sanity_check.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# With the \"static\" method, you should find the same results if you set e.g. T=10 in the old code, compared to T=1 and num_sensors=10 in the new code"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# New Code:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\n10907700\\Anaconda3\\envs\\dad_vtorch\\lib\\site-packages\\tqdm\\auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    }
   ],
   "source": [
    "import warnings\n",
    "from location_finding_class import DAD\n",
    "import matplotlib.pyplot as plt\n",
    "import torch\n",
    "import numpy as np\n",
    "\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Loss: 0.499 : 100%|██████████| 10000/10000 [08:34<00:00, 19.44it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Example run 1\n",
      "*True Theta: tensor([[ 2.2828,  0.5623],\n",
      "        [-0.3015, -0.4345]])*\n",
      "xi1: tensor([[ 0.5416,  0.0524],\n",
      "        [-0.6090,  0.4837],\n",
      "        [-0.3368, -0.5050],\n",
      "        [ 0.1861,  0.4724],\n",
      "        [-0.1601,  0.4651],\n",
      "        [ 0.5813,  0.5394],\n",
      "        [ 0.5650, -0.5549],\n",
      "        [-0.5738, -0.3827],\n",
      "        [-0.4830,  0.1742],\n",
      "        [ 0.0174, -0.5035]], requires_grad=True)\n",
      " y1: tensor([ 0.4243,  0.2061,  4.3710, -0.3696, -0.0197, -0.6374,  0.7873,  2.0807,\n",
      "         1.1068,  1.9285])\n",
      "\n",
      "Example run 2\n",
      "*True Theta: tensor([[-0.2456,  1.9529],\n",
      "        [ 0.3205, -0.7407]])*\n",
      "xi1: tensor([[ 0.5416,  0.0524],\n",
      "        [-0.6090,  0.4837],\n",
      "        [-0.3368, -0.5050],\n",
      "        [ 0.1861,  0.4724],\n",
      "        [-0.1601,  0.4651],\n",
      "        [ 0.5813,  0.5394],\n",
      "        [ 0.5650, -0.5549],\n",
      "        [-0.5738, -0.3827],\n",
      "        [-0.4830,  0.1742],\n",
      "        [ 0.0174, -0.5035]], requires_grad=True)\n",
      " y1: tensor([ 0.4021, -0.5927,  1.5143,  0.6281,  0.2794, -0.0591,  2.7211,  0.2168,\n",
      "        -0.1691,  1.6407])\n",
      "\n",
      "Example run 3\n",
      "*True Theta: tensor([[-1.7495,  1.6867],\n",
      "        [-1.5444, -0.5566]])*\n",
      "xi1: tensor([[ 0.5416,  0.0524],\n",
      "        [-0.6090,  0.4837],\n",
      "        [-0.3368, -0.5050],\n",
      "        [ 0.1861,  0.4724],\n",
      "        [-0.1601,  0.4651],\n",
      "        [ 0.5813,  0.5394],\n",
      "        [ 0.5650, -0.5549],\n",
      "        [-0.5738, -0.3827],\n",
      "        [-0.4830,  0.1742],\n",
      "        [ 0.0174, -0.5035]], requires_grad=True)\n",
      " y1: tensor([-0.6149,  0.0600, -0.3558, -0.4718, -0.1401, -0.7248,  0.2171, -0.5375,\n",
      "        -0.5143, -1.2751])\n",
      "       xi_0      xi_1  observations  n  order  run_id\n",
      "0  0.541591  0.052357      0.424276  0      1       1\n",
      "1 -0.608995  0.483652      0.206058  1      1       1\n",
      "2 -0.336753 -0.505043      4.371039  2      1       1\n",
      "3  0.186138  0.472409     -0.369593  3      1       1\n",
      "4 -0.160059  0.465129     -0.019744  4      1       1\n",
      "5  0.581313  0.539374     -0.637443  5      1       1\n",
      "6  0.565034 -0.554893      0.787274  6      1       1\n",
      "7 -0.573774 -0.382739      2.080685  7      1       1\n",
      "8 -0.482988  0.174184      1.106816  8      1       1\n",
      "9  0.017398 -0.503502      1.928478  9      1       1\n",
      "0  0.541591  0.052357      0.402087  0      1       2\n",
      "1 -0.608995  0.483652     -0.592729  1      1       2\n",
      "2 -0.336753 -0.505043      1.514284  2      1       2\n",
      "3  0.186138  0.472409      0.628150  3      1       2\n",
      "4 -0.160059  0.465129      0.279428  4      1       2\n",
      "5  0.581313  0.539374     -0.059127  5      1       2\n",
      "6  0.565034 -0.554893      2.721147  6      1       2\n",
      "7 -0.573774 -0.382739      0.216766  7      1       2\n",
      "8 -0.482988  0.174184     -0.169145  8      1       2\n",
      "9  0.017398 -0.503502      1.640656  9      1       2\n",
      "0  0.541591  0.052357     -0.614860  0      1       3\n",
      "1 -0.608995  0.483652      0.059974  1      1       3\n",
      "2 -0.336753 -0.505043     -0.355824  2      1       3\n",
      "3  0.186138  0.472409     -0.471764  3      1       3\n",
      "4 -0.160059  0.465129     -0.140115  4      1       3\n",
      "5  0.581313  0.539374     -0.724810  5      1       3\n",
      "6  0.565034 -0.554893      0.217093  6      1       3\n",
      "7 -0.573774 -0.382739     -0.537484  7      1       3\n",
      "8 -0.482988  0.174184     -0.514260  8      1       3\n",
      "9  0.017398 -0.503502     -1.275076  9      1       3\n",
      "Saved. The experiment-id of this run is 936708953841566871\n"
     ]
    }
   ],
   "source": [
    "design = DAD(\n",
    "    n=10,  # num_sensors\n",
    "    T=1,\n",
    "    num_steps=10000,\n",
    "    num_inner_samples=64,\n",
    "    num_outer_samples=128,\n",
    "    design_network_type=\"static\",\n",
    ")\n",
    "fitted_design = design.fit()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_design(fitted_design, run_id=1, df=None):\n",
    "    # get the designs\n",
    "    df = fitted_design.ho_model.run_df.copy() if df is None else df\n",
    "    df = df[df[\"run_id\"] == run_id]\n",
    "\n",
    "    # generate 2 2d grids for the x & y bounds\n",
    "    x0s, x1s = np.meshgrid(np.linspace(-3, 3, 100), np.linspace(-3, 3, 100))\n",
    "\n",
    "    # get true thetas\n",
    "    tt = torch.tensor(fitted_design.ho_model.true_thetas[run_id - 1])\n",
    "\n",
    "    # generate contours\n",
    "    z = [\n",
    "        fitted_design.ho_model.forward_map(torch.tensor([x0, x1]).unsqueeze(-2), tt)\n",
    "        for (x0, x1) in zip(x0s.flatten(), x1s.flatten())\n",
    "    ]\n",
    "\n",
    "    # plot\n",
    "    fig, ax = plt.subplots()\n",
    "    c = ax.pcolormesh(x0s, x1s, np.array(z).reshape(100, 100), cmap=\"RdBu\")\n",
    "    ax.set_title(\n",
    "        \"Design - {} (n={})\".format(fitted_design.design_network_type, fitted_design.n)\n",
    "    )\n",
    "    fig.colorbar(c, ax=ax)\n",
    "    ax.scatter(df.xi_0, df.xi_1, s=2, c=\"black\")\n",
    "    for i, txt in enumerate(df.order):\n",
    "        ax.annotate(txt, (df.xi_0[i], df.xi_1[i]))\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_design(fitted_design)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Old Code:\n",
    "\n",
    "copied from: https://github.com/ae-foster/dad/blob/main/location_finding.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Example run 1\n",
      "*True Theta: tensor([[-0.1228,  0.8208],\n",
      "        [-1.3081,  0.6779]])*\n",
      "xi1: tensor([0.5696, 0.0404], requires_grad=True)\n",
      " y1: 0.8355629444122314\n",
      "xi2: tensor([-0.2931,  0.5897], requires_grad=True)\n",
      " y2: 2.249709367752075\n",
      "xi3: tensor([-0.5183, -0.5681], requires_grad=True)\n",
      " y3: 0.2053459882736206\n",
      "xi4: tensor([0.0989, 0.4882], requires_grad=True)\n",
      " y4: 1.6867191791534424\n",
      "xi5: tensor([ 0.2811, -0.4125], requires_grad=True)\n",
      " y5: 0.08608085662126541\n",
      "xi6: tensor([ 0.5207, -0.5528], requires_grad=True)\n",
      " y6: 0.7695834636688232\n",
      "xi7: tensor([-0.5893,  0.3612], requires_grad=True)\n",
      " y7: 1.2912057638168335\n",
      "xi8: tensor([0.5331, 0.5335], requires_grad=True)\n",
      " y8: 0.46314847469329834\n",
      "xi9: tensor([-0.1041, -0.4840], requires_grad=True)\n",
      " y9: 0.0810835212469101\n",
      "xi10: tensor([-0.5263, -0.1477], requires_grad=True)\n",
      " y10: 0.15569037199020386\n",
      "\n",
      "Example run 2\n",
      "*True Theta: tensor([[ 0.2143, -0.8467],\n",
      "        [-0.3639,  1.1265]])*\n",
      "xi1: tensor([0.5696, 0.0404], requires_grad=True)\n",
      " y1: 0.7459218502044678\n",
      "xi2: tensor([-0.2931,  0.5897], requires_grad=True)\n",
      " y2: 1.4425334930419922\n",
      "xi3: tensor([-0.5183, -0.5681], requires_grad=True)\n",
      " y3: -0.133087158203125\n",
      "xi4: tensor([0.0989, 0.4882], requires_grad=True)\n",
      " y4: 1.1267839670181274\n",
      "xi5: tensor([ 0.2811, -0.4125], requires_grad=True)\n",
      " y5: 1.9109206199645996\n",
      "xi6: tensor([ 0.5207, -0.5528], requires_grad=True)\n",
      " y6: 2.2925267219543457\n",
      "xi7: tensor([-0.5893,  0.3612], requires_grad=True)\n",
      " y7: 0.7713184356689453\n",
      "xi8: tensor([0.5331, 0.5335], requires_grad=True)\n",
      " y8: 0.41305914521217346\n",
      "xi9: tensor([-0.1041, -0.4840], requires_grad=True)\n",
      " y9: 1.6709489822387695\n",
      "xi10: tensor([-0.5263, -0.1477], requires_grad=True)\n",
      " y10: 0.09819445013999939\n",
      "\n",
      "Example run 3\n",
      "*True Theta: tensor([[-1.0412,  0.0423],\n",
      "        [ 1.0288, -0.3849]])*\n",
      "xi1: tensor([0.5696, 0.0404], requires_grad=True)\n",
      " y1: 0.7293790578842163\n",
      "xi2: tensor([-0.2931,  0.5897], requires_grad=True)\n",
      " y2: 0.17562028765678406\n",
      "xi3: tensor([-0.5183, -0.5681], requires_grad=True)\n",
      " y3: 0.6933813691139221\n",
      "xi4: tensor([0.0989, 0.4882], requires_grad=True)\n",
      " y4: 0.48270636796951294\n",
      "xi5: tensor([ 0.2811, -0.4125], requires_grad=True)\n",
      " y5: 1.2579445838928223\n",
      "xi6: tensor([ 0.5207, -0.5528], requires_grad=True)\n",
      " y6: 2.17256498336792\n",
      "xi7: tensor([-0.5893,  0.3612], requires_grad=True)\n",
      " y7: 0.6325554847717285\n",
      "xi8: tensor([0.5331, 0.5335], requires_grad=True)\n",
      " y8: -0.276590496301651\n",
      "xi9: tensor([-0.1041, -0.4840], requires_grad=True)\n",
      " y9: 0.43859440088272095\n",
      "xi10: tensor([-0.5263, -0.1477], requires_grad=True)\n",
      " y10: 1.5196144580841064\n",
      "       xi_0      xi_1  observations  order  run_id\n",
      "0  0.569597  0.040414      0.835563      1       1\n",
      "1 -0.293097  0.589723      2.249709      2       1\n",
      "2 -0.518251 -0.568128      0.205346      3       1\n",
      "3  0.098950  0.488185      1.686719      4       1\n",
      "4  0.281130 -0.412470      0.086081      5       1\n",
      "5  0.520696 -0.552815      0.769583      6       1\n",
      "6 -0.589286  0.361165      1.291206      7       1\n",
      "7  0.533144  0.533530      0.463148      8       1\n",
      "8 -0.104150 -0.484022      0.081084      9       1\n",
      "9 -0.526312 -0.147747      0.155690     10       1\n",
      "0  0.569597  0.040414      0.745922      1       2\n",
      "1 -0.293097  0.589723      1.442533      2       2\n",
      "2 -0.518251 -0.568128     -0.133087      3       2\n",
      "3  0.098950  0.488185      1.126784      4       2\n",
      "4  0.281130 -0.412470      1.910921      5       2\n",
      "5  0.520696 -0.552815      2.292527      6       2\n",
      "6 -0.589286  0.361165      0.771318      7       2\n",
      "7  0.533144  0.533530      0.413059      8       2\n",
      "8 -0.104150 -0.484022      1.670949      9       2\n",
      "9 -0.526312 -0.147747      0.098194     10       2\n",
      "0  0.569597  0.040414      0.729379      1       3\n",
      "1 -0.293097  0.589723      0.175620      2       3\n",
      "2 -0.518251 -0.568128      0.693381      3       3\n",
      "3  0.098950  0.488185      0.482706      4       3\n",
      "4  0.281130 -0.412470      1.257945      5       3\n",
      "5  0.520696 -0.552815      2.172565      6       3\n",
      "6 -0.589286  0.361165      0.632555      7       3\n",
      "7  0.533144  0.533530     -0.276590      8       3\n",
      "8 -0.104150 -0.484022      0.438594      9       3\n",
      "9 -0.526312 -0.147747      1.519614     10       3\n",
      "Storing model to MlFlow... Model sotred in mlruns/482487252312563380/d282d034a0254a4a90084a09d5be6778/artifacts/model. Done.\n",
      "The experiment-id of this run is 482487252312563380\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\n10907700\\Anaconda3\\envs\\dad_vtorch\\lib\\site-packages\\pyro\\distributions\\torch_patch.py:81: UserWarning: torch.symeig is deprecated in favor of torch.linalg.eigh and will be removed in a future PyTorch release.\n",
      "The default behavior has changed from using the upper triangular portion of the matrix by default to using the lower triangular portion.\n",
      "L, _ = torch.symeig(A, upper=upper)\n",
      "should be replaced with\n",
      "L = torch.linalg.eigvalsh(A, UPLO='U' if upper else 'L')\n",
      "and\n",
      "L, V = torch.symeig(A, eigenvectors=True)\n",
      "should be replaced with\n",
      "L, V = torch.linalg.eigh(A, UPLO='U' if upper else 'L') (Triggered internally at  ..\\aten\\src\\ATen\\native\\BatchLinearAlgebra.cpp:2524.)\n",
      "  return torch.stack([v.symeig(eigenvectors=False)[0][:1] > 0.0\n",
      "\n",
      "Loss: 0.000 :   0%|          | 0/10000 [00:00<?, ?it/s]\n",
      "Loss: 2.490 :   0%|          | 0/10000 [00:00<?, ?it/s]\n",
      "Loss: 2.490 :   0%|          | 2/10000 [00:00<12:14, 13.62it/s]\n",
      "Loss: 2.490 :   0%|          | 5/10000 [00:00<09:46, 17.04it/s]\n",
      "Loss: 2.490 :   0%|          | 7/10000 [00:00<09:16, 17.96it/s]\n",
      "Loss: 2.490 :   0%|          | 9/10000 [00:00<08:56, 18.61it/s]\n",
      "Loss: 2.490 :   0%|          | 11/10000 [00:00<09:13, 18.05it/s]\n",
      "Loss: 2.490 :   0%|          | 13/10000 [00:00<09:06, 18.28it/s]\n",
      "Loss: 2.490 :   0%|          | 15/10000 [00:00<09:02, 18.41it/s]\n",
      "Loss: 2.490 :   0%|          | 17/10000 [00:00<09:17, 17.92it/s]\n",
      "Loss: 2.490 :   0%|          | 19/10000 [00:01<09:16, 17.93it/s]\n",
      "Loss: 2.490 :   0%|          | 21/10000 [00:01<09:22, 17.74it/s]\n",
      "Loss: 2.490 :   0%|          | 23/10000 [00:01<09:22, 17.73it/s]\n",
      "Loss: 2.490 :   0%|          | 25/10000 [00:01<09:24, 17.66it/s]\n",
      "Loss: 2.490 :   0%|          | 27/10000 [00:01<09:28, 17.54it/s]\n",
      "Loss: 2.490 :   0%|          | 29/10000 [00:01<09:30, 17.47it/s]\n",
      "Loss: 2.490 :   0%|          | 31/10000 [00:01<09:26, 17.61it/s]\n",
      "Loss: 2.490 :   0%|          | 33/10000 [00:01<09:19, 17.82it/s]\n",
      "Loss: 2.490 :   0%|          | 35/10000 [00:01<09:24, 17.66it/s]\n",
      "Loss: 2.490 :   0%|          | 37/10000 [00:02<09:31, 17.45it/s]\n",
      "Loss: 2.490 :   0%|          | 39/10000 [00:02<09:47, 16.94it/s]\n",
      "Loss: 2.490 :   0%|          | 41/10000 [00:02<09:45, 17.02it/s]\n",
      "Loss: 2.490 :   0%|          | 43/10000 [00:02<09:40, 17.15it/s]\n",
      "Loss: 2.490 :   0%|          | 45/10000 [00:02<09:44, 17.03it/s]\n",
      "Loss: 2.490 :   0%|          | 47/10000 [00:02<10:08, 16.35it/s]\n",
      "Loss: 2.490 :   0%|          | 49/10000 [00:02<10:53, 15.24it/s]\n",
      "Loss: 2.455 :   0%|          | 49/10000 [00:02<10:53, 15.24it/s]\n",
      "Loss: 2.455 :   1%|          | 51/10000 [00:03<12:02, 13.78it/s]\n",
      "Loss: 2.455 :   1%|          | 53/10000 [00:03<14:40, 11.30it/s]\n",
      "Loss: 2.455 :   1%|          | 55/10000 [00:03<13:21, 12.41it/s]\n",
      "Loss: 2.455 :   1%|          | 57/10000 [00:03<12:38, 13.11it/s]\n",
      "Loss: 2.455 :   1%|          | 59/10000 [00:03<11:55, 13.89it/s]\n",
      "Loss: 2.455 :   1%|          | 61/10000 [00:03<11:23, 14.54it/s]\n",
      "Loss: 2.455 :   1%|          | 63/10000 [00:03<10:52, 15.24it/s]\n",
      "Loss: 2.455 :   1%|          | 65/10000 [00:04<10:31, 15.73it/s]\n",
      "Loss: 2.455 :   1%|          | 67/10000 [00:04<10:17, 16.09it/s]\n",
      "Loss: 2.455 :   1%|          | 69/10000 [00:04<10:06, 16.36it/s]\n",
      "Loss: 2.455 :   1%|          | 71/10000 [00:04<09:54, 16.69it/s]\n",
      "Loss: 2.455 :   1%|          | 73/10000 [00:04<09:56, 16.64it/s]\n",
      "Loss: 2.455 :   1%|          | 75/10000 [00:04<09:45, 16.94it/s]\n",
      "Loss: 2.455 :   1%|          | 77/10000 [00:04<09:36, 17.21it/s]\n",
      "Loss: 2.455 :   1%|          | 79/10000 [00:04<09:24, 17.57it/s]\n",
      "Loss: 2.455 :   1%|          | 81/10000 [00:04<09:13, 17.92it/s]\n",
      "Loss: 2.455 :   1%|          | 83/10000 [00:05<09:13, 17.93it/s]\n",
      "Loss: 2.455 :   1%|          | 85/10000 [00:05<09:05, 18.16it/s]\n",
      "Loss: 2.455 :   1%|          | 87/10000 [00:05<09:07, 18.10it/s]\n",
      "Loss: 2.455 :   1%|          | 89/10000 [00:05<09:09, 18.04it/s]\n",
      "Loss: 2.455 :   1%|          | 91/10000 [00:05<09:02, 18.27it/s]\n",
      "Loss: 2.455 :   1%|          | 93/10000 [00:05<09:12, 17.94it/s]\n",
      "Loss: 2.455 :   1%|          | 95/10000 [00:05<09:08, 18.04it/s]\n",
      "Loss: 2.455 :   1%|          | 97/10000 [00:05<09:06, 18.11it/s]\n",
      "Loss: 2.455 :   1%|          | 99/10000 [00:05<09:10, 18.00it/s]\n",
      "Loss: 2.515 :   1%|          | 99/10000 [00:06<09:10, 18.00it/s]\n",
      "Loss: 2.515 :   1%|          | 101/10000 [00:06<10:09, 16.24it/s]\n",
      "Loss: 2.515 :   1%|          | 103/10000 [00:06<09:54, 16.63it/s]\n",
      "Loss: 2.515 :   1%|          | 105/10000 [00:06<09:53, 16.67it/s]\n",
      "Loss: 2.515 :   1%|          | 107/10000 [00:06<09:38, 17.09it/s]\n",
      "Loss: 2.515 :   1%|          | 109/10000 [00:06<09:32, 17.28it/s]\n",
      "Loss: 2.515 :   1%|          | 111/10000 [00:06<09:17, 17.73it/s]\n",
