(prototxt)
+        current_top1_accuracy = get_the_accuracy(caffe_bin, prototxt, model_weights, iterations, detection, blob_name)
+
+
+def check_blob_name_existence(prototxt, blob_name):
+    net = read_prototxt(prototxt)
+    if not net.layer:
+        print 'Please check the model prototxt integrity.'
+        sys.exit(-1)
+
+    for i in net.layer[::-1]:
+        for _, value in enumerate(i.top):
+            if value == blob_name:
+                return True
+    return False
 
 
 if __name__ == '__main__':
@@ -265,36 +328,41 @@ def tuning_quantized_topology(base_top1_accuracy, quantized_file, caffe_bin, mod
                    ' -w pre-trained-fp32 weights ' \
                    ' -m typology ' \
                    ' -i iterations ' \
-                   ' -t resnet|inceptionv3|ssd\n '
+                   ' -l acceptable accuracy loss value, the default value is 0.01(stands for 1%)' \
+                   ' -d 1(0 means classification while 1 means detection, the default value is 0' \
+                   ' -n blob name which means accuracy.\n '
 
     parser = argparse.ArgumentParser(add_help=False)
     parser.add_argument('-h', '--help', action='help', help=usage_string)
 
     parser.add_argument('-i', '--iterations', action='store', dest='iterations', default=10,
-                        help='iterations')
+                        help='equal to the number to complete one epoch.')
 
     parser.add_argument('-w', '--weights', action='store', dest='weights', default='',
-                        help='pre-trained-fp32-weights')
+                        help='pre-trained-fp32-weights.')
 
     parser.add_argument('-m', '--model', action='store', dest='model', default='',
-                        help='model')
+                        help='topology definition prototxt.')
 
     parser.add_argument('-l', '--accuracy_loss', action='store', dest='loss', default=0.01,
-                        help='accuracy-loss')
+                        help='the acceptable accuracy loss that raised by 8-Bit quantization, '
+                             'default value is 0.01(1%).')
 
-    parser.add_argument('-t', '--type', action='store', dest='input_model_type', default='',
-                        help='model type')
+    parser.add_argument('-d', '--detection', action='store', dest='is_detection', default=0,
+                        help='0 for classification while 1 for detection, default value is 0.')
 
     parser.add_argument('-r', '--root', action='store', dest='root', default='',
                         help='caffe build path')
-    
+
+    parser.add_argument('-n', '--blob_name', action='store', dest='blob_name', default='',
+                        help='top blob name which stands for accuracy')
     params = parser.parse_args()
 
     try:
-        iterations = int(params.iterations)
+        user_input_iterations = int(params.iterations)
     except:
         print 'Set the iterations to the default value 1000'
-        iterations = 1000
+        user_input_iterations = 1000
 
     try:
         toleration = float(params.loss)
@@ -304,41 +372,41 @@ def tuning_quantized_topology(base_top1_accuracy, quantized_file, caffe_bin, mod
         print 'Set the toleration to 1%.'
         toleration = 0.01
 
+    try:
+        detection_flag = 1 if int(params.is_detection) == 1 else 0
+    except:
+        print 'Set the test type to classification.'
+        detection_flag = 0
+
     model = os.path.abspath(params.model)
-    weights = os.path.abspath(params.weights)
+    user_input_weights = os.path.abspath(params.weights)
     sample = os.path.abspath(params.root + 'tools/sample')
-    caffe_bin = os.path.abspath(params.root + 'tools/caffe')
+    caffe_bin_path = os.path.abspath(params.root + 'tools/caffe')
     setup_env()
 
-    if params.input_model_type == 'resnet':
-        model_type = 1
-    elif params.input_model_type == 'inceptionv3':
-        model_type = 2
-    elif params.input_model_type == 'ssd':
-        model_type = 3
-    else:
-        print 'Invalid model type!'
+    if not check_existence(model) or not check_existence(user_input_weights) or not check_existence(sample) \
+            or not check_existence(caffe_bin_path):
+        print 'Please check model/weights/sample existence.'
         sys.exit(-1)
 
-    if check_existence(model) is False or check_existence(weights) is False or check_existence(sample) is False or \
-            check_existence(caffe_bin) is False:
-        print 'Please check model/weights/sample existence.'
+    target_blob_name = params.blob_name
+    if not target_blob_name or not check_blob_name_existence(model, target_blob_name):
+        print 'Please specify valid blob name and rerun the script.'
         sys.exit(-1)
 
     sys.path.insert(0, params.root + '../python')
     quantized_prototxt = model.rsplit('.')[0] + '_quantized.prototxt'
-    quantized_weights = weights.rsplit('.')[0] + '_quantized.caffemodel'
+    quantized_weights = user_input_weights.rsplit('.')[0] + '_quantized.caffemodel'
     enable_floating_point = 0
     print 'Sampling...'
-    generate_sample(sample, model, weights,
-                    quantized_prototxt, model_type, 10, 100 * toleration, enable_floating_point)
-
+    generate_sample(sample, model, user_input_weights,
+                    quantized_prototxt, detection_flag, 10, 100 * toleration, enable_floating_point)
     print 'Sampling done'
     print 'Generating the FP32 accuracy...'
-    top_1 = get_the_accuracy(caffe_bin, model, weights, iterations, model_type)
+    top_1 = get_the_accuracy(caffe_bin_path, model, user_input_weights, user_input_iterations, detection_flag,
+                             target_blob_name)
     print 'FP32 accuracy is: {}'.format(top_1)
-
-    tuning_quantized_topology(top_1, quantized_prototxt, caffe_bin, weights, iterations, enable_floating_point,
-                              toleration, model_type)
+    tuning_quantized_topology(top_1, quantized_prototxt, caffe_bin_path, user_input_weights, user_input_iterations,
+                              enable_floating_point, toleration, detection_flag, target_blob_name)
 
     print 'Updated prototxt {} is generated.'.format(quantized_prototxt)