      "Loss: 2.515 :   1%|          | 113/10000 [00:06<09:24, 17.51it/s]\n",
      "Loss: 2.515 :   1%|          | 115/10000 [00:06<09:19, 17.67it/s]\n",
      "Loss: 2.515 :   1%|          | 117/10000 [00:06<09:12, 17.87it/s]\n",
      "Loss: 2.515 :   1%|          | 119/10000 [00:07<09:17, 17.72it/s]\n",
      "Loss: 2.515 :   1%|          | 121/10000 [00:07<09:13, 17.85it/s]\n",
      "Loss: 2.515 :   1%|          | 123/10000 [00:07<09:13, 17.85it/s]\n",
      "Loss: 2.515 :   1%|▏         | 125/10000 [00:07<09:06, 18.07it/s]\n",
      "Loss: 2.515 :   1%|▏         | 127/10000 [00:07<09:06, 18.06it/s]\n",
      "Loss: 2.515 :   1%|▏         | 129/10000 [00:07<08:54, 18.46it/s]\n",
      "Loss: 2.515 :   1%|▏         | 131/10000 [00:07<08:52, 18.52it/s]\n",
      "Loss: 2.515 :   1%|▏         | 133/10000 [00:07<08:58, 18.33it/s]\n",
      "Loss: 2.515 :   1%|▏         | 135/10000 [00:07<09:09, 17.97it/s]\n",
      "Loss: 2.515 :   1%|▏         | 137/10000 [00:08<10:12, 16.11it/s]\n",
      "Loss: 2.515 :   1%|▏         | 139/10000 [00:08<10:36, 15.50it/s]\n",
      "Loss: 2.515 :   1%|▏         | 141/10000 [00:08<10:37, 15.46it/s]\n",
      "Loss: 2.515 :   1%|▏         | 143/10000 [00:08<11:54, 13.80it/s]\n",
      "Loss: 2.515 :   1%|▏         | 145/10000 [00:08<11:14, 14.61it/s]\n",
      "Loss: 2.515 :   1%|▏         | 147/10000 [00:08<10:57, 14.99it/s]\n",
      "Loss: 2.515 :   1%|▏         | 149/10000 [00:08<10:23, 15.80it/s]\n",
      "Loss: 2.325 :   1%|▏         | 149/10000 [00:09<10:23, 15.80it/s]\n",
      "Loss: 2.325 :   2%|▏         | 151/10000 [00:09<10:59, 14.93it/s]\n",
      "Loss: 2.325 :   2%|▏         | 153/10000 [00:09<10:34, 15.52it/s]\n",
      "Loss: 2.325 :   2%|▏         | 155/10000 [00:09<10:17, 15.94it/s]\n",
      "Loss: 2.325 :   2%|▏         | 157/10000 [00:09<09:54, 16.56it/s]\n",
      "Loss: 2.325 :   2%|▏         | 159/10000 [00:09<09:44, 16.83it/s]\n",
      "Loss: 2.325 :   2%|▏         | 161/10000 [00:09<09:30, 17.25it/s]\n",
      "Loss: 2.325 :   2%|▏         | 163/10000 [00:09<09:16, 17.69it/s]\n",
      "Loss: 2.325 :   2%|▏         | 165/10000 [00:09<09:21, 17.50it/s]\n",
      "Loss: 2.325 :   2%|▏         | 167/10000 [00:09<09:14, 17.75it/s]\n",
      "Loss: 2.325 :   2%|▏         | 169/10000 [00:10<09:08, 17.92it/s]\n",
      "Loss: 2.325 :   2%|▏         | 171/10000 [00:10<09:03, 18.08it/s]\n",
      "Loss: 2.325 :   2%|▏         | 173/10000 [00:10<09:02, 18.12it/s]\n",
      "Loss: 2.325 :   2%|▏         | 175/10000 [00:10<08:59, 18.21it/s]\n",
      "Loss: 2.325 :   2%|▏         | 177/10000 [00:10<09:08, 17.90it/s]\n",
      "Loss: 2.325 :   2%|▏         | 179/10000 [00:10<09:02, 18.10it/s]\n",
      "Loss: 2.325 :   2%|▏         | 181/10000 [00:10<08:53, 18.39it/s]\n",
      "Loss: 2.325 :   2%|▏         | 183/10000 [00:10<08:58, 18.24it/s]\n",
      "Loss: 2.325 :   2%|▏         | 185/10000 [00:10<08:54, 18.36it/s]\n",
      "Loss: 2.325 :   2%|▏         | 187/10000 [00:11<08:57, 18.26it/s]\n",
      "Loss: 2.325 :   2%|▏         | 189/10000 [00:11<09:00, 18.16it/s]\n",
      "Loss: 2.325 :   2%|▏         | 191/10000 [00:11<08:52, 18.42it/s]\n",
      "Loss: 2.325 :   2%|▏         | 193/10000 [00:11<08:56, 18.28it/s]\n",
      "Loss: 2.325 :   2%|▏         | 195/10000 [00:11<09:05, 17.97it/s]\n",
      "Loss: 2.325 :   2%|▏         | 197/10000 [00:11<09:00, 18.12it/s]\n",
      "Loss: 2.325 :   2%|▏         | 199/10000 [00:11<09:04, 17.99it/s]\n",
      "Loss: 2.413 :   2%|▏         | 199/10000 [00:11<09:04, 17.99it/s]\n",
      "Loss: 2.413 :   2%|▏         | 201/10000 [00:11<09:59, 16.34it/s]\n",
      "Loss: 2.413 :   2%|▏         | 203/10000 [00:12<09:45, 16.74it/s]\n",
      "Loss: 2.413 :   2%|▏         | 205/10000 [00:12<09:46, 16.70it/s]\n",
      "Loss: 2.413 :   2%|▏         | 207/10000 [00:12<09:38, 16.94it/s]\n",
      "Loss: 2.413 :   2%|▏         | 209/10000 [00:12<09:34, 17.05it/s]\n",
      "Loss: 2.413 :   2%|▏         | 211/10000 [00:12<09:17, 17.57it/s]\n",
      "Loss: 2.413 :   2%|▏         | 213/10000 [00:12<09:05, 17.94it/s]\n",
      "Loss: 2.413 :   2%|▏         | 215/10000 [00:12<09:09, 17.80it/s]\n",
      "Loss: 2.413 :   2%|▏         | 217/10000 [00:12<09:05, 17.93it/s]\n",
      "Loss: 2.413 :   2%|▏         | 219/10000 [00:12<09:04, 17.95it/s]\n",
      "Loss: 2.413 :   2%|▏         | 221/10000 [00:13<09:14, 17.63it/s]\n",
      "Loss: 2.413 :   2%|▏         | 223/10000 [00:13<09:08, 17.83it/s]\n",
      "Loss: 2.413 :   2%|▏         | 225/10000 [00:13<09:11, 17.72it/s]\n",
      "Loss: 2.413 :   2%|▏         | 227/10000 [00:13<10:02, 16.22it/s]\n",
      "Loss: 2.413 :   2%|▏         | 229/10000 [00:13<10:43, 15.18it/s]\n",
      "Loss: 2.413 :   2%|▏         | 231/10000 [00:13<10:55, 14.91it/s]\n",
      "Loss: 2.413 :   2%|▏         | 233/10000 [00:13<10:37, 15.32it/s]\n",
      "Loss: 2.413 :   2%|▏         | 235/10000 [00:13<10:23, 15.66it/s]\n",
      "Loss: 2.413 :   2%|▏         | 237/10000 [00:14<10:03, 16.17it/s]\n",
      "Loss: 2.413 :   2%|▏         | 239/10000 [00:14<09:38, 16.87it/s]\n",
      "Loss: 2.413 :   2%|▏         | 241/10000 [00:14<09:32, 17.06it/s]\n",
      "Loss: 2.413 :   2%|▏         | 243/10000 [00:14<09:29, 17.13it/s]\n",
      "Loss: 2.413 :   2%|▏         | 245/10000 [00:14<09:29, 17.14it/s]\n",
      "Loss: 2.413 :   2%|▏         | 247/10000 [00:14<09:24, 17.29it/s]\n",
      "Loss: 2.413 :   2%|▏         | 249/10000 [00:14<09:29, 17.12it/s]\n",
      "Loss: 2.516 :   2%|▏         | 249/10000 [00:14<09:29, 17.12it/s]\n",
      "Loss: 2.516 :   3%|▎         | 251/10000 [00:14<10:12, 15.92it/s]\n",
      "Loss: 2.516 :   3%|▎         | 253/10000 [00:14<09:44, 16.68it/s]\n",
      "Loss: 2.516 :   3%|▎         | 255/10000 [00:15<09:33, 16.99it/s]\n",
      "Loss: 2.516 :   3%|▎         | 257/10000 [00:15<09:20, 17.38it/s]\n",
      "Loss: 2.516 :   3%|▎         | 259/10000 [00:15<09:13, 17.61it/s]\n",
      "Loss: 2.516 :   3%|▎         | 261/10000 [00:15<09:08, 17.77it/s]\n",
      "Loss: 2.516 :   3%|▎         | 263/10000 [00:15<09:13, 17.59it/s]\n",
      "Loss: 2.516 :   3%|▎         | 265/10000 [00:15<09:08, 17.75it/s]\n",
      "Loss: 2.516 :   3%|▎         | 267/10000 [00:15<09:06, 17.80it/s]\n",
      "Loss: 2.516 :   3%|▎         | 269/10000 [00:15<09:04, 17.87it/s]\n",
      "Loss: 2.516 :   3%|▎         | 271/10000 [00:15<08:56, 18.13it/s]\n",
      "Loss: 2.516 :   3%|▎         | 273/10000 [00:16<08:54, 18.20it/s]\n",
      "Loss: 2.516 :   3%|▎         | 275/10000 [00:16<08:54, 18.18it/s]\n",
      "Loss: 2.516 :   3%|▎         | 277/10000 [00:16<08:49, 18.35it/s]\n",
      "Loss: 2.516 :   3%|▎         | 279/10000 [00:16<08:52, 18.27it/s]\n",
      "Loss: 2.516 :   3%|▎         | 281/10000 [00:16<08:47, 18.43it/s]\n",
      "Loss: 2.516 :   3%|▎         | 283/10000 [00:16<09:02, 17.90it/s]\n",
      "Loss: 2.516 :   3%|▎         | 285/10000 [00:16<08:58, 18.05it/s]\n",
      "Loss: 2.516 :   3%|▎         | 287/10000 [00:16<08:59, 18.00it/s]\n",
      "Loss: 2.516 :   3%|▎         | 289/10000 [00:16<08:55, 18.13it/s]\n",
      "Loss: 2.516 :   3%|▎         | 291/10000 [00:17<08:50, 18.31it/s]\n",
      "Loss: 2.516 :   3%|▎         | 293/10000 [00:17<08:51, 18.26it/s]\n",
      "Loss: 2.516 :   3%|▎         | 295/10000 [00:17<08:58, 18.02it/s]\n",
      "Loss: 2.516 :   3%|▎         | 297/10000 [00:17<09:13, 17.54it/s]\n",
      "Loss: 2.516 :   3%|▎         | 299/10000 [00:17<09:13, 17.54it/s]\n",
      "Loss: 2.471 :   3%|▎         | 299/10000 [00:17<09:13, 17.54it/s]\n",
      "Loss: 2.471 :   3%|▎         | 301/10000 [00:17<09:59, 16.18it/s]\n",
      "Loss: 2.471 :   3%|▎         | 303/10000 [00:17<09:38, 16.78it/s]\n",
      "Loss: 2.471 :   3%|▎         | 305/10000 [00:17<09:12, 17.56it/s]\n",
      "Loss: 2.471 :   3%|▎         | 307/10000 [00:18<09:15, 17.46it/s]\n",
      "Loss: 2.471 :   3%|▎         | 309/10000 [00:18<09:03, 17.82it/s]\n",
      "Loss: 2.471 :   3%|▎         | 311/10000 [00:18<09:02, 17.85it/s]\n",
      "Loss: 2.471 :   3%|▎         | 313/10000 [00:18<09:03, 17.81it/s]\n",
      "Loss: 2.471 :   3%|▎         | 315/10000 [00:18<08:47, 18.35it/s]\n",
      "Loss: 2.471 :   3%|▎         | 317/10000 [00:18<08:57, 18.03it/s]\n",
      "Loss: 2.471 :   3%|▎         | 319/10000 [00:18<09:15, 17.43it/s]\n",
      "Loss: 2.471 :   3%|▎         | 321/10000 [00:18<10:08, 15.90it/s]\n",
      "Loss: 2.471 :   3%|▎         | 323/10000 [00:18<10:19, 15.62it/s]\n",
      "Loss: 2.471 :   3%|▎         | 325/10000 [00:19<10:08, 15.90it/s]\n",
      "Loss: 2.471 :   3%|▎         | 327/10000 [00:19<09:44, 16.54it/s]\n",
      "Loss: 2.471 :   3%|▎         | 329/10000 [00:19<09:49, 16.41it/s]\n",
      "Loss: 2.471 :   3%|▎         | 331/10000 [00:19<09:39, 16.70it/s]\n",
      "Loss: 2.471 :   3%|▎         | 333/10000 [00:19<09:23, 17.17it/s]\n",
      "Loss: 2.471 :   3%|▎         | 335/10000 [00:19<09:17, 17.35it/s]\n",
      "Loss: 2.471 :   3%|▎         | 337/10000 [00:19<09:14, 17.42it/s]\n",
      "Loss: 2.471 :   3%|▎         | 339/10000 [00:19<09:07, 17.65it/s]\n",
      "Loss: 2.471 :   3%|▎         | 341/10000 [00:19<09:02, 17.82it/s]\n",
      "Loss: 2.471 :   3%|▎         | 343/10000 [00:20<09:04, 17.74it/s]\n",
      "Loss: 2.471 :   3%|▎         | 345/10000 [00:20<09:01, 17.82it/s]\n",
      "Loss: 2.471 :   3%|▎         | 347/10000 [00:20<09:16, 17.36it/s]\n",
      "Loss: 2.471 :   3%|▎         | 349/10000 [00:20<09:10, 17.53it/s]\n",
      "Loss: 2.372 :   3%|▎         | 349/10000 [00:20<09:10, 17.53it/s]\n",
      "Loss: 2.372 :   4%|▎         | 351/10000 [00:20<10:01, 16.04it/s]\n",
      "Loss: 2.372 :   4%|▎         | 353/10000 [00:20<09:38, 16.68it/s]\n",
      "Loss: 2.372 :   4%|▎         | 355/10000 [00:20<09:22, 17.15it/s]\n",
      "Loss: 2.372 :   4%|▎         | 357/10000 [00:20<09:16, 17.31it/s]\n",
      "Loss: 2.372 :   4%|▎         | 359/10000 [00:21<09:01, 17.80it/s]\n",
      "Loss: 2.372 :   4%|▎         | 361/10000 [00:21<09:03, 17.74it/s]\n",
      "Loss: 2.372 :   4%|▎         | 363/10000 [00:21<08:56, 17.96it/s]\n",
      "Loss: 2.372 :   4%|▎         | 365/10000 [00:21<08:54, 18.02it/s]\n",
      "Loss: 2.372 :   4%|▎         | 367/10000 [00:21<08:59, 17.87it/s]\n",
      "Loss: 2.372 :   4%|▎         | 369/10000 [00:21<08:54, 18.01it/s]\n",
      "Loss: 2.372 :   4%|▎         | 371/10000 [00:21<09:04, 17.69it/s]\n",
      "Loss: 2.372 :   4%|▎         | 373/10000 [00:21<09:05, 17.65it/s]\n",
      "Loss: 2.372 :   4%|▍         | 375/10000 [00:21<08:59, 17.83it/s]\n",
      "Loss: 2.372 :   4%|▍         | 377/10000 [00:22<09:01, 17.77it/s]\n",
      "Loss: 2.372 :   4%|▍         | 379/10000 [00:22<09:01, 17.78it/s]\n",
      "Loss: 2.372 :   4%|▍         | 381/10000 [00:22<08:55, 17.95it/s]\n",
      "Loss: 2.372 :   4%|▍         | 383/10000 [00:22<08:55, 17.97it/s]\n",
      "Loss: 2.372 :   4%|▍         | 385/10000 [00:22<08:52, 18.07it/s]\n",
      "Loss: 2.372 :   4%|▍         | 387/10000 [00:22<08:43, 18.35it/s]\n",
      "Loss: 2.372 :   4%|▍         | 389/10000 [00:22<08:42, 18.40it/s]\n",
      "Loss: 2.372 :   4%|▍         | 391/10000 [00:22<08:40, 18.47it/s]\n",
      "Loss: 2.372 :   4%|▍         | 393/10000 [00:22<08:42, 18.37it/s]\n",
      "Loss: 2.372 :   4%|▍         | 395/10000 [00:23<08:40, 18.45it/s]\n",
      "Loss: 2.372 :   4%|▍         | 397/10000 [00:23<08:40, 18.44it/s]\n",
      "Loss: 2.372 :   4%|▍         | 399/10000 [00:23<08:36, 18.59it/s]\n",
      "Loss: 2.491 :   4%|▍         | 399/10000 [00:23<08:36, 18.59it/s]\n",
      "Loss: 2.491 :   4%|▍         | 401/10000 [00:23<09:47, 16.35it/s]\n",
      "Loss: 2.491 :   4%|▍         | 403/10000 [00:23<09:22, 17.06it/s]\n",
      "Loss: 2.491 :   4%|▍         | 405/10000 [00:23<09:17, 17.21it/s]\n",
      "Loss: 2.491 :   4%|▍         | 407/10000 [00:23<09:00, 17.76it/s]\n",
      "Loss: 2.491 :   4%|▍         | 409/10000 [00:23<08:48, 18.16it/s]\n",
      "Loss: 2.491 :   4%|▍         | 411/10000 [00:23<09:03, 17.65it/s]\n",
      "Loss: 2.491 :   4%|▍         | 413/10000 [00:24<09:51, 16.21it/s]\n",
      "Loss: 2.491 :   4%|▍         | 415/10000 [00:24<10:14, 15.61it/s]\n",
      "Loss: 2.491 :   4%|▍         | 417/10000 [00:24<10:10, 15.70it/s]\n",
      "Loss: 2.491 :   4%|▍         | 419/10000 [00:24<09:55, 16.09it/s]\n",
      "Loss: 2.491 :   4%|▍         | 421/10000 [00:24<09:46, 16.34it/s]\n",
      "Loss: 2.491 :   4%|▍         | 423/10000 [00:24<09:29, 16.82it/s]\n",
      "Loss: 2.491 :   4%|▍         | 425/10000 [00:24<09:26, 16.91it/s]\n",
      "Loss: 2.491 :   4%|▍         | 427/10000 [00:24<09:25, 16.94it/s]\n",
      "Loss: 2.491 :   4%|▍         | 429/10000 [00:25<09:28, 16.85it/s]\n",
      "Loss: 2.491 :   4%|▍         | 431/10000 [00:25<09:24, 16.95it/s]\n",
      "Loss: 2.491 :   4%|▍         | 433/10000 [00:25<09:16, 17.20it/s]\n",
      "Loss: 2.491 :   4%|▍         | 435/10000 [00:25<09:13, 17.28it/s]\n",
      "Loss: 2.491 :   4%|▍         | 437/10000 [00:25<09:05, 17.52it/s]\n",
      "Loss: 2.491 :   4%|▍         | 439/10000 [00:25<09:05, 17.53it/s]\n",
      "Loss: 2.491 :   4%|▍         | 441/10000 [00:25<09:07, 17.46it/s]\n",
      "Loss: 2.491 :   4%|▍         | 443/10000 [00:25<09:05, 17.51it/s]\n",
      "Loss: 2.491 :   4%|▍         | 445/10000 [00:25<08:56, 17.81it/s]\n",
      "Loss: 2.491 :   4%|▍         | 447/10000 [00:26<08:54, 17.88it/s]\n",
      "Loss: 2.491 :   4%|▍         | 449/10000 [00:26<08:55, 17.84it/s]\n",
      "Loss: 2.406 :   4%|▍         | 449/10000 [00:26<08:55, 17.84it/s]\n",
      "Loss: 2.406 :   5%|▍         | 451/10000 [00:26<09:50, 16.16it/s]\n",
      "Loss: 2.406 :   5%|▍         | 453/10000 [00:26<09:34, 16.61it/s]\n",
      "Loss: 2.406 :   5%|▍         | 455/10000 [00:26<09:13, 17.25it/s]\n",
      "Loss: 2.406 :   5%|▍         | 457/10000 [00:26<08:58, 17.73it/s]\n",
      "Loss: 2.406 :   5%|▍         | 459/10000 [00:26<08:42, 18.25it/s]\n",
      "Loss: 2.406 :   5%|▍         | 461/10000 [00:26<08:45, 18.15it/s]\n",
      "Loss: 2.406 :   5%|▍         | 463/10000 [00:26<08:39, 18.34it/s]\n",
      "Loss: 2.406 :   5%|▍         | 465/10000 [00:27<08:49, 18.00it/s]\n",
      "Loss: 2.406 :   5%|▍         | 467/10000 [00:27<08:48, 18.03it/s]\n",
      "Loss: 2.406 :   5%|▍         | 469/10000 [00:27<08:46, 18.10it/s]\n",
      "Loss: 2.406 :   5%|▍         | 471/10000 [00:27<08:44, 18.16it/s]\n",
      "Loss: 2.406 :   5%|▍         | 473/10000 [00:27<08:47, 18.06it/s]\n",
      "Loss: 2.406 :   5%|▍         | 475/10000 [00:27<09:00, 17.62it/s]\n",
      "Loss: 2.406 :   5%|▍         | 477/10000 [00:27<08:56, 17.75it/s]\n",
      "Loss: 2.406 :   5%|▍         | 479/10000 [00:27<08:42, 18.23it/s]\n",
      "Loss: 2.406 :   5%|▍         | 481/10000 [00:27<08:33, 18.55it/s]\n",
      "Loss: 2.406 :   5%|▍         | 483/10000 [00:28<08:33, 18.54it/s]\n",
      "Loss: 2.406 :   5%|▍         | 485/10000 [00:28<08:31, 18.59it/s]\n",
      "Loss: 2.406 :   5%|▍         | 487/10000 [00:28<08:37, 18.40it/s]\n",
      "Loss: 2.406 :   5%|▍         | 489/10000 [00:28<08:44, 18.12it/s]\n",
      "Loss: 2.406 :   5%|▍         | 491/10000 [00:28<08:51, 17.88it/s]\n",
      "Loss: 2.406 :   5%|▍         | 493/10000 [00:28<08:49, 17.96it/s]\n",
      "Loss: 2.406 :   5%|▍         | 495/10000 [00:28<08:54, 17.79it/s]\n",
      "Loss: 2.406 :   5%|▍         | 497/10000 [00:28<08:47, 18.02it/s]\n",
      "Loss: 2.406 :   5%|▍         | 499/10000 [00:28<08:43, 18.13it/s]\n",
      "Loss: 2.245 :   5%|▍         | 499/10000 [00:29<08:43, 18.13it/s]\n",
      "Loss: 2.245 :   5%|▌         | 501/10000 [00:29<09:32, 16.58it/s]\n",
      "Loss: 2.245 :   5%|▌         | 503/10000 [00:29<09:21, 16.92it/s]\n",
      "Loss: 2.245 :   5%|▌         | 505/10000 [00:29<10:11, 15.52it/s]\n",
      "Loss: 2.245 :   5%|▌         | 507/10000 [00:29<10:25, 15.17it/s]\n",
      "Loss: 2.245 :   5%|▌         | 509/10000 [00:29<10:20, 15.30it/s]\n",
      "Loss: 2.245 :   5%|▌         | 511/10000 [00:29<10:05, 15.67it/s]\n",
      "Loss: 2.245 :   5%|▌         | 513/10000 [00:29<09:57, 15.89it/s]\n",
      "Loss: 2.245 :   5%|▌         | 515/10000 [00:30<09:58, 15.85it/s]\n",
      "Loss: 2.245 :   5%|▌         | 517/10000 [00:30<09:47, 16.13it/s]\n",
      "Loss: 2.245 :   5%|▌         | 519/10000 [00:30<09:39, 16.35it/s]\n",
      "Loss: 2.245 :   5%|▌         | 521/10000 [00:30<09:36, 16.44it/s]\n",
      "Loss: 2.245 :   5%|▌         | 523/10000 [00:30<09:18, 16.98it/s]\n",
      "Loss: 2.245 :   5%|▌         | 525/10000 [00:30<09:09, 17.24it/s]\n",
      "Loss: 2.245 :   5%|▌         | 527/10000 [00:30<09:04, 17.40it/s]\n",
      "Loss: 2.245 :   5%|▌         | 529/10000 [00:30<09:10, 17.20it/s]\n",
      "Loss: 2.245 :   5%|▌         | 531/10000 [00:30<09:10, 17.20it/s]\n",
      "Loss: 2.245 :   5%|▌         | 533/10000 [00:31<09:12, 17.13it/s]\n",
      "Loss: 2.245 :   5%|▌         | 535/10000 [00:31<09:06, 17.32it/s]\n",
      "Loss: 2.245 :   5%|▌         | 537/10000 [00:31<08:58, 17.59it/s]\n",
      "Loss: 2.245 :   5%|▌         | 539/10000 [00:31<08:52, 17.77it/s]\n",
      "Loss: 2.245 :   5%|▌         | 541/10000 [00:31<08:42, 18.11it/s]\n",
      "Loss: 2.245 :   5%|▌         | 543/10000 [00:31<08:47, 17.93it/s]\n",
      "Loss: 2.245 :   5%|▌         | 545/10000 [00:31<08:40, 18.17it/s]\n",
      "Loss: 2.245 :   5%|▌         | 547/10000 [00:31<08:33, 18.40it/s]\n",
      "Loss: 2.245 :   5%|▌         | 549/10000 [00:31<08:49, 17.86it/s]\n",
      "Loss: 2.609 :   5%|▌         | 549/10000 [00:32<08:49, 17.86it/s]\n",
      "Loss: 2.609 :   6%|▌         | 551/10000 [00:32<09:47, 16.08it/s]\n",
      "Loss: 2.609 :   6%|▌         | 553/10000 [00:32<09:37, 16.35it/s]\n",
      "Loss: 2.609 :   6%|▌         | 555/10000 [00:32<09:19, 16.89it/s]\n",
      "Loss: 2.609 :   6%|▌         | 557/10000 [00:32<09:12, 17.08it/s]\n",
      "Loss: 2.609 :   6%|▌         | 559/10000 [00:32<09:03, 17.37it/s]\n",
      "Loss: 2.609 :   6%|▌         | 561/10000 [00:32<08:53, 17.70it/s]\n",
      "Loss: 2.609 :   6%|▌         | 563/10000 [00:32<08:56, 17.60it/s]\n",
      "Loss: 2.609 :   6%|▌         | 565/10000 [00:32<08:49, 17.82it/s]\n",
      "Loss: 2.609 :   6%|▌         | 567/10000 [00:32<08:47, 17.87it/s]\n",
      "Loss: 2.609 :   6%|▌         | 569/10000 [00:33<09:04, 17.31it/s]\n",
      "Loss: 2.609 :   6%|▌         | 571/10000 [00:33<09:15, 16.97it/s]\n",
      "Loss: 2.609 :   6%|▌         | 573/10000 [00:33<09:22, 16.76it/s]\n",
      "Loss: 2.609 :   6%|▌         | 575/10000 [00:33<09:40, 16.24it/s]\n",
      "Loss: 2.609 :   6%|▌         | 577/10000 [00:33<09:38, 16.30it/s]\n",
      "Loss: 2.609 :   6%|▌         | 579/10000 [00:33<09:33, 16.41it/s]\n",
      "Loss: 2.609 :   6%|▌         | 581/10000 [00:33<09:38, 16.28it/s]\n",
      "Loss: 2.609 :   6%|▌         | 583/10000 [00:33<09:51, 15.91it/s]\n",
      "Loss: 2.609 :   6%|▌         | 585/10000 [00:34<09:51, 15.91it/s]\n",
      "Loss: 2.609 :   6%|▌         | 587/10000 [00:34<09:56, 15.78it/s]\n",
      "Loss: 2.609 :   6%|▌         | 589/10000 [00:34<10:11, 15.40it/s]\n",
      "Loss: 2.609 :   6%|▌         | 591/10000 [00:34<10:00, 15.68it/s]\n",
      "Loss: 2.609 :   6%|▌         | 593/10000 [00:34<10:52, 14.42it/s]\n",
      "Loss: 2.609 :   6%|▌         | 595/10000 [00:34<11:45, 13.33it/s]\n",
      "Loss: 2.609 :   6%|▌         | 597/10000 [00:34<11:57, 13.10it/s]\n",
      "Loss: 2.609 :   6%|▌         | 599/10000 [00:35<12:06, 12.94it/s]\n",
      "Loss: 2.426 :   6%|▌         | 599/10000 [00:35<12:06, 12.94it/s]\n",
      "Loss: 2.426 :   6%|▌         | 601/10000 [00:35<13:27, 11.64it/s]\n",
      "Loss: 2.426 :   6%|▌         | 603/10000 [00:35<12:41, 12.33it/s]\n",
      "Loss: 2.426 :   6%|▌         | 605/10000 [00:35<11:54, 13.16it/s]\n",
      "Loss: 2.426 :   6%|▌         | 607/10000 [00:35<11:25, 13.70it/s]\n",
      "Loss: 2.426 :   6%|▌         | 609/10000 [00:35<10:44, 14.57it/s]\n",
      "Loss: 2.426 :   6%|▌         | 611/10000 [00:36<10:12, 15.32it/s]\n",
      "Loss: 2.426 :   6%|▌         | 613/10000 [00:36<09:46, 16.01it/s]\n",
      "Loss: 2.426 :   6%|▌         | 615/10000 [00:36<09:28, 16.51it/s]\n",
      "Loss: 2.426 :   6%|▌         | 617/10000 [00:36<09:21, 16.72it/s]\n",
      "Loss: 2.426 :   6%|▌         | 619/10000 [00:36<09:03, 17.25it/s]\n",
      "Loss: 2.426 :   6%|▌         | 621/10000 [00:36<08:55, 17.51it/s]\n",
      "Loss: 2.426 :   6%|▌         | 623/10000 [00:36<08:55, 17.52it/s]\n",
      "Loss: 2.426 :   6%|▋         | 625/10000 [00:36<08:45, 17.85it/s]\n",
      "Loss: 2.426 :   6%|▋         | 627/10000 [00:36<08:41, 17.97it/s]\n",
      "Loss: 2.426 :   6%|▋         | 629/10000 [00:37<08:44, 17.86it/s]\n",
      "Loss: 2.426 :   6%|▋         | 631/10000 [00:37<08:35, 18.17it/s]\n",
      "Loss: 2.426 :   6%|▋         | 633/10000 [00:37<08:33, 18.26it/s]\n",
      "Loss: 2.426 :   6%|▋         | 635/10000 [00:37<08:31, 18.29it/s]\n",
      "Loss: 2.426 :   6%|▋         | 637/10000 [00:37<08:25, 18.51it/s]\n",
      "Loss: 2.426 :   6%|▋         | 639/10000 [00:37<08:28, 18.41it/s]\n",
      "Loss: 2.426 :   6%|▋         | 641/10000 [00:37<08:22, 18.61it/s]\n",
      "Loss: 2.426 :   6%|▋         | 643/10000 [00:37<08:18, 18.78it/s]\n",
      "Loss: 2.426 :   6%|▋         | 645/10000 [00:37<08:27, 18.45it/s]\n",
      "Loss: 2.426 :   6%|▋         | 647/10000 [00:37<08:19, 18.71it/s]\n",
      "Loss: 2.426 :   6%|▋         | 649/10000 [00:38<08:15, 18.88it/s]\n",
      "Loss: 2.480 :   6%|▋         | 649/10000 [00:38<08:15, 18.88it/s]\n",
      "Loss: 2.480 :   7%|▋         | 651/10000 [00:38<09:07, 17.07it/s]\n",
      "Loss: 2.480 :   7%|▋         | 653/10000 [00:38<08:55, 17.44it/s]\n",
      "Loss: 2.480 :   7%|▋         | 655/10000 [00:38<08:44, 17.81it/s]\n",
      "Loss: 2.480 :   7%|▋         | 657/10000 [00:38<08:40, 17.94it/s]\n",
      "Loss: 2.480 :   7%|▋         | 659/10000 [00:38<08:41, 17.90it/s]\n",
      "Loss: 2.480 :   7%|▋         | 661/10000 [00:38<08:39, 17.97it/s]\n",
      "Loss: 2.480 :   7%|▋         | 663/10000 [00:38<08:32, 18.23it/s]\n",
      "Loss: 2.480 :   7%|▋         | 665/10000 [00:38<08:25, 18.48it/s]\n",
      "Loss: 2.480 :   7%|▋         | 667/10000 [00:39<08:24, 18.49it/s]\n",
      "Loss: 2.480 :   7%|▋         | 669/10000 [00:39<08:20, 18.66it/s]\n",
      "Loss: 2.480 :   7%|▋         | 671/10000 [00:39<08:20, 18.66it/s]\n",
      "Loss: 2.480 :   7%|▋         | 673/10000 [00:39<08:12, 18.95it/s]\n",
      "Loss: 2.480 :   7%|▋         | 675/10000 [00:39<08:09, 19.05it/s]\n",
      "Loss: 2.480 :   7%|▋         | 677/10000 [00:39<08:11, 18.95it/s]\n",
      "Loss: 2.480 :   7%|▋         | 679/10000 [00:39<08:16, 18.77it/s]\n",
      "Loss: 2.480 :   7%|▋         | 681/10000 [00:39<08:29, 18.30it/s]\n",
      "Loss: 2.480 :   7%|▋         | 683/10000 [00:39<08:43, 17.78it/s]\n",
      "Loss: 2.480 :   7%|▋         | 685/10000 [00:40<09:27, 16.42it/s]\n",
      "Loss: 2.480 :   7%|▋         | 687/10000 [00:40<09:48, 15.81it/s]\n",
      "Loss: 2.480 :   7%|▋         | 689/10000 [00:40<09:43, 15.96it/s]\n",
      "Loss: 2.480 :   7%|▋         | 691/10000 [00:40<09:26, 16.44it/s]\n",
      "Loss: 2.480 :   7%|▋         | 693/10000 [00:40<09:17, 16.68it/s]\n",
      "Loss: 2.480 :   7%|▋         | 695/10000 [00:40<09:08, 16.95it/s]\n",
      "Loss: 2.480 :   7%|▋         | 697/10000 [00:40<09:14, 16.77it/s]\n",
      "Loss: 2.480 :   7%|▋         | 699/10000 [00:40<09:04, 17.07it/s]\n",
      "Loss: 2.417 :   7%|▋         | 699/10000 [00:41<09:04, 17.07it/s]\n",
      "Loss: 2.417 :   7%|▋         | 701/10000 [00:41<09:51, 15.73it/s]\n",
      "Loss: 2.417 :   7%|▋         | 703/10000 [00:41<09:20, 16.57it/s]\n",
      "Loss: 2.417 :   7%|▋         | 705/10000 [00:41<09:04, 17.08it/s]\n",
      "Loss: 2.417 :   7%|▋         | 707/10000 [00:41<09:01, 17.17it/s]\n",
      "Loss: 2.417 :   7%|▋         | 709/10000 [00:41<08:56, 17.31it/s]\n",
      "Loss: 2.417 :   7%|▋         | 711/10000 [00:41<08:46, 17.65it/s]\n",
      "Loss: 2.417 :   7%|▋         | 713/10000 [00:41<08:49, 17.54it/s]\n",
      "Loss: 2.417 :   7%|▋         | 715/10000 [00:41<08:43, 17.73it/s]\n",
      "Loss: 2.417 :   7%|▋         | 717/10000 [00:41<08:35, 17.99it/s]\n",
      "Loss: 2.417 :   7%|▋         | 719/10000 [00:42<08:29, 18.23it/s]\n",
      "Loss: 2.417 :   7%|▋         | 721/10000 [00:42<08:19, 18.58it/s]\n",
      "Loss: 2.417 :   7%|▋         | 723/10000 [00:42<08:19, 18.57it/s]\n",
      "Loss: 2.417 :   7%|▋         | 725/10000 [00:42<08:13, 18.78it/s]\n",
      "Loss: 2.417 :   7%|▋         | 727/10000 [00:42<08:16, 18.68it/s]\n",
      "Loss: 2.417 :   7%|▋         | 729/10000 [00:42<08:19, 18.56it/s]\n",
      "Loss: 2.417 :   7%|▋         | 731/10000 [00:42<08:16, 18.69it/s]\n",
      "Loss: 2.417 :   7%|▋         | 733/10000 [00:42<08:17, 18.64it/s]\n",
      "Loss: 2.417 :   7%|▋         | 735/10000 [00:42<08:21, 18.48it/s]\n",
      "Loss: 2.417 :   7%|▋         | 737/10000 [00:43<08:20, 18.52it/s]\n",
      "Loss: 2.417 :   7%|▋         | 739/10000 [00:43<08:14, 18.72it/s]\n",
      "Loss: 2.417 :   7%|▋         | 741/10000 [00:43<08:18, 18.59it/s]\n",
      "Loss: 2.417 :   7%|▋         | 743/10000 [00:43<08:17, 18.61it/s]\n",
      "Loss: 2.417 :   7%|▋         | 745/10000 [00:43<08:23, 18.37it/s]\n",
      "Loss: 2.417 :   7%|▋         | 747/10000 [00:43<08:23, 18.37it/s]\n",
      "Loss: 2.417 :   7%|▋         | 749/10000 [00:43<08:33, 18.01it/s]\n",
      "Loss: 2.249 :   7%|▋         | 749/10000 [00:43<08:33, 18.01it/s]\n",
      "Loss: 2.249 :   8%|▊         | 751/10000 [00:43<09:42, 15.87it/s]\n",
      "Loss: 2.249 :   8%|▊         | 753/10000 [00:43<09:28, 16.27it/s]\n",
      "Loss: 2.249 :   8%|▊         | 755/10000 [00:44<09:23, 16.40it/s]\n",
      "Loss: 2.249 :   8%|▊         | 757/10000 [00:44<09:39, 15.96it/s]\n",
      "Loss: 2.249 :   8%|▊         | 759/10000 [00:44<10:11, 15.12it/s]\n",
      "Loss: 2.249 :   8%|▊         | 761/10000 [00:44<10:21, 14.85it/s]\n",
      "Loss: 2.249 :   8%|▊         | 763/10000 [00:44<10:36, 14.52it/s]\n",
      "Loss: 2.249 :   8%|▊         | 765/10000 [00:44<10:48, 14.23it/s]\n",
      "Loss: 2.249 :   8%|▊         | 767/10000 [00:44<11:14, 13.68it/s]\n",
      "Loss: 2.249 :   8%|▊         | 769/10000 [00:45<11:41, 13.17it/s]\n",
      "Loss: 2.249 :   8%|▊         | 771/10000 [00:45<12:23, 12.41it/s]\n",
      "Loss: 2.249 :   8%|▊         | 773/10000 [00:45<13:10, 11.68it/s]\n",
      "Loss: 2.249 :   8%|▊         | 775/10000 [00:45<13:38, 11.26it/s]\n",
      "Loss: 2.249 :   8%|▊         | 777/10000 [00:45<13:20, 11.52it/s]\n",
      "Loss: 2.249 :   8%|▊         | 779/10000 [00:46<13:08, 11.70it/s]\n",
      "Loss: 2.249 :   8%|▊         | 781/10000 [00:46<12:54, 11.91it/s]\n",
      "Loss: 2.249 :   8%|▊         | 783/10000 [00:46<12:38, 12.16it/s]\n",
      "Loss: 2.249 :   8%|▊         | 785/10000 [00:46<13:24, 11.45it/s]\n",
      "Loss: 2.249 :   8%|▊         | 787/10000 [00:46<12:25, 12.36it/s]\n",
      "Loss: 2.249 :   8%|▊         | 789/10000 [00:46<11:35, 13.24it/s]\n",
      "Loss: 2.249 :   8%|▊         | 791/10000 [00:46<11:08, 13.78it/s]\n",
      "Loss: 2.249 :   8%|▊         | 793/10000 [00:47<10:41, 14.35it/s]\n",
      "Loss: 2.249 :   8%|▊         | 795/10000 [00:47<10:11, 15.06it/s]\n",
      "Loss: 2.249 :   8%|▊         | 797/10000 [00:47<09:51, 15.55it/s]\n",
      "Loss: 2.249 :   8%|▊         | 799/10000 [00:47<09:40, 15.86it/s]\n",
      "Loss: 2.364 :   8%|▊         | 799/10000 [00:47<09:40, 15.86it/s]\n",
      "Loss: 2.364 :   8%|▊         | 801/10000 [00:47<10:29, 14.61it/s]\n",
      "Loss: 2.364 :   8%|▊         | 803/10000 [00:47<10:03, 15.25it/s]\n",
      "Loss: 2.364 :   8%|▊         | 805/10000 [00:47<09:39, 15.87it/s]\n",
      "Loss: 2.364 :   8%|▊         | 807/10000 [00:47<09:34, 15.99it/s]\n",
      "Loss: 2.364 :   8%|▊         | 809/10000 [00:48<09:27, 16.20it/s]\n",
      "Loss: 2.364 :   8%|▊         | 811/10000 [00:48<09:36, 15.94it/s]\n",
      "Loss: 2.364 :   8%|▊         | 813/10000 [00:48<09:39, 15.86it/s]\n",
      "Loss: 2.364 :   8%|▊         | 815/10000 [00:48<09:27, 16.19it/s]\n",
      "Loss: 2.364 :   8%|▊         | 817/10000 [00:48<09:14, 16.55it/s]\n",
      "Loss: 2.364 :   8%|▊         | 819/10000 [00:48<09:06, 16.80it/s]\n",
      "Loss: 2.364 :   8%|▊         | 821/10000 [00:48<09:03, 16.89it/s]\n",
      "Loss: 2.364 :   8%|▊         | 823/10000 [00:48<09:11, 16.65it/s]\n",
      "Loss: 2.364 :   8%|▊         | 825/10000 [00:49<09:01, 16.95it/s]\n",
      "Loss: 2.364 :   8%|▊         | 827/10000 [00:49<08:58, 17.05it/s]\n",
      "Loss: 2.364 :   8%|▊         | 829/10000 [00:49<08:49, 17.31it/s]\n",
      "Loss: 2.364 :   8%|▊         | 831/10000 [00:49<08:58, 17.04it/s]\n",
      "Loss: 2.364 :   8%|▊         | 833/10000 [00:49<08:45, 17.45it/s]\n",
      "Loss: 2.364 :   8%|▊         | 835/10000 [00:49<08:38, 17.67it/s]\n",
      "Loss: 2.364 :   8%|▊         | 837/10000 [00:49<08:49, 17.30it/s]\n",
      "Loss: 2.364 :   8%|▊         | 839/10000 [00:49<09:11, 16.62it/s]\n",
      "Loss: 2.364 :   8%|▊         | 841/10000 [00:50<10:34, 14.43it/s]\n",
      "Loss: 2.364 :   8%|▊         | 843/10000 [00:50<12:53, 11.83it/s]\n",
      "Loss: 2.364 :   8%|▊         | 845/10000 [00:50<11:58, 12.74it/s]\n",
      "Loss: 2.364 :   8%|▊         | 847/10000 [00:50<11:17, 13.52it/s]\n",
      "Loss: 2.364 :   8%|▊         | 849/10000 [00:50<10:36, 14.37it/s]\n",
      "Loss: 2.238 :   8%|▊         | 849/10000 [00:50<10:36, 14.37it/s]\n",
      "Loss: 2.238 :   9%|▊         | 851/10000 [00:50<12:23, 12.31it/s]\n",
      "Loss: 2.238 :   9%|▊         | 853/10000 [00:50<11:59, 12.72it/s]\n",
      "Loss: 2.238 :   9%|▊         | 855/10000 [00:51<11:08, 13.67it/s]\n",
      "Loss: 2.238 :   9%|▊         | 857/10000 [00:51<10:35, 14.39it/s]\n",
      "Loss: 2.238 :   9%|▊         | 859/10000 [00:51<10:18, 14.78it/s]\n",
      "Loss: 2.238 :   9%|▊         | 861/10000 [00:51<10:17, 14.80it/s]\n",
      "Loss: 2.238 :   9%|▊         | 863/10000 [00:51<10:20, 14.72it/s]\n",
      "Loss: 2.238 :   9%|▊         | 865/10000 [00:51<10:22, 14.67it/s]\n",
      "Loss: 2.238 :   9%|▊         | 867/10000 [00:51<10:30, 14.48it/s]\n",
      "Loss: 2.238 :   9%|▊         | 869/10000 [00:52<10:36, 14.35it/s]\n",
      "Loss: 2.238 :   9%|▊         | 871/10000 [00:52<10:50, 14.04it/s]\n",
      "Loss: 2.238 :   9%|▊         | 873/10000 [00:52<10:56, 13.91it/s]\n",
      "Loss: 2.238 :   9%|▉         | 875/10000 [00:52<10:37, 14.32it/s]\n",
      "Loss: 2.238 :   9%|▉         | 877/10000 [00:52<10:25, 14.58it/s]\n",
      "Loss: 2.238 :   9%|▉         | 879/10000 [00:52<10:11, 14.91it/s]\n",
      "Loss: 2.238 :   9%|▉         | 881/10000 [00:52<09:58, 15.24it/s]\n",
      "Loss: 2.238 :   9%|▉         | 883/10000 [00:52<09:54, 15.33it/s]\n",
      "Loss: 2.238 :   9%|▉         | 885/10000 [00:53<09:45, 15.56it/s]\n",
      "Loss: 2.238 :   9%|▉         | 887/10000 [00:53<09:27, 16.05it/s]\n",
      "Loss: 2.238 :   9%|▉         | 889/10000 [00:53<09:28, 16.02it/s]\n",
      "Loss: 2.238 :   9%|▉         | 891/10000 [00:53<09:18, 16.32it/s]\n",
      "Loss: 2.238 :   9%|▉         | 893/10000 [00:53<09:02, 16.80it/s]\n",
      "Loss: 2.238 :   9%|▉         | 895/10000 [00:53<08:56, 16.99it/s]\n",
      "Loss: 2.238 :   9%|▉         | 897/10000 [00:53<08:54, 17.04it/s]\n",
      "Loss: 2.238 :   9%|▉         | 899/10000 [00:53<08:52, 17.08it/s]\n",
      "Loss: 2.400 :   9%|▉         | 899/10000 [00:54<08:52, 17.08it/s]\n",
      "Loss: 2.400 :   9%|▉         | 901/10000 [00:54<09:39, 15.70it/s]\n",
      "Loss: 2.400 :   9%|▉         | 903/10000 [00:54<09:20, 16.23it/s]\n",
      "Loss: 2.400 :   9%|▉         | 905/10000 [00:54<09:32, 15.89it/s]\n",
      "Loss: 2.400 :   9%|▉         | 907/10000 [00:54<09:48, 15.46it/s]\n",
      "Loss: 2.400 :   9%|▉         | 909/10000 [00:54<09:51, 15.37it/s]\n",
      "Loss: 2.400 :   9%|▉         | 911/10000 [00:54<09:51, 15.37it/s]\n",
      "Loss: 2.400 :   9%|▉         | 913/10000 [00:54<10:06, 14.99it/s]\n",
      "Loss: 2.400 :   9%|▉         | 915/10000 [00:54<10:04, 15.04it/s]\n",
      "Loss: 2.400 :   9%|▉         | 917/10000 [00:55<10:06, 14.97it/s]\n",
      "Loss: 2.400 :   9%|▉         | 919/10000 [00:55<10:11, 14.85it/s]\n",
      "Loss: 2.400 :   9%|▉         | 921/10000 [00:55<10:45, 14.07it/s]\n",
      "Loss: 2.400 :   9%|▉         | 923/10000 [00:55<15:33,  9.72it/s]\n",
      "Loss: 2.400 :   9%|▉         | 925/10000 [00:56<18:48,  8.04it/s]\n",
      "Loss: 2.400 :   9%|▉         | 926/10000 [00:56<22:49,  6.62it/s]\n",
      "Loss: 2.400 :   9%|▉         | 927/10000 [00:56<32:30,  4.65it/s]\n",
      "Loss: 2.400 :   9%|▉         | 928/10000 [00:57<35:32,  4.25it/s]\n",
      "Loss: 2.400 :   9%|▉         | 929/10000 [00:57<31:25,  4.81it/s]\n",
      "Loss: 2.400 :   9%|▉         | 930/10000 [00:57<27:29,  5.50it/s]\n",
      "Loss: 2.400 :   9%|▉         | 932/10000 [00:57<21:35,  7.00it/s]\n",
      "Loss: 2.400 :   9%|▉         | 934/10000 [00:57<17:53,  8.45it/s]\n",
      "Loss: 2.400 :   9%|▉         | 936/10000 [00:57<15:22,  9.83it/s]\n",
      "Loss: 2.400 :   9%|▉         | 938/10000 [00:58<13:37, 11.09it/s]\n",
      "Loss: 2.400 :   9%|▉         | 940/10000 [00:58<12:28, 12.11it/s]\n",
      "Loss: 2.400 :   9%|▉         | 942/10000 [00:58<11:34, 13.04it/s]\n",
      "Loss: 2.400 :   9%|▉         | 944/10000 [00:58<10:50, 13.93it/s]\n",
      "Loss: 2.400 :   9%|▉         | 946/10000 [00:58<10:24, 14.49it/s]\n",
      "Loss: 2.400 :   9%|▉         | 948/10000 [00:58<09:58, 15.12it/s]\n",
      "Loss: 2.400 :  10%|▉         | 950/10000 [00:58<09:53, 15.26it/s]\n",
      "Loss: 2.340 :  10%|▉         | 950/10000 [00:58<09:53, 15.26it/s]\n",
      "Loss: 2.340 :  10%|▉         | 952/10000 [00:58<10:45, 14.01it/s]\n",
      "Loss: 2.340 :  10%|▉         | 954/10000 [00:59<10:07, 14.90it/s]\n",
      "Loss: 2.340 :  10%|▉         | 956/10000 [00:59<09:35, 15.70it/s]\n",
      "Loss: 2.340 :  10%|▉         | 958/10000 [00:59<09:18, 16.18it/s]\n",
      "Loss: 2.340 :  10%|▉         | 960/10000 [00:59<09:13, 16.32it/s]\n",
      "Loss: 2.340 :  10%|▉         | 962/10000 [00:59<10:00, 15.04it/s]\n",
      "Loss: 2.340 :  10%|▉         | 964/10000 [00:59<10:07, 14.87it/s]\n",
      "Loss: 2.340 :  10%|▉         | 966/10000 [00:59<10:31, 14.30it/s]\n",
      "Loss: 2.340 :  10%|▉         | 968/10000 [01:00<11:05, 13.56it/s]\n",
      "Loss: 2.340 :  10%|▉         | 970/10000 [01:00<10:59, 13.70it/s]\n",
      "Loss: 2.340 :  10%|▉         | 972/10000 [01:00<11:27, 13.13it/s]\n",
      "Loss: 2.340 :  10%|▉         | 974/10000 [01:00<12:11, 12.34it/s]\n",
      "Loss: 2.340 :  10%|▉         | 976/10000 [01:00<12:25, 12.11it/s]\n",
      "Loss: 2.340 :  10%|▉         | 978/10000 [01:00<12:36, 11.93it/s]\n",
      "Loss: 2.340 :  10%|▉         | 980/10000 [01:01<12:58, 11.59it/s]\n",
      "Loss: 2.340 :  10%|▉         | 982/10000 [01:01<13:08, 11.44it/s]\n",
      "Loss: 2.340 :  10%|▉         | 984/10000 [01:01<13:32, 11.09it/s]\n",
      "Loss: 2.340 :  10%|▉         | 986/10000 [01:01<13:52, 10.83it/s]\n",
      "Loss: 2.340 :  10%|▉         | 988/10000 [01:01<13:43, 10.94it/s]\n",
      "Loss: 2.340 :  10%|▉         | 990/10000 [01:02<14:20, 10.48it/s]\n",
      "Loss: 2.340 :  10%|▉         | 992/10000 [01:02<14:49, 10.13it/s]\n",
      "Loss: 2.340 :  10%|▉         | 994/10000 [01:02<14:41, 10.21it/s]\n",
      "Loss: 2.340 :  10%|▉         | 996/10000 [01:02<14:37, 10.26it/s]\n",
      "Loss: 2.340 :  10%|▉         | 998/10000 [01:02<14:34, 10.29it/s]\n",
      "Loss: 2.340 :  10%|█         | 1000/10000 [01:02<14:42, 10.20it/s]\n",
      "Loss: 2.306 :  10%|█         | 1000/10000 [01:03<14:42, 10.20it/s]\n",
      "Loss: 2.306 :  10%|█         | 1002/10000 [01:03<15:53,  9.44it/s]\n",
      "Loss: 2.306 :  10%|█         | 1004/10000 [01:03<14:52, 10.08it/s]\n",
      "Loss: 2.306 :  10%|█         | 1006/10000 [01:03<14:20, 10.45it/s]\n",
      "Loss: 2.306 :  10%|█         | 1008/10000 [01:03<13:29, 11.11it/s]\n",
      "Loss: 2.306 :  10%|█         | 1010/10000 [01:03<12:44, 11.76it/s]\n",
      "Loss: 2.306 :  10%|█         | 1012/10000 [01:04<12:13, 12.26it/s]\n",
      "Loss: 2.306 :  10%|█         | 1014/10000 [01:04<11:38, 12.87it/s]\n",
      "Loss: 2.306 :  10%|█         | 1016/10000 [01:04<11:16, 13.28it/s]\n",
      "Loss: 2.306 :  10%|█         | 1018/10000 [01:04<11:04, 13.51it/s]\n",
      "Loss: 2.306 :  10%|█         | 1020/10000 [01:04<10:42, 13.99it/s]\n",
      "Loss: 2.306 :  10%|█         | 1022/10000 [01:04<10:36, 14.11it/s]\n",
      "Loss: 2.306 :  10%|█         | 1024/10000 [01:04<10:14, 14.60it/s]\n",
      "Loss: 2.306 :  10%|█         | 1026/10000 [01:04<10:09, 14.72it/s]\n",
      "Loss: 2.306 :  10%|█         | 1028/10000 [01:05<10:03, 14.86it/s]\n",
      "Loss: 2.306 :  10%|█         | 1030/10000 [01:05<09:58, 14.99it/s]\n",
      "Loss: 2.306 :  10%|█         | 1032/10000 [01:05<10:00, 14.93it/s]\n",
      "Loss: 2.306 :  10%|█         | 1034/10000 [01:05<09:57, 15.01it/s]\n",
      "Loss: 2.306 :  10%|█         | 1036/10000 [01:05<10:27, 14.28it/s]\n",
      "Loss: 2.306 :  10%|█         | 1038/10000 [01:05<11:03, 13.51it/s]\n",
      "Loss: 2.306 :  10%|█         | 1040/10000 [01:05<11:22, 13.14it/s]\n",
      "Loss: 2.306 :  10%|█         | 1042/10000 [01:06<11:16, 13.24it/s]\n",
      "Loss: 2.306 :  10%|█         | 1044/10000 [01:06<11:16, 13.25it/s]\n",
      "Loss: 2.306 :  10%|█         | 1046/10000 [01:06<11:02, 13.51it/s]\n",
      "Loss: 2.306 :  10%|█         | 1048/10000 [01:06<10:56, 13.63it/s]\n",
      "Loss: 2.306 :  10%|█         | 1050/10000 [01:06<11:14, 13.28it/s]\n",
      "Loss: 2.371 :  10%|█         | 1050/10000 [01:06<11:14, 13.28it/s]\n",
      "Loss: 2.371 :  11%|█         | 1052/10000 [01:06<12:26, 11.98it/s]\n",
      "Loss: 2.371 :  11%|█         | 1054/10000 [01:07<12:07, 12.30it/s]\n",
      "Loss: 2.371 :  11%|█         | 1056/10000 [01:07<11:57, 12.47it/s]\n",
      "Loss: 2.371 :  11%|█         | 1058/10000 [01:07<12:02, 12.37it/s]\n",
      "Loss: 2.371 :  11%|█         | 1060/10000 [01:07<12:02, 12.37it/s]\n",
      "Loss: 2.371 :  11%|█         | 1062/10000 [01:07<11:54, 12.51it/s]\n",
      "Loss: 2.371 :  11%|█         | 1064/10000 [01:07<12:06, 12.29it/s]\n",
      "Loss: 2.371 :  11%|█         | 1066/10000 [01:08<13:05, 11.37it/s]\n",
      "Loss: 2.371 :  11%|█         | 1068/10000 [01:08<13:23, 11.12it/s]\n",
      "Loss: 2.371 :  11%|█         | 1070/10000 [01:08<13:24, 11.10it/s]\n",
      "Loss: 2.371 :  11%|█         | 1072/10000 [01:08<13:44, 10.83it/s]\n",
      "Loss: 2.371 :  11%|█         | 1074/10000 [01:08<14:03, 10.58it/s]\n",
      "Loss: 2.371 :  11%|█         | 1076/10000 [01:09<13:43, 10.84it/s]\n",
      "Loss: 2.371 :  11%|█         | 1078/10000 [01:09<13:08, 11.31it/s]\n",
      "Loss: 2.371 :  11%|█         | 1080/10000 [01:09<12:28, 11.92it/s]\n",
      "Loss: 2.371 :  11%|█         | 1082/10000 [01:09<11:53, 12.50it/s]\n",
      "Loss: 2.371 :  11%|█         | 1084/10000 [01:09<11:32, 12.88it/s]\n",
      "Loss: 2.371 :  11%|█         | 1086/10000 [01:09<11:05, 13.39it/s]\n",
      "Loss: 2.371 :  11%|█         | 1088/10000 [01:09<10:49, 13.72it/s]\n",
      "Loss: 2.371 :  11%|█         | 1090/10000 [01:10<10:32, 14.09it/s]\n",
      "Loss: 2.371 :  11%|█         | 1092/10000 [01:10<10:18, 14.41it/s]\n",
      "Loss: 2.371 :  11%|█         | 1094/10000 [01:10<10:00, 14.84it/s]\n",
      "Loss: 2.371 :  11%|█         | 1096/10000 [01:10<10:03, 14.74it/s]\n",
      "Loss: 2.371 :  11%|█         | 1098/10000 [01:10<09:55, 14.94it/s]\n",
      "Loss: 2.371 :  11%|█         | 1100/10000 [01:10<09:44, 15.24it/s]\n",
      "Loss: 2.241 :  11%|█         | 1100/10000 [01:10<09:44, 15.24it/s]\n",
      "Loss: 2.241 :  11%|█         | 1102/10000 [01:10<10:33, 14.04it/s]\n",
      "Loss: 2.241 :  11%|█         | 1104/10000 [01:10<10:03, 14.75it/s]\n",
      "Loss: 2.241 :  11%|█         | 1106/10000 [01:11<09:53, 14.99it/s]\n",
      "Loss: 2.241 :  11%|█         | 1108/10000 [01:11<09:41, 15.30it/s]\n",
      "Loss: 2.241 :  11%|█         | 1110/10000 [01:11<09:37, 15.38it/s]\n",
      "Loss: 2.241 :  11%|█         | 1112/10000 [01:11<09:31, 15.55it/s]\n",
      "Loss: 2.241 :  11%|█         | 1114/10000 [01:11<09:22, 15.81it/s]\n",
      "Loss: 2.241 :  11%|█         | 1116/10000 [01:11<09:21, 15.83it/s]\n",
      "Loss: 2.241 :  11%|█         | 1118/10000 [01:11<09:18, 15.91it/s]\n",
      "Loss: 2.241 :  11%|█         | 1120/10000 [01:11<09:23, 15.76it/s]\n",
      "Loss: 2.241 :  11%|█         | 1122/10000 [01:12<09:16, 15.94it/s]\n",
      "Loss: 2.241 :  11%|█         | 1124/10000 [01:12<09:10, 16.13it/s]\n",
      "Loss: 2.241 :  11%|█▏        | 1126/10000 [01:12<09:04, 16.30it/s]\n",
      "Loss: 2.241 :  11%|█▏        | 1128/10000 [01:12<09:04, 16.31it/s]\n",
      "Loss: 2.241 :  11%|█▏        | 1130/10000 [01:12<08:57, 16.52it/s]\n",
      "Loss: 2.241 :  11%|█▏        | 1132/10000 [01:12<08:51, 16.70it/s]\n",
      "Loss: 2.241 :  11%|█▏        | 1134/10000 [01:12<08:44, 16.89it/s]\n",
      "Loss: 2.241 :  11%|█▏        | 1136/10000 [01:12<08:48, 16.77it/s]\n",
      "Loss: 2.241 :  11%|█▏        | 1138/10000 [01:13<08:37, 17.12it/s]\n",
      "Loss: 2.241 :  11%|█▏        | 1140/10000 [01:13<08:34, 17.23it/s]\n",
      "Loss: 2.241 :  11%|█▏        | 1142/10000 [01:13<08:29, 17.40it/s]\n",
      "Loss: 2.241 :  11%|█▏        | 1144/10000 [01:13<08:27, 17.47it/s]\n",
      "Loss: 2.241 :  11%|█▏        | 1146/10000 [01:13<08:37, 17.10it/s]\n",
      "Loss: 2.241 :  11%|█▏        | 1148/10000 [01:13<08:36, 17.12it/s]\n",
      "Loss: 2.241 :  12%|█▏        | 1150/10000 [01:13<08:45, 16.85it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1150/10000 [01:13<08:45, 16.85it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1152/10000 [01:13<09:39, 15.27it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1154/10000 [01:14<09:25, 15.63it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1156/10000 [01:14<09:12, 16.01it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1158/10000 [01:14<09:00, 16.36it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1160/10000 [01:14<08:57, 16.44it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1162/10000 [01:14<08:59, 16.38it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1164/10000 [01:14<09:03, 16.26it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1166/10000 [01:14<09:11, 16.02it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1168/10000 [01:14<09:26, 15.59it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1170/10000 [01:15<09:55, 14.84it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1172/10000 [01:15<10:19, 14.25it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1174/10000 [01:15<11:48, 12.46it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1176/10000 [01:15<11:38, 12.64it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1178/10000 [01:15<11:19, 12.98it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1180/10000 [01:15<11:08, 13.20it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1182/10000 [01:16<12:46, 11.50it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1184/10000 [01:16<13:48, 10.65it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1186/10000 [01:16<13:36, 10.79it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1188/10000 [01:16<12:43, 11.54it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1190/10000 [01:16<12:02, 12.19it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1192/10000 [01:16<11:31, 12.73it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1194/10000 [01:17<10:56, 13.40it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1196/10000 [01:17<10:43, 13.68it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1198/10000 [01:17<10:26, 14.05it/s]\n",
      "Loss: 2.161 :  12%|█▏        | 1200/10000 [01:17<10:24, 14.09it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1200/10000 [01:17<10:24, 14.09it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1202/10000 [01:17<11:08, 13.16it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1204/10000 [01:17<10:36, 13.82it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1206/10000 [01:17<10:20, 14.17it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1208/10000 [01:18<10:03, 14.58it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1210/10000 [01:18<09:57, 14.72it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1212/10000 [01:18<09:43, 15.07it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1214/10000 [01:18<09:51, 14.85it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1216/10000 [01:18<10:08, 14.43it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1218/10000 [01:18<10:03, 14.56it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1220/10000 [01:18<10:08, 14.44it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1222/10000 [01:18<10:06, 14.47it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1224/10000 [01:19<09:57, 14.68it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1226/10000 [01:19<09:46, 14.95it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1228/10000 [01:19<09:33, 15.29it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1230/10000 [01:19<09:19, 15.68it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1232/10000 [01:19<09:01, 16.18it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1234/10000 [01:19<08:53, 16.44it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1236/10000 [01:19<08:45, 16.69it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1238/10000 [01:19<08:47, 16.61it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1240/10000 [01:20<08:38, 16.90it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1242/10000 [01:20<08:53, 16.41it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1244/10000 [01:20<09:13, 15.81it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1246/10000 [01:20<09:12, 15.84it/s]\n",
      "Loss: 2.179 :  12%|█▏        | 1248/10000 [01:20<09:06, 16.00it/s]\n",
      "Loss: 2.179 :  12%|█▎        | 1250/10000 [01:20<09:13, 15.81it/s]\n",
      "Loss: 2.116 :  12%|█▎        | 1250/10000 [01:20<09:13, 15.81it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1252/10000 [01:20<10:44, 13.58it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1254/10000 [01:21<11:00, 13.24it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1256/10000 [01:21<12:13, 11.92it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1258/10000 [01:21<15:31,  9.39it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1260/10000 [01:21<14:51,  9.81it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1262/10000 [01:21<13:46, 10.58it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1264/10000 [01:22<12:19, 11.81it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1266/10000 [01:22<11:31, 12.64it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1268/10000 [01:22<10:45, 13.52it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1270/10000 [01:22<10:08, 14.34it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1272/10000 [01:22<09:28, 15.36it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1274/10000 [01:22<09:13, 15.78it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1276/10000 [01:22<08:52, 16.39it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1278/10000 [01:22<08:39, 16.78it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1280/10000 [01:22<08:28, 17.16it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1282/10000 [01:23<08:15, 17.60it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1284/10000 [01:23<08:01, 18.10it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1286/10000 [01:23<07:53, 18.40it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1288/10000 [01:23<08:02, 18.07it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1290/10000 [01:23<08:04, 17.97it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1293/10000 [01:23<07:43, 18.80it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1295/10000 [01:23<07:43, 18.79it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1297/10000 [01:23<07:40, 18.89it/s]\n",
      "Loss: 2.116 :  13%|█▎        | 1299/10000 [01:23<07:36, 19.06it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1299/10000 [01:24<07:36, 19.06it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1301/10000 [01:24<08:13, 17.62it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1304/10000 [01:24<07:45, 18.69it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1306/10000 [01:24<07:39, 18.93it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1308/10000 [01:24<07:34, 19.14it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1311/10000 [01:24<07:24, 19.55it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1313/10000 [01:24<07:25, 19.52it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1315/10000 [01:24<07:25, 19.51it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1318/10000 [01:24<07:21, 19.68it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1320/10000 [01:25<07:21, 19.66it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1323/10000 [01:25<07:08, 20.25it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1326/10000 [01:25<07:07, 20.31it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1329/10000 [01:25<07:00, 20.60it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1332/10000 [01:25<07:05, 20.39it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1335/10000 [01:25<07:15, 19.90it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1337/10000 [01:25<07:17, 19.80it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1339/10000 [01:26<07:18, 19.76it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1341/10000 [01:26<07:18, 19.77it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1343/10000 [01:26<07:23, 19.52it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1345/10000 [01:26<07:32, 19.13it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1347/10000 [01:26<08:19, 17.32it/s]\n",
      "Loss: 2.108 :  13%|█▎        | 1349/10000 [01:26<09:36, 15.02it/s]\n",
      "Loss: 2.179 :  13%|█▎        | 1349/10000 [01:26<09:36, 15.02it/s]\n",
      "Loss: 2.179 :  14%|█▎        | 1351/10000 [01:26<10:58, 13.14it/s]\n",
      "Loss: 2.179 :  14%|█▎        | 1353/10000 [01:27<10:48, 13.33it/s]\n",
      "Loss: 2.179 :  14%|█▎        | 1355/10000 [01:27<10:30, 13.71it/s]\n",
      "Loss: 2.179 :  14%|█▎        | 1357/10000 [01:27<10:10, 14.17it/s]\n",
      "Loss: 2.179 :  14%|█▎        | 1359/10000 [01:27<09:53, 14.57it/s]\n",
      "Loss: 2.179 :  14%|█▎        | 1361/10000 [01:27<09:31, 15.12it/s]\n",
      "Loss: 2.179 :  14%|█▎        | 1363/10000 [01:27<09:13, 15.59it/s]\n",
      "Loss: 2.179 :  14%|█▎        | 1365/10000 [01:27<09:05, 15.84it/s]\n",
      "Loss: 2.179 :  14%|█▎        | 1367/10000 [01:27<09:12, 15.63it/s]\n",
      "Loss: 2.179 :  14%|█▎        | 1369/10000 [01:28<09:26, 15.24it/s]\n",
      "Loss: 2.179 :  14%|█▎        | 1371/10000 [01:28<09:19, 15.42it/s]\n",
      "Loss: 2.179 :  14%|█▎        | 1373/10000 [01:28<09:02, 15.91it/s]\n",
      "Loss: 2.179 :  14%|█▍        | 1375/10000 [01:28<09:02, 15.91it/s]\n",
      "Loss: 2.179 :  14%|█▍        | 1377/10000 [01:28<08:40, 16.56it/s]\n",
      "Loss: 2.179 :  14%|█▍        | 1379/10000 [01:28<08:27, 16.97it/s]\n",
      "Loss: 2.179 :  14%|█▍        | 1381/10000 [01:28<08:06, 17.72it/s]\n",
      "Loss: 2.179 :  14%|█▍        | 1383/10000 [01:28<07:58, 18.00it/s]\n",
      "Loss: 2.179 :  14%|█▍        | 1385/10000 [01:28<07:56, 18.08it/s]\n",
      "Loss: 2.179 :  14%|█▍        | 1387/10000 [01:29<08:39, 16.57it/s]\n",
      "Loss: 2.179 :  14%|█▍        | 1389/10000 [01:29<11:44, 12.22it/s]\n",
      "Loss: 2.179 :  14%|█▍        | 1391/10000 [01:29<15:58,  8.98it/s]\n",
      "Loss: 2.179 :  14%|█▍        | 1393/10000 [01:29<16:17,  8.81it/s]\n",
      "Loss: 2.179 :  14%|█▍        | 1395/10000 [01:30<16:03,  8.93it/s]\n",
      "Loss: 2.179 :  14%|█▍        | 1397/10000 [01:30<15:48,  9.07it/s]\n",
      "Loss: 2.179 :  14%|█▍        | 1398/10000 [01:30<15:36,  9.19it/s]\n",
      "Loss: 2.179 :  14%|█▍        | 1399/10000 [01:30<16:37,  8.62it/s]\n",
      "Loss: 2.179 :  14%|█▍        | 1400/10000 [01:30<20:41,  6.93it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1400/10000 [01:30<20:41,  6.93it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1401/10000 [01:31<22:13,  6.45it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1403/10000 [01:31<18:28,  7.75it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1405/10000 [01:31<16:34,  8.65it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1407/10000 [01:31<15:09,  9.44it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1409/10000 [01:31<13:27, 10.64it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1411/10000 [01:31<12:22, 11.57it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1413/10000 [01:32<15:12,  9.41it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1415/10000 [01:32<14:55,  9.58it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1417/10000 [01:32<14:07, 10.13it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1419/10000 [01:32<13:36, 10.51it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1421/10000 [01:32<12:55, 11.07it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1423/10000 [01:33<12:24, 11.51it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1425/10000 [01:33<13:39, 10.47it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1427/10000 [01:33<12:42, 11.25it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1429/10000 [01:33<11:56, 11.97it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1431/10000 [01:33<11:12, 12.74it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1433/10000 [01:33<10:46, 13.26it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1435/10000 [01:33<10:19, 13.83it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1437/10000 [01:34<09:59, 14.28it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1439/10000 [01:34<09:34, 14.91it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1441/10000 [01:34<09:30, 15.01it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1443/10000 [01:34<09:18, 15.32it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1445/10000 [01:34<09:10, 15.54it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1447/10000 [01:34<09:20, 15.27it/s]\n",
      "Loss: 1.970 :  14%|█▍        | 1449/10000 [01:34<09:53, 14.41it/s]\n",
      "Loss: 2.188 :  14%|█▍        | 1449/10000 [01:35<09:53, 14.41it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1451/10000 [01:35<10:44, 13.26it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1453/10000 [01:35<09:55, 14.34it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1455/10000 [01:35<09:42, 14.68it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1457/10000 [01:35<09:16, 15.35it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1459/10000 [01:35<08:53, 16.01it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1461/10000 [01:35<08:41, 16.38it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1463/10000 [01:35<08:30, 16.71it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1465/10000 [01:35<08:24, 16.92it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1467/10000 [01:35<08:21, 17.01it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1469/10000 [01:36<08:16, 17.20it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1471/10000 [01:36<08:13, 17.30it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1473/10000 [01:36<08:11, 17.34it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1475/10000 [01:36<08:04, 17.60it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1477/10000 [01:36<08:05, 17.55it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1479/10000 [01:36<08:13, 17.26it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1481/10000 [01:36<08:07, 17.47it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1483/10000 [01:36<08:02, 17.64it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1485/10000 [01:37<08:04, 17.59it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1487/10000 [01:37<08:09, 17.41it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1489/10000 [01:37<08:12, 17.29it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1491/10000 [01:37<08:18, 17.07it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1493/10000 [01:37<08:17, 17.10it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1495/10000 [01:37<08:36, 16.47it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1497/10000 [01:37<08:47, 16.12it/s]\n",
      "Loss: 2.188 :  15%|█▍        | 1499/10000 [01:37<08:53, 15.94it/s]\n",
      "Loss: 2.218 :  15%|█▍        | 1499/10000 [01:37<08:53, 15.94it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1501/10000 [01:38<09:56, 14.26it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1503/10000 [01:38<09:49, 14.42it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1505/10000 [01:38<09:44, 14.54it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1507/10000 [01:38<09:47, 14.45it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1509/10000 [01:38<09:52, 14.34it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1511/10000 [01:38<09:40, 14.62it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1513/10000 [01:38<09:33, 14.79it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1515/10000 [01:38<09:21, 15.12it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1517/10000 [01:39<09:09, 15.44it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1519/10000 [01:39<08:59, 15.72it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1521/10000 [01:39<09:21, 15.09it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1523/10000 [01:39<09:07, 15.49it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1525/10000 [01:39<09:05, 15.55it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1527/10000 [01:39<08:57, 15.78it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1529/10000 [01:39<08:50, 15.97it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1531/10000 [01:40<09:45, 14.46it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1533/10000 [01:40<10:22, 13.61it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1535/10000 [01:40<10:14, 13.78it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1537/10000 [01:40<09:54, 14.25it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1539/10000 [01:40<10:10, 13.85it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1541/10000 [01:40<10:07, 13.91it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1543/10000 [01:40<10:28, 13.45it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1545/10000 [01:41<10:55, 12.89it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1547/10000 [01:41<11:21, 12.41it/s]\n",
      "Loss: 2.218 :  15%|█▌        | 1549/10000 [01:41<11:00, 12.79it/s]\n",
      "Loss: 2.085 :  15%|█▌        | 1549/10000 [01:41<11:00, 12.79it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1551/10000 [01:41<11:58, 11.75it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1553/10000 [01:41<11:18, 12.46it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1555/10000 [01:41<10:36, 13.27it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1557/10000 [01:42<10:11, 13.80it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1559/10000 [01:42<09:55, 14.19it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1561/10000 [01:42<09:42, 14.48it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1563/10000 [01:42<09:29, 14.82it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1565/10000 [01:42<09:17, 15.14it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1567/10000 [01:42<08:57, 15.69it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1569/10000 [01:42<08:49, 15.93it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1571/10000 [01:42<08:39, 16.21it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1573/10000 [01:43<08:38, 16.25it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1575/10000 [01:43<08:27, 16.61it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1577/10000 [01:43<08:18, 16.90it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1579/10000 [01:43<08:13, 17.06it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1581/10000 [01:43<08:05, 17.35it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1583/10000 [01:43<08:01, 17.50it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1585/10000 [01:43<07:50, 17.88it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1587/10000 [01:43<07:53, 17.78it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1589/10000 [01:43<08:02, 17.44it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1591/10000 [01:44<08:00, 17.50it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1593/10000 [01:44<08:06, 17.28it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1595/10000 [01:44<08:26, 16.61it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1597/10000 [01:44<08:44, 16.03it/s]\n",
      "Loss: 2.085 :  16%|█▌        | 1599/10000 [01:44<09:00, 15.55it/s]\n",
      "Loss: 1.989 :  16%|█▌        | 1599/10000 [01:44<09:00, 15.55it/s]\n",
      "Loss: 1.989 :  16%|█▌        | 1601/10000 [01:44<10:12, 13.72it/s]\n",
      "Loss: 1.989 :  16%|█▌        | 1603/10000 [01:44<10:32, 13.27it/s]\n",
      "Loss: 1.989 :  16%|█▌        | 1605/10000 [01:45<10:25, 13.42it/s]\n",
      "Loss: 1.989 :  16%|█▌        | 1607/10000 [01:45<10:24, 13.44it/s]\n",
      "Loss: 1.989 :  16%|█▌        | 1609/10000 [01:45<11:40, 11.99it/s]\n",
      "Loss: 1.989 :  16%|█▌        | 1611/10000 [01:45<12:46, 10.94it/s]\n",
      "Loss: 1.989 :  16%|█▌        | 1613/10000 [01:45<12:50, 10.89it/s]\n",
      "Loss: 1.989 :  16%|█▌        | 1615/10000 [01:45<12:40, 11.03it/s]\n",
      "Loss: 1.989 :  16%|█▌        | 1617/10000 [01:46<12:39, 11.04it/s]\n",
      "Loss: 1.989 :  16%|█▌        | 1619/10000 [01:46<12:39, 11.04it/s]\n",
      "Loss: 1.989 :  16%|█▌        | 1621/10000 [01:46<12:20, 11.31it/s]\n",
      "Loss: 1.989 :  16%|█▌        | 1623/10000 [01:46<11:42, 11.92it/s]\n",
      "Loss: 1.989 :  16%|█▋        | 1625/10000 [01:46<11:03, 12.63it/s]\n",
      "Loss: 1.989 :  16%|█▋        | 1627/10000 [01:46<10:48, 12.90it/s]\n",
      "Loss: 1.989 :  16%|█▋        | 1629/10000 [01:47<10:27, 13.35it/s]\n",
      "Loss: 1.989 :  16%|█▋        | 1631/10000 [01:47<09:47, 14.24it/s]\n",
      "Loss: 1.989 :  16%|█▋        | 1633/10000 [01:47<09:17, 15.01it/s]\n",
      "Loss: 1.989 :  16%|█▋        | 1635/10000 [01:47<09:14, 15.10it/s]\n",
      "Loss: 1.989 :  16%|█▋        | 1637/10000 [01:47<08:54, 15.65it/s]\n",
      "Loss: 1.989 :  16%|█▋        | 1639/10000 [01:47<08:48, 15.82it/s]\n",
      "Loss: 1.989 :  16%|█▋        | 1641/10000 [01:47<08:48, 15.82it/s]\n",
      "Loss: 1.989 :  16%|█▋        | 1643/10000 [01:47<08:54, 15.65it/s]\n",
      "Loss: 1.989 :  16%|█▋        | 1645/10000 [01:48<09:08, 15.24it/s]\n",
      "Loss: 1.989 :  16%|█▋        | 1647/10000 [01:48<09:07, 15.24it/s]\n",
      "Loss: 1.989 :  16%|█▋        | 1649/10000 [01:48<09:00, 15.44it/s]\n",
      "Loss: 1.922 :  16%|█▋        | 1649/10000 [01:48<09:00, 15.44it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1651/10000 [01:48<09:57, 13.97it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1653/10000 [01:48<09:30, 14.62it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1655/10000 [01:48<09:08, 15.22it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1657/10000 [01:48<08:53, 15.63it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1659/10000 [01:49<08:42, 15.97it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1661/10000 [01:49<08:34, 16.21it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1663/10000 [01:49<11:57, 11.63it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1665/10000 [01:49<11:25, 12.16it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1667/10000 [01:49<10:49, 12.82it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1669/10000 [01:49<10:30, 13.21it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1671/10000 [01:49<10:27, 13.28it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1673/10000 [01:50<10:28, 13.26it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1675/10000 [01:50<10:17, 13.48it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1677/10000 [01:50<10:25, 13.31it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1679/10000 [01:50<10:27, 13.26it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1681/10000 [01:50<10:39, 13.02it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1683/10000 [01:50<11:41, 11.85it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1685/10000 [01:51<12:03, 11.48it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1687/10000 [01:51<12:13, 11.33it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1689/10000 [01:51<12:48, 10.81it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1691/10000 [01:51<12:04, 11.47it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1693/10000 [01:51<11:33, 11.98it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1695/10000 [01:51<11:32, 12.00it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1697/10000 [01:52<11:15, 12.29it/s]\n",
      "Loss: 1.922 :  17%|█▋        | 1699/10000 [01:52<10:45, 12.86it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1699/10000 [01:52<10:45, 12.86it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1701/10000 [01:52<11:22, 12.16it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1703/10000 [01:52<10:33, 13.10it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1705/10000 [01:52<09:54, 13.96it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1707/10000 [01:52<09:23, 14.71it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1709/10000 [01:52<09:04, 15.23it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1711/10000 [01:53<08:56, 15.45it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1713/10000 [01:53<08:53, 15.52it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1715/10000 [01:53<08:44, 15.79it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1717/10000 [01:53<08:54, 15.49it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1719/10000 [01:53<08:52, 15.54it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1721/10000 [01:53<08:53, 15.52it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1723/10000 [01:53<08:47, 15.69it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1725/10000 [01:53<08:42, 15.85it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1727/10000 [01:54<08:35, 16.06it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1729/10000 [01:54<08:19, 16.57it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1731/10000 [01:54<08:20, 16.52it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1733/10000 [01:54<08:20, 16.53it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1735/10000 [01:54<08:21, 16.49it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1737/10000 [01:54<08:21, 16.47it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1739/10000 [01:54<08:27, 16.27it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1741/10000 [01:54<08:29, 16.22it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1743/10000 [01:55<08:20, 16.51it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1745/10000 [01:55<08:35, 16.01it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1747/10000 [01:55<08:57, 15.36it/s]\n",
      "Loss: 1.971 :  17%|█▋        | 1749/10000 [01:55<08:50, 15.55it/s]\n",
      "Loss: 1.902 :  17%|█▋        | 1749/10000 [01:55<08:50, 15.55it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1751/10000 [01:55<09:29, 14.47it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1753/10000 [01:55<09:11, 14.95it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1755/10000 [01:55<08:52, 15.49it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1757/10000 [01:55<08:44, 15.71it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1759/10000 [01:56<09:20, 14.71it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1761/10000 [01:56<10:01, 13.70it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1763/10000 [01:56<10:09, 13.51it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1765/10000 [01:56<10:49, 12.69it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1767/10000 [01:56<11:03, 12.40it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1769/10000 [01:56<10:35, 12.94it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1771/10000 [01:57<10:10, 13.48it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1773/10000 [01:57<09:58, 13.76it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1775/10000 [01:57<10:36, 12.93it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1777/10000 [01:57<12:19, 11.11it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1779/10000 [01:57<13:33, 10.10it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1781/10000 [01:58<12:59, 10.54it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1783/10000 [01:58<12:25, 11.02it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1785/10000 [01:58<11:35, 11.81it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1787/10000 [01:58<10:46, 12.71it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1789/10000 [01:58<10:12, 13.41it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1791/10000 [01:58<09:43, 14.06it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1793/10000 [01:58<09:24, 14.53it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1795/10000 [01:58<09:07, 14.99it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1797/10000 [01:59<08:52, 15.42it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1799/10000 [01:59<08:44, 15.65it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1799/10000 [01:59<08:44, 15.65it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1801/10000 [01:59<09:30, 14.36it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1803/10000 [01:59<09:09, 14.91it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1805/10000 [01:59<08:53, 15.36it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1807/10000 [01:59<08:36, 15.87it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1809/10000 [01:59<08:19, 16.40it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1811/10000 [01:59<08:18, 16.44it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1813/10000 [02:00<08:11, 16.67it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1815/10000 [02:00<08:07, 16.79it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1817/10000 [02:00<08:06, 16.84it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1819/10000 [02:00<07:57, 17.12it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1821/10000 [02:00<07:54, 17.24it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1823/10000 [02:00<07:55, 17.21it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1825/10000 [02:00<07:55, 17.19it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1827/10000 [02:00<07:54, 17.21it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1829/10000 [02:01<08:01, 16.98it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1831/10000 [02:01<07:59, 17.03it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1833/10000 [02:01<07:57, 17.11it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1835/10000 [02:01<07:56, 17.13it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1837/10000 [02:01<07:49, 17.40it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1839/10000 [02:01<07:47, 17.46it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1841/10000 [02:01<07:47, 17.45it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1843/10000 [02:01<08:06, 16.78it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1845/10000 [02:02<08:51, 15.34it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1847/10000 [02:02<09:22, 14.49it/s]\n",
      "Loss: 1.902 :  18%|█▊        | 1849/10000 [02:02<09:21, 14.52it/s]\n",
      "Loss: 1.879 :  18%|█▊        | 1849/10000 [02:02<09:21, 14.52it/s]\n",
      "Loss: 1.879 :  19%|█▊        | 1851/10000 [02:02<10:05, 13.47it/s]\n",
      "Loss: 1.879 :  19%|█▊        | 1853/10000 [02:02<09:32, 14.22it/s]\n",
      "Loss: 1.879 :  19%|█▊        | 1855/10000 [02:02<09:11, 14.77it/s]\n",
      "Loss: 1.879 :  19%|█▊        | 1857/10000 [02:02<08:56, 15.17it/s]\n",
      "Loss: 1.879 :  19%|█▊        | 1859/10000 [02:02<08:49, 15.36it/s]\n",
      "Loss: 1.879 :  19%|█▊        | 1861/10000 [02:03<08:28, 16.02it/s]\n",
      "Loss: 1.879 :  19%|█▊        | 1863/10000 [02:03<08:25, 16.10it/s]\n",
      "Loss: 1.879 :  19%|█▊        | 1865/10000 [02:03<08:26, 16.06it/s]\n",
      "Loss: 1.879 :  19%|█▊        | 1867/10000 [02:03<08:23, 16.15it/s]\n",
      "Loss: 1.879 :  19%|█▊        | 1869/10000 [02:03<08:20, 16.24it/s]\n",
      "Loss: 1.879 :  19%|█▊        | 1871/10000 [02:03<08:14, 16.45it/s]\n",
      "Loss: 1.879 :  19%|█▊        | 1873/10000 [02:03<08:19, 16.29it/s]\n",
      "Loss: 1.879 :  19%|█▉        | 1875/10000 [02:03<08:14, 16.44it/s]\n",
      "Loss: 1.879 :  19%|█▉        | 1877/10000 [02:04<08:04, 16.77it/s]\n",
      "Loss: 1.879 :  19%|█▉        | 1879/10000 [02:04<07:59, 16.94it/s]\n",
      "Loss: 1.879 :  19%|█▉        | 1881/10000 [02:04<07:54, 17.10it/s]\n",
      "Loss: 1.879 :  19%|█▉        | 1883/10000 [02:04<07:51, 17.22it/s]\n",
      "Loss: 1.879 :  19%|█▉        | 1885/10000 [02:04<07:43, 17.51it/s]\n",
      "Loss: 1.879 :  19%|█▉        | 1887/10000 [02:04<07:50, 17.23it/s]\n",
      "Loss: 1.879 :  19%|█▉        | 1889/10000 [02:04<07:52, 17.15it/s]\n",
      "Loss: 1.879 :  19%|█▉        | 1891/10000 [02:04<07:52, 17.15it/s]\n",
      "Loss: 1.879 :  19%|█▉        | 1893/10000 [02:04<07:54, 17.07it/s]\n",
      "Loss: 1.879 :  19%|█▉        | 1895/10000 [02:05<08:01, 16.85it/s]\n",
      "Loss: 1.879 :  19%|█▉        | 1897/10000 [02:05<07:53, 17.13it/s]\n",
      "Loss: 1.879 :  19%|█▉        | 1899/10000 [02:05<07:53, 17.10it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1899/10000 [02:05<07:53, 17.10it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1901/10000 [02:05<08:41, 15.54it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1903/10000 [02:05<08:27, 15.94it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1905/10000 [02:05<08:21, 16.14it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1907/10000 [02:05<08:06, 16.62it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1909/10000 [02:05<08:02, 16.78it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1911/10000 [02:06<07:54, 17.06it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1913/10000 [02:06<07:47, 17.29it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1915/10000 [02:06<07:47, 17.29it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1917/10000 [02:06<07:43, 17.45it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1919/10000 [02:06<07:49, 17.21it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1921/10000 [02:06<07:48, 17.24it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1923/10000 [02:06<07:48, 17.25it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1925/10000 [02:06<07:54, 17.03it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1927/10000 [02:06<07:43, 17.41it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1929/10000 [02:07<07:43, 17.43it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1931/10000 [02:07<08:08, 16.52it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1933/10000 [02:07<08:42, 15.45it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1935/10000 [02:07<08:56, 15.03it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1937/10000 [02:07<08:47, 15.27it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1939/10000 [02:07<08:36, 15.60it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1941/10000 [02:07<08:32, 15.74it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1943/10000 [02:08<08:20, 16.11it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1945/10000 [02:08<08:08, 16.49it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1947/10000 [02:08<08:03, 16.65it/s]\n",
      "Loss: 1.892 :  19%|█▉        | 1949/10000 [02:08<07:50, 17.11it/s]\n",
      "Loss: 1.865 :  19%|█▉        | 1949/10000 [02:08<07:50, 17.11it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1951/10000 [02:08<08:39, 15.49it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1953/10000 [02:08<08:37, 15.54it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1955/10000 [02:08<08:37, 15.55it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1957/10000 [02:08<08:38, 15.52it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1959/10000 [02:09<08:43, 15.37it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1961/10000 [02:09<08:48, 15.20it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1963/10000 [02:09<08:45, 15.30it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1965/10000 [02:09<08:36, 15.56it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1967/10000 [02:09<08:21, 16.02it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1969/10000 [02:09<08:08, 16.43it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1971/10000 [02:09<08:07, 16.46it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1973/10000 [02:09<08:03, 16.60it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1975/10000 [02:10<08:02, 16.65it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1977/10000 [02:10<07:58, 16.76it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1979/10000 [02:10<07:51, 17.00it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1981/10000 [02:10<07:51, 17.02it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1983/10000 [02:10<07:45, 17.23it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1985/10000 [02:10<07:41, 17.38it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1987/10000 [02:10<07:39, 17.44it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1989/10000 [02:10<07:39, 17.45it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1991/10000 [02:10<07:30, 17.77it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1993/10000 [02:11<07:33, 17.66it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1995/10000 [02:11<07:31, 17.71it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1997/10000 [02:11<07:25, 17.96it/s]\n",
      "Loss: 1.865 :  20%|█▉        | 1999/10000 [02:11<07:28, 17.86it/s]\n",
      "Loss: 1.847 :  20%|█▉        | 1999/10000 [02:11<07:28, 17.86it/s]\n",
      "Loss: 1.847 :  20%|██        | 2001/10000 [02:11<08:26, 15.79it/s]\n",
      "Loss: 1.847 :  20%|██        | 2003/10000 [02:11<08:21, 15.96it/s]\n",
      "Loss: 1.847 :  20%|██        | 2005/10000 [02:11<08:07, 16.40it/s]\n",
      "Loss: 1.847 :  20%|██        | 2007/10000 [02:11<08:03, 16.53it/s]\n",
      "Loss: 1.847 :  20%|██        | 2009/10000 [02:11<07:50, 16.97it/s]\n",
      "Loss: 1.847 :  20%|██        | 2011/10000 [02:12<07:50, 16.98it/s]\n",
      "Loss: 1.847 :  20%|██        | 2013/10000 [02:12<08:04, 16.50it/s]\n",
      "Loss: 1.847 :  20%|██        | 2015/10000 [02:12<07:56, 16.76it/s]\n",
      "Loss: 1.847 :  20%|██        | 2017/10000 [02:12<07:58, 16.68it/s]\n",
      "Loss: 1.847 :  20%|██        | 2019/10000 [02:12<08:47, 15.12it/s]\n",
      "Loss: 1.847 :  20%|██        | 2021/10000 [02:12<09:11, 14.47it/s]\n",
      "Loss: 1.847 :  20%|██        | 2023/10000 [02:12<09:02, 14.69it/s]\n",
      "Loss: 1.847 :  20%|██        | 2025/10000 [02:13<08:50, 15.04it/s]\n",
      "Loss: 1.847 :  20%|██        | 2027/10000 [02:13<08:47, 15.13it/s]\n",
      "Loss: 1.847 :  20%|██        | 2029/10000 [02:13<08:26, 15.75it/s]\n",
      "Loss: 1.847 :  20%|██        | 2031/10000 [02:13<08:17, 16.01it/s]\n",
      "Loss: 1.847 :  20%|██        | 2033/10000 [02:13<08:11, 16.21it/s]\n",
      "Loss: 1.847 :  20%|██        | 2035/10000 [02:13<08:05, 16.41it/s]\n",
      "Loss: 1.847 :  20%|██        | 2037/10000 [02:13<07:56, 16.72it/s]\n",
      "Loss: 1.847 :  20%|██        | 2039/10000 [02:13<07:52, 16.86it/s]\n",
      "Loss: 1.847 :  20%|██        | 2041/10000 [02:13<07:45, 17.09it/s]\n",
      "Loss: 1.847 :  20%|██        | 2043/10000 [02:14<07:37, 17.40it/s]\n",
      "Loss: 1.847 :  20%|██        | 2045/10000 [02:14<07:42, 17.20it/s]\n",
      "Loss: 1.847 :  20%|██        | 2047/10000 [02:14<07:43, 17.16it/s]\n",
      "Loss: 1.847 :  20%|██        | 2049/10000 [02:14<07:44, 17.13it/s]\n",
      "Loss: 1.660 :  20%|██        | 2049/10000 [02:14<07:44, 17.13it/s]\n",
      "Loss: 1.660 :  21%|██        | 2051/10000 [02:14<08:31, 15.55it/s]\n",
      "Loss: 1.660 :  21%|██        | 2053/10000 [02:14<08:08, 16.26it/s]\n",
      "Loss: 1.660 :  21%|██        | 2055/10000 [02:14<08:04, 16.39it/s]\n",
      "Loss: 1.660 :  21%|██        | 2057/10000 [02:14<07:47, 17.00it/s]\n",
      "Loss: 1.660 :  21%|██        | 2059/10000 [02:15<07:45, 17.07it/s]\n",
      "Loss: 1.660 :  21%|██        | 2061/10000 [02:15<07:38, 17.31it/s]\n",
      "Loss: 1.660 :  21%|██        | 2063/10000 [02:15<07:37, 17.36it/s]\n",
      "Loss: 1.660 :  21%|██        | 2065/10000 [02:15<08:44, 15.14it/s]\n",
      "Loss: 1.660 :  21%|██        | 2067/10000 [02:15<08:31, 15.52it/s]\n",
      "Loss: 1.660 :  21%|██        | 2069/10000 [02:15<08:20, 15.84it/s]\n",
      "Loss: 1.660 :  21%|██        | 2071/10000 [02:15<08:06, 16.30it/s]\n",
      "Loss: 1.660 :  21%|██        | 2073/10000 [02:15<07:55, 16.66it/s]\n",
      "Loss: 1.660 :  21%|██        | 2075/10000 [02:16<07:49, 16.89it/s]\n",
      "Loss: 1.660 :  21%|██        | 2077/10000 [02:16<07:48, 16.93it/s]\n",
      "Loss: 1.660 :  21%|██        | 2079/10000 [02:16<07:39, 17.23it/s]\n",
      "Loss: 1.660 :  21%|██        | 2081/10000 [02:16<07:37, 17.31it/s]\n",
      "Loss: 1.660 :  21%|██        | 2083/10000 [02:16<07:39, 17.21it/s]\n",
      "Loss: 1.660 :  21%|██        | 2085/10000 [02:16<07:46, 16.97it/s]\n",
      "Loss: 1.660 :  21%|██        | 2087/10000 [02:16<07:45, 16.99it/s]\n",
      "Loss: 1.660 :  21%|██        | 2089/10000 [02:16<07:42, 17.09it/s]\n",
      "Loss: 1.660 :  21%|██        | 2091/10000 [02:16<07:38, 17.27it/s]\n",
      "Loss: 1.660 :  21%|██        | 2093/10000 [02:17<07:26, 17.70it/s]\n",
      "Loss: 1.660 :  21%|██        | 2095/10000 [02:17<07:26, 17.72it/s]\n",
      "Loss: 1.660 :  21%|██        | 2097/10000 [02:17<07:34, 17.38it/s]\n",
      "Loss: 1.660 :  21%|██        | 2099/10000 [02:17<07:41, 17.13it/s]\n",
      "Loss: 1.953 :  21%|██        | 2099/10000 [02:17<07:41, 17.13it/s]\n",
      "Loss: 1.953 :  21%|██        | 2101/10000 [02:17<08:29, 15.51it/s]\n",
      "Loss: 1.953 :  21%|██        | 2103/10000 [02:17<08:11, 16.06it/s]\n",
      "Loss: 1.953 :  21%|██        | 2105/10000 [02:17<08:15, 15.94it/s]\n",
      "Loss: 1.953 :  21%|██        | 2107/10000 [02:17<08:39, 15.20it/s]\n",
      "Loss: 1.953 :  21%|██        | 2109/10000 [02:18<08:47, 14.96it/s]\n",
      "Loss: 1.953 :  21%|██        | 2111/10000 [02:18<08:54, 14.76it/s]\n",
      "Loss: 1.953 :  21%|██        | 2113/10000 [02:18<08:57, 14.67it/s]\n",
      "Loss: 1.953 :  21%|██        | 2115/10000 [02:18<09:07, 14.40it/s]\n",
      "Loss: 1.953 :  21%|██        | 2117/10000 [02:18<09:07, 14.41it/s]\n",
      "Loss: 1.953 :  21%|██        | 2119/10000 [02:18<08:53, 14.78it/s]\n",
      "Loss: 1.953 :  21%|██        | 2121/10000 [02:18<08:44, 15.02it/s]\n",
      "Loss: 1.953 :  21%|██        | 2123/10000 [02:19<08:42, 15.06it/s]\n",
      "Loss: 1.953 :  21%|██▏       | 2125/10000 [02:19<08:37, 15.21it/s]\n",
      "Loss: 1.953 :  21%|██▏       | 2127/10000 [02:19<08:39, 15.15it/s]\n",
      "Loss: 1.953 :  21%|██▏       | 2129/10000 [02:19<08:42, 15.07it/s]\n",
      "Loss: 1.953 :  21%|██▏       | 2131/10000 [02:19<08:49, 14.87it/s]\n",
      "Loss: 1.953 :  21%|██▏       | 2133/10000 [02:19<08:53, 14.75it/s]\n",
      "Loss: 1.953 :  21%|██▏       | 2135/10000 [02:19<08:53, 14.74it/s]\n",
      "Loss: 1.953 :  21%|██▏       | 2137/10000 [02:20<08:45, 14.96it/s]\n",
      "Loss: 1.953 :  21%|██▏       | 2139/10000 [02:20<08:25, 15.56it/s]\n",
      "Loss: 1.953 :  21%|██▏       | 2141/10000 [02:20<08:20, 15.71it/s]\n",
      "Loss: 1.953 :  21%|██▏       | 2143/10000 [02:20<08:20, 15.69it/s]\n",
      "Loss: 1.953 :  21%|██▏       | 2145/10000 [02:20<08:12, 15.94it/s]\n",
      "Loss: 1.953 :  21%|██▏       | 2147/10000 [02:20<08:02, 16.28it/s]\n",
      "Loss: 1.953 :  21%|██▏       | 2149/10000 [02:20<08:04, 16.21it/s]\n",
      "Loss: 1.772 :  21%|██▏       | 2149/10000 [02:20<08:04, 16.21it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2151/10000 [02:20<09:55, 13.17it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2153/10000 [02:21<10:38, 12.30it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2155/10000 [02:21<11:16, 11.59it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2157/10000 [02:21<11:28, 11.39it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2159/10000 [02:21<11:39, 11.21it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2161/10000 [02:21<11:44, 11.12it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2163/10000 [02:22<12:06, 10.79it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2165/10000 [02:22<11:37, 11.23it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2167/10000 [02:22<11:18, 11.55it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2169/10000 [02:22<10:53, 11.98it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2171/10000 [02:22<10:32, 12.37it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2173/10000 [02:22<10:23, 12.55it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2175/10000 [02:23<10:19, 12.64it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2177/10000 [02:23<10:20, 12.60it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2179/10000 [02:23<10:47, 12.08it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2181/10000 [02:23<10:33, 12.34it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2183/10000 [02:23<10:05, 12.92it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2185/10000 [02:23<09:47, 13.30it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2187/10000 [02:23<09:30, 13.69it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2189/10000 [02:24<09:13, 14.11it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2191/10000 [02:24<08:55, 14.59it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2193/10000 [02:24<08:39, 15.02it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2195/10000 [02:24<08:31, 15.27it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2197/10000 [02:24<08:20, 15.60it/s]\n",
      "Loss: 1.772 :  22%|██▏       | 2199/10000 [02:24<08:05, 16.08it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2199/10000 [02:24<08:05, 16.08it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2201/10000 [02:24<08:55, 14.57it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2203/10000 [02:24<08:33, 15.19it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2205/10000 [02:25<08:17, 15.67it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2207/10000 [02:25<08:10, 15.90it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2209/10000 [02:25<07:51, 16.51it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2211/10000 [02:25<07:43, 16.80it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2213/10000 [02:25<07:39, 16.96it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2215/10000 [02:25<07:39, 16.93it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2217/10000 [02:25<07:31, 17.25it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2219/10000 [02:25<07:31, 17.22it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2221/10000 [02:25<07:28, 17.35it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2223/10000 [02:26<07:30, 17.27it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2225/10000 [02:26<07:29, 17.31it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2227/10000 [02:26<07:26, 17.40it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2229/10000 [02:26<07:26, 17.40it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2231/10000 [02:26<07:36, 17.03it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2233/10000 [02:26<07:37, 16.96it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2235/10000 [02:26<07:26, 17.37it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2237/10000 [02:26<07:20, 17.62it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2239/10000 [02:27<07:16, 17.79it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2241/10000 [02:27<07:21, 17.58it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2243/10000 [02:27<07:19, 17.65it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2245/10000 [02:27<07:23, 17.50it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2247/10000 [02:27<07:18, 17.66it/s]\n",
      "Loss: 1.910 :  22%|██▏       | 2249/10000 [02:27<07:16, 17.78it/s]\n",
      "Loss: 1.840 :  22%|██▏       | 2249/10000 [02:27<07:16, 17.78it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2251/10000 [02:27<08:09, 15.85it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2253/10000 [02:27<07:59, 16.17it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2255/10000 [02:27<07:44, 16.69it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2257/10000 [02:28<07:37, 16.94it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2259/10000 [02:28<07:31, 17.16it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2261/10000 [02:28<07:39, 16.84it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2263/10000 [02:28<07:38, 16.86it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2265/10000 [02:28<08:16, 15.58it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2267/10000 [02:28<08:42, 14.81it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2269/10000 [02:28<08:41, 14.84it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2271/10000 [02:29<08:40, 14.86it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2273/10000 [02:29<08:27, 15.23it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2275/10000 [02:29<08:19, 15.48it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2277/10000 [02:29<08:16, 15.55it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2279/10000 [02:29<08:04, 15.95it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2281/10000 [02:29<07:56, 16.20it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2283/10000 [02:29<07:41, 16.73it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2285/10000 [02:29<07:43, 16.65it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2287/10000 [02:29<07:47, 16.50it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2289/10000 [02:30<07:46, 16.52it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2291/10000 [02:30<07:43, 16.64it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2293/10000 [02:30<07:37, 16.85it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2295/10000 [02:30<07:31, 17.08it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2297/10000 [02:30<07:26, 17.25it/s]\n",
      "Loss: 1.840 :  23%|██▎       | 2299/10000 [02:30<07:19, 17.51it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2299/10000 [02:30<07:19, 17.51it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2301/10000 [02:30<08:14, 15.58it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2303/10000 [02:30<08:04, 15.90it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2305/10000 [02:31<08:00, 16.03it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2307/10000 [02:31<07:46, 16.50it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2309/10000 [02:31<07:42, 16.63it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2311/10000 [02:31<07:38, 16.77it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2313/10000 [02:31<08:00, 15.99it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2315/10000 [02:31<08:00, 16.01it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2317/10000 [02:31<07:53, 16.24it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2319/10000 [02:31<08:01, 15.95it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2321/10000 [02:32<08:05, 15.81it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2323/10000 [02:32<08:26, 15.17it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2325/10000 [02:32<08:32, 14.98it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2327/10000 [02:32<08:26, 15.16it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2329/10000 [02:32<08:21, 15.31it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2331/10000 [02:32<08:16, 15.44it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2333/10000 [02:32<08:08, 15.70it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2335/10000 [02:32<08:01, 15.91it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2337/10000 [02:33<08:00, 15.95it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2339/10000 [02:33<10:09, 12.57it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2341/10000 [02:33<10:14, 12.46it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2343/10000 [02:33<10:03, 12.69it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2345/10000 [02:33<10:12, 12.50it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2347/10000 [02:34<11:10, 11.41it/s]\n",
      "Loss: 1.545 :  23%|██▎       | 2349/10000 [02:34<11:35, 11.00it/s]\n",
      "Loss: 1.921 :  23%|██▎       | 2349/10000 [02:34<11:35, 11.00it/s]\n",
      "Loss: 1.921 :  24%|██▎       | 2351/10000 [02:34<12:55,  9.87it/s]\n",
      "Loss: 1.921 :  24%|██▎       | 2353/10000 [02:34<12:22, 10.31it/s]\n",
      "Loss: 1.921 :  24%|██▎       | 2355/10000 [02:34<11:59, 10.63it/s]\n",
      "Loss: 1.921 :  24%|██▎       | 2357/10000 [02:35<11:41, 10.89it/s]\n",
      "Loss: 1.921 :  24%|██▎       | 2359/10000 [02:35<11:34, 10.99it/s]\n",
      "Loss: 1.921 :  24%|██▎       | 2361/10000 [02:35<11:15, 11.31it/s]\n",
      "Loss: 1.921 :  24%|██▎       | 2363/10000 [02:35<10:58, 11.60it/s]\n",
      "Loss: 1.921 :  24%|██▎       | 2365/10000 [02:35<10:34, 12.03it/s]\n",
      "Loss: 1.921 :  24%|██▎       | 2367/10000 [02:35<10:12, 12.46it/s]\n",
      "Loss: 1.921 :  24%|██▎       | 2369/10000 [02:35<09:54, 12.83it/s]\n",
      "Loss: 1.921 :  24%|██▎       | 2371/10000 [02:36<09:49, 12.94it/s]\n",
      "Loss: 1.921 :  24%|██▎       | 2373/10000 [02:36<09:35, 13.24it/s]\n",
      "Loss: 1.921 :  24%|██▍       | 2375/10000 [02:36<09:20, 13.62it/s]\n",
      "Loss: 1.921 :  24%|██▍       | 2377/10000 [02:36<09:15, 13.73it/s]\n",
      "Loss: 1.921 :  24%|██▍       | 2379/10000 [02:36<09:06, 13.94it/s]\n",
      "Loss: 1.921 :  24%|██▍       | 2381/10000 [02:36<08:59, 14.13it/s]\n",
      "Loss: 1.921 :  24%|██▍       | 2383/10000 [02:36<08:59, 14.11it/s]\n",
      "Loss: 1.921 :  24%|██▍       | 2385/10000 [02:37<08:59, 14.12it/s]\n",
      "Loss: 1.921 :  24%|██▍       | 2387/10000 [02:37<08:56, 14.20it/s]\n",
      "Loss: 1.921 :  24%|██▍       | 2389/10000 [02:37<08:49, 14.39it/s]\n",
      "Loss: 1.921 :  24%|██▍       | 2391/10000 [02:37<08:52, 14.28it/s]\n",
      "Loss: 1.921 :  24%|██▍       | 2393/10000 [02:37<08:52, 14.28it/s]\n",
      "Loss: 1.921 :  24%|██▍       | 2395/10000 [02:37<08:46, 14.44it/s]\n",
      "Loss: 1.921 :  24%|██▍       | 2397/10000 [02:37<08:40, 14.60it/s]\n",
      "Loss: 1.921 :  24%|██▍       | 2399/10000 [02:38<08:45, 14.48it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2399/10000 [02:38<08:45, 14.48it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2401/10000 [02:38<10:06, 12.53it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2403/10000 [02:38<10:27, 12.10it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2405/10000 [02:38<11:04, 11.42it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2407/10000 [02:38<10:59, 11.51it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2409/10000 [02:39<11:38, 10.87it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2411/10000 [02:39<11:39, 10.85it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2413/10000 [02:39<12:00, 10.53it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2415/10000 [02:39<12:02, 10.50it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2417/10000 [02:39<11:34, 10.91it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2419/10000 [02:39<11:03, 11.43it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2421/10000 [02:40<10:32, 11.99it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2423/10000 [02:40<10:13, 12.36it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2425/10000 [02:40<09:44, 12.95it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2427/10000 [02:40<09:25, 13.39it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2429/10000 [02:40<09:02, 13.95it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2431/10000 [02:40<08:40, 14.54it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2433/10000 [02:40<08:24, 14.99it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2435/10000 [02:40<08:13, 15.33it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2437/10000 [02:41<07:49, 16.10it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2439/10000 [02:41<07:43, 16.31it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2441/10000 [02:41<07:43, 16.30it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2443/10000 [02:41<07:38, 16.48it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2445/10000 [02:41<07:44, 16.25it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2447/10000 [02:41<07:38, 16.47it/s]\n",
      "Loss: 1.803 :  24%|██▍       | 2449/10000 [02:41<07:36, 16.54it/s]\n",
      "Loss: 1.779 :  24%|██▍       | 2449/10000 [02:41<07:36, 16.54it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2451/10000 [02:41<08:09, 15.43it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2453/10000 [02:42<07:49, 16.09it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2455/10000 [02:42<07:38, 16.45it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2457/10000 [02:42<07:30, 16.75it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2459/10000 [02:42<07:24, 16.98it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2461/10000 [02:42<07:30, 16.75it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2463/10000 [02:42<07:21, 17.06it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2465/10000 [02:42<07:13, 17.39it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2467/10000 [02:42<07:16, 17.27it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2469/10000 [02:43<07:21, 17.07it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2471/10000 [02:43<07:15, 17.31it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2473/10000 [02:43<07:15, 17.30it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2475/10000 [02:43<07:17, 17.21it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2477/10000 [02:43<07:20, 17.09it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2479/10000 [02:43<07:17, 17.18it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2481/10000 [02:43<07:17, 17.20it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2483/10000 [02:43<07:14, 17.29it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2485/10000 [02:43<07:12, 17.37it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2487/10000 [02:44<07:14, 17.28it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2489/10000 [02:44<07:16, 17.20it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2491/10000 [02:44<07:16, 17.21it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2493/10000 [02:44<07:16, 17.19it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2495/10000 [02:44<07:13, 17.32it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2497/10000 [02:44<07:14, 17.27it/s]\n",
      "Loss: 1.779 :  25%|██▍       | 2499/10000 [02:44<07:53, 15.84it/s]\n",
      "Loss: 1.905 :  25%|██▍       | 2499/10000 [02:44<07:53, 15.84it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2501/10000 [02:44<08:55, 14.00it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2503/10000 [02:45<08:39, 14.44it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2505/10000 [02:45<08:22, 14.92it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2507/10000 [02:45<08:04, 15.47it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2509/10000 [02:45<07:56, 15.73it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2511/10000 [02:45<07:53, 15.81it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2513/10000 [02:45<07:47, 16.03it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2515/10000 [02:45<07:38, 16.32it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2517/10000 [02:45<07:31, 16.58it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2519/10000 [02:46<07:30, 16.59it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2521/10000 [02:46<07:26, 16.74it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2523/10000 [02:46<07:23, 16.85it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2525/10000 [02:46<07:19, 17.00it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2527/10000 [02:46<07:17, 17.09it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2529/10000 [02:46<07:12, 17.29it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2531/10000 [02:46<07:11, 17.29it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2533/10000 [02:46<07:05, 17.54it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2535/10000 [02:46<07:07, 17.48it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2537/10000 [02:47<07:07, 17.47it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2539/10000 [02:47<07:00, 17.74it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2541/10000 [02:47<07:04, 17.57it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2543/10000 [02:47<07:13, 17.22it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2545/10000 [02:47<07:20, 16.94it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2547/10000 [02:47<07:13, 17.21it/s]\n",
      "Loss: 1.905 :  25%|██▌       | 2549/10000 [02:47<07:09, 17.36it/s]\n",
      "Loss: 1.551 :  25%|██▌       | 2549/10000 [02:47<07:09, 17.36it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2551/10000 [02:47<07:55, 15.68it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2553/10000 [02:48<07:45, 16.01it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2555/10000 [02:48<07:29, 16.57it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2557/10000 [02:48<07:18, 16.96it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2559/10000 [02:48<07:23, 16.76it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2561/10000 [02:48<07:21, 16.86it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2563/10000 [02:48<07:14, 17.12it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2565/10000 [02:48<07:03, 17.58it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2567/10000 [02:48<07:07, 17.40it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2569/10000 [02:48<07:04, 17.51it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2571/10000 [02:49<07:13, 17.15it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2573/10000 [02:49<07:12, 17.16it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2575/10000 [02:49<07:11, 17.20it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2577/10000 [02:49<07:07, 17.35it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2579/10000 [02:49<07:08, 17.33it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2581/10000 [02:49<07:11, 17.21it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2583/10000 [02:49<07:04, 17.48it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2585/10000 [02:49<07:04, 17.47it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2587/10000 [02:50<07:30, 16.46it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2589/10000 [02:50<08:07, 15.21it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2591/10000 [02:50<08:14, 14.99it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2593/10000 [02:50<08:07, 15.21it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2595/10000 [02:50<07:52, 15.67it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2597/10000 [02:50<07:48, 15.80it/s]\n",
      "Loss: 1.551 :  26%|██▌       | 2599/10000 [02:50<07:37, 16.18it/s]\n",
      "Loss: 1.546 :  26%|██▌       | 2599/10000 [02:50<07:37, 16.18it/s]\n",
      "Loss: 1.546 :  26%|██▌       | 2601/10000 [02:50<08:23, 14.69it/s]\n",
      "Loss: 1.546 :  26%|██▌       | 2603/10000 [02:51<08:00, 15.39it/s]\n",
      "Loss: 1.546 :  26%|██▌       | 2605/10000 [02:51<07:43, 15.95it/s]\n",
      "Loss: 1.546 :  26%|██▌       | 2607/10000 [02:51<07:36, 16.21it/s]\n",
      "Loss: 1.546 :  26%|██▌       | 2609/10000 [02:51<07:27, 16.50it/s]\n",
      "Loss: 1.546 :  26%|██▌       | 2611/10000 [02:51<07:25, 16.57it/s]\n",
      "Loss: 1.546 :  26%|██▌       | 2613/10000 [02:51<07:22, 16.68it/s]\n",
      "Loss: 1.546 :  26%|██▌       | 2615/10000 [02:51<07:19, 16.81it/s]\n",
      "Loss: 1.546 :  26%|██▌       | 2617/10000 [02:51<07:18, 16.84it/s]\n",
      "Loss: 1.546 :  26%|██▌       | 2619/10000 [02:52<07:28, 16.45it/s]\n",
      "Loss: 1.546 :  26%|██▌       | 2621/10000 [02:52<07:35, 16.21it/s]\n",
      "Loss: 1.546 :  26%|██▌       | 2623/10000 [02:52<07:36, 16.16it/s]\n",
      "Loss: 1.546 :  26%|██▋       | 2625/10000 [02:52<07:34, 16.23it/s]\n",
      "Loss: 1.546 :  26%|██▋       | 2627/10000 [02:52<07:29, 16.42it/s]\n",
      "Loss: 1.546 :  26%|██▋       | 2629/10000 [02:52<07:22, 16.67it/s]\n",
      "Loss: 1.546 :  26%|██▋       | 2631/10000 [02:52<07:17, 16.83it/s]\n",
      "Loss: 1.546 :  26%|██▋       | 2633/10000 [02:52<07:13, 16.99it/s]\n",
      "Loss: 1.546 :  26%|██▋       | 2635/10000 [02:53<07:12, 17.04it/s]\n",
      "Loss: 1.546 :  26%|██▋       | 2637/10000 [02:53<07:10, 17.11it/s]\n",
      "Loss: 1.546 :  26%|██▋       | 2639/10000 [02:53<07:09, 17.16it/s]\n",
      "Loss: 1.546 :  26%|██▋       | 2641/10000 [02:53<07:04, 17.34it/s]\n",
      "Loss: 1.546 :  26%|██▋       | 2643/10000 [02:53<06:55, 17.72it/s]\n",
      "Loss: 1.546 :  26%|██▋       | 2645/10000 [02:53<06:46, 18.11it/s]\n",
      "Loss: 1.546 :  26%|██▋       | 2647/10000 [02:53<06:55, 17.72it/s]\n",
      "Loss: 1.546 :  26%|██▋       | 2649/10000 [02:53<07:01, 17.42it/s]\n",
      "Loss: 1.574 :  26%|██▋       | 2649/10000 [02:53<07:01, 17.42it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2651/10000 [02:53<07:46, 15.76it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2653/10000 [02:54<07:37, 16.04it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2655/10000 [02:54<07:19, 16.72it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2657/10000 [02:54<07:07, 17.16it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2659/10000 [02:54<07:02, 17.38it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2661/10000 [02:54<06:59, 17.49it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2663/10000 [02:54<06:57, 17.59it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2665/10000 [02:54<06:55, 17.66it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2667/10000 [02:54<06:55, 17.65it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2669/10000 [02:54<06:49, 17.92it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2671/10000 [02:55<06:49, 17.90it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2673/10000 [02:55<06:56, 17.61it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2675/10000 [02:55<07:15, 16.81it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2677/10000 [02:55<07:59, 15.26it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2679/10000 [02:55<08:31, 14.32it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2681/10000 [02:55<08:20, 14.61it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2683/10000 [02:55<08:05, 15.07it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2685/10000 [02:56<08:03, 15.14it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2687/10000 [02:56<08:03, 15.13it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2689/10000 [02:56<07:57, 15.30it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2691/10000 [02:56<07:50, 15.54it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2693/10000 [02:56<07:41, 15.82it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2695/10000 [02:56<07:33, 16.09it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2697/10000 [02:56<07:21, 16.54it/s]\n",
      "Loss: 1.574 :  27%|██▋       | 2699/10000 [02:56<07:11, 16.90it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2699/10000 [02:56<07:11, 16.90it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2701/10000 [02:57<07:56, 15.31it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2703/10000 [02:57<07:47, 15.61it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2705/10000 [02:57<08:39, 14.05it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2707/10000 [02:57<08:22, 14.52it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2709/10000 [02:57<08:13, 14.76it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2711/10000 [02:57<08:02, 15.09it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2713/10000 [02:57<07:50, 15.49it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2715/10000 [02:57<07:40, 15.83it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2717/10000 [02:58<07:27, 16.29it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2719/10000 [02:58<07:25, 16.33it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2721/10000 [02:58<07:38, 15.87it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2723/10000 [02:58<07:51, 15.42it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2725/10000 [02:58<08:06, 14.95it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2727/10000 [02:58<08:23, 14.44it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2729/10000 [02:58<08:33, 14.15it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2731/10000 [02:59<08:37, 14.05it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2733/10000 [02:59<08:22, 14.46it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2735/10000 [02:59<08:06, 14.93it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2737/10000 [02:59<07:57, 15.20it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2739/10000 [02:59<07:49, 15.47it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2741/10000 [02:59<07:42, 15.71it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2743/10000 [02:59<07:38, 15.84it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2745/10000 [02:59<07:31, 16.08it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2747/10000 [03:00<07:16, 16.62it/s]\n",
      "Loss: 1.693 :  27%|██▋       | 2749/10000 [03:00<07:13, 16.72it/s]\n",
      "Loss: 1.670 :  27%|██▋       | 2749/10000 [03:00<07:13, 16.72it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2751/10000 [03:00<08:07, 14.87it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2753/10000 [03:00<07:47, 15.49it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2755/10000 [03:00<07:33, 15.96it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2757/10000 [03:00<07:52, 15.32it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2759/10000 [03:00<08:14, 14.64it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2761/10000 [03:00<08:14, 14.65it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2763/10000 [03:01<08:00, 15.08it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2765/10000 [03:01<07:50, 15.39it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2767/10000 [03:01<07:51, 15.35it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2769/10000 [03:01<07:44, 15.56it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2771/10000 [03:01<07:29, 16.08it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2773/10000 [03:01<07:16, 16.54it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2775/10000 [03:01<07:17, 16.52it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2777/10000 [03:01<07:07, 16.90it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2779/10000 [03:02<07:06, 16.92it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2781/10000 [03:02<07:03, 17.06it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2783/10000 [03:02<07:09, 16.79it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2785/10000 [03:02<07:03, 17.05it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2787/10000 [03:02<06:53, 17.43it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2789/10000 [03:02<06:57, 17.25it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2791/10000 [03:02<07:21, 16.33it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2793/10000 [03:02<07:17, 16.48it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2795/10000 [03:03<07:16, 16.52it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2797/10000 [03:03<07:08, 16.83it/s]\n",
      "Loss: 1.670 :  28%|██▊       | 2799/10000 [03:03<07:05, 16.93it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2799/10000 [03:03<07:05, 16.93it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2801/10000 [03:03<07:58, 15.05it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2803/10000 [03:03<07:43, 15.52it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2805/10000 [03:03<07:37, 15.71it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2807/10000 [03:03<07:25, 16.14it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2809/10000 [03:03<07:22, 16.26it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2811/10000 [03:04<07:18, 16.41it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2813/10000 [03:04<07:12, 16.62it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2815/10000 [03:04<07:13, 16.56it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2817/10000 [03:04<07:09, 16.71it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2819/10000 [03:04<07:07, 16.81it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2821/10000 [03:04<06:55, 17.29it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2823/10000 [03:04<06:57, 17.20it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2825/10000 [03:04<06:53, 17.34it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2827/10000 [03:04<06:47, 17.61it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2829/10000 [03:05<06:47, 17.60it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2831/10000 [03:05<06:43, 17.79it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2833/10000 [03:05<06:43, 17.77it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2835/10000 [03:05<06:43, 17.77it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2837/10000 [03:05<06:46, 17.62it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2839/10000 [03:05<06:45, 17.65it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2841/10000 [03:05<06:49, 17.49it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2843/10000 [03:05<06:47, 17.56it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2845/10000 [03:05<06:52, 17.35it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2847/10000 [03:06<07:32, 15.79it/s]\n",
      "Loss: 1.361 :  28%|██▊       | 2849/10000 [03:06<07:52, 15.14it/s]\n",
      "Loss: 1.480 :  28%|██▊       | 2849/10000 [03:06<07:52, 15.14it/s]\n",
      "Loss: 1.480 :  29%|██▊       | 2851/10000 [03:06<08:34, 13.88it/s]\n",
      "Loss: 1.480 :  29%|██▊       | 2853/10000 [03:06<08:12, 14.52it/s]\n",
      "Loss: 1.480 :  29%|██▊       | 2855/10000 [03:06<07:59, 14.90it/s]\n",
      "Loss: 1.480 :  29%|██▊       | 2857/10000 [03:06<07:43, 15.40it/s]\n",
      "Loss: 1.480 :  29%|██▊       | 2859/10000 [03:06<07:34, 15.71it/s]\n",
      "Loss: 1.480 :  29%|██▊       | 2861/10000 [03:07<07:35, 15.68it/s]\n",
      "Loss: 1.480 :  29%|██▊       | 2863/10000 [03:07<07:25, 16.01it/s]\n",
      "Loss: 1.480 :  29%|██▊       | 2865/10000 [03:07<07:20, 16.19it/s]\n",
      "Loss: 1.480 :  29%|██▊       | 2867/10000 [03:07<07:20, 16.18it/s]\n",
      "Loss: 1.480 :  29%|██▊       | 2869/10000 [03:07<07:16, 16.32it/s]\n",
      "Loss: 1.480 :  29%|██▊       | 2871/10000 [03:07<07:14, 16.42it/s]\n",
      "Loss: 1.480 :  29%|██▊       | 2873/10000 [03:07<07:19, 16.23it/s]\n",
      "Loss: 1.480 :  29%|██▉       | 2875/10000 [03:07<07:13, 16.44it/s]\n",
      "Loss: 1.480 :  29%|██▉       | 2877/10000 [03:08<07:14, 16.38it/s]\n",
      "Loss: 1.480 :  29%|██▉       | 2879/10000 [03:08<07:08, 16.62it/s]\n",
      "Loss: 1.480 :  29%|██▉       | 2881/10000 [03:08<06:58, 17.00it/s]\n",
      "Loss: 1.480 :  29%|██▉       | 2883/10000 [03:08<07:13, 16.43it/s]\n",
      "Loss: 1.480 :  29%|██▉       | 2885/10000 [03:08<07:07, 16.64it/s]\n",
      "Loss: 1.480 :  29%|██▉       | 2887/10000 [03:08<07:02, 16.85it/s]\n",
      "Loss: 1.480 :  29%|██▉       | 2889/10000 [03:08<06:58, 16.99it/s]\n",
      "Loss: 1.480 :  29%|██▉       | 2891/10000 [03:08<06:56, 17.07it/s]\n",
      "Loss: 1.480 :  29%|██▉       | 2893/10000 [03:08<06:55, 17.11it/s]\n",
      "Loss: 1.480 :  29%|██▉       | 2895/10000 [03:09<06:55, 17.10it/s]\n",
      "Loss: 1.480 :  29%|██▉       | 2897/10000 [03:09<06:54, 17.15it/s]\n",
      "Loss: 1.480 :  29%|██▉       | 2899/10000 [03:09<06:56, 17.04it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2899/10000 [03:09<06:56, 17.04it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2901/10000 [03:09<07:40, 15.42it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2903/10000 [03:09<07:25, 15.92it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2905/10000 [03:09<07:19, 16.15it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2907/10000 [03:09<07:03, 16.76it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2909/10000 [03:09<06:59, 16.91it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2911/10000 [03:10<06:55, 17.06it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2913/10000 [03:10<06:48, 17.35it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2915/10000 [03:10<06:51, 17.22it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2917/10000 [03:10<06:48, 17.35it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2919/10000 [03:10<06:42, 17.61it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2921/10000 [03:10<06:43, 17.56it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2923/10000 [03:10<06:37, 17.79it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2925/10000 [03:10<06:37, 17.81it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2927/10000 [03:10<06:41, 17.61it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2929/10000 [03:11<06:42, 17.55it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2931/10000 [03:11<06:47, 17.33it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2933/10000 [03:11<07:09, 16.47it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2935/10000 [03:11<07:48, 15.07it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2937/10000 [03:11<07:55, 14.84it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2939/10000 [03:11<07:49, 15.04it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2941/10000 [03:11<07:45, 15.17it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2943/10000 [03:11<07:34, 15.52it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2945/10000 [03:12<07:26, 15.81it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2947/10000 [03:12<07:21, 15.97it/s]\n",
      "Loss: 1.739 :  29%|██▉       | 2949/10000 [03:12<07:20, 16.00it/s]\n",
      "Loss: 1.488 :  29%|██▉       | 2949/10000 [03:12<07:20, 16.00it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2951/10000 [03:12<08:01, 14.63it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2953/10000 [03:12<07:41, 15.26it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2955/10000 [03:12<07:22, 15.91it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2957/10000 [03:12<07:13, 16.23it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2959/10000 [03:12<07:08, 16.43it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2961/10000 [03:13<07:02, 16.67it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2963/10000 [03:13<06:51, 17.10it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2965/10000 [03:13<06:44, 17.37it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2967/10000 [03:13<06:43, 17.44it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2969/10000 [03:13<06:36, 17.73it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2971/10000 [03:13<06:33, 17.87it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2973/10000 [03:13<06:34, 17.80it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2975/10000 [03:13<06:36, 17.71it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2977/10000 [03:14<06:42, 17.46it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2979/10000 [03:14<06:39, 17.57it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2981/10000 [03:14<06:40, 17.53it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2983/10000 [03:14<06:41, 17.48it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2985/10000 [03:14<06:39, 17.55it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2987/10000 [03:14<06:37, 17.65it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2989/10000 [03:14<06:27, 18.08it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2991/10000 [03:14<06:32, 17.87it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2993/10000 [03:14<06:47, 17.18it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2995/10000 [03:15<06:55, 16.86it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2997/10000 [03:15<06:50, 17.04it/s]\n",
      "Loss: 1.488 :  30%|██▉       | 2999/10000 [03:15<07:01, 16.62it/s]\n",
      "Loss: 1.676 :  30%|██▉       | 2999/10000 [03:15<07:01, 16.62it/s]\n",
      "Loss: 1.676 :  30%|███       | 3001/10000 [03:15<07:39, 15.22it/s]\n",
      "Loss: 1.676 :  30%|███       | 3003/10000 [03:15<07:26, 15.66it/s]\n",
      "Loss: 1.676 :  30%|███       | 3005/10000 [03:15<07:17, 15.98it/s]\n",
      "Loss: 1.676 :  30%|███       | 3007/10000 [03:15<07:13, 16.13it/s]\n",
      "Loss: 1.676 :  30%|███       | 3009/10000 [03:15<07:02, 16.56it/s]\n",
      "Loss: 1.676 :  30%|███       | 3011/10000 [03:16<07:12, 16.14it/s]\n",
      "Loss: 1.676 :  30%|███       | 3013/10000 [03:16<07:25, 15.69it/s]\n",
      "Loss: 1.676 :  30%|███       | 3015/10000 [03:16<07:26, 15.64it/s]\n",
      "Loss: 1.676 :  30%|███       | 3017/10000 [03:16<07:29, 15.53it/s]\n",
      "Loss: 1.676 :  30%|███       | 3019/10000 [03:16<07:27, 15.61it/s]\n",
      "Loss: 1.676 :  30%|███       | 3021/10000 [03:16<08:02, 14.45it/s]\n",
      "Loss: 1.676 :  30%|███       | 3023/10000 [03:16<08:15, 14.09it/s]\n",
      "Loss: 1.676 :  30%|███       | 3025/10000 [03:17<08:12, 14.17it/s]\n",
      "Loss: 1.676 :  30%|███       | 3027/10000 [03:17<08:08, 14.29it/s]\n",
      "Loss: 1.676 :  30%|███       | 3029/10000 [03:17<08:13, 14.13it/s]\n",
      "Loss: 1.676 :  30%|███       | 3031/10000 [03:17<08:06, 14.33it/s]\n",
      "Loss: 1.676 :  30%|███       | 3033/10000 [03:17<08:01, 14.48it/s]\n",
      "Loss: 1.676 :  30%|███       | 3035/10000 [03:17<07:54, 14.67it/s]\n",
      "Loss: 1.676 :  30%|███       | 3037/10000 [03:17<07:49, 14.82it/s]\n",
      "Loss: 1.676 :  30%|███       | 3039/10000 [03:17<07:40, 15.13it/s]\n",
      "Loss: 1.676 :  30%|███       | 3041/10000 [03:18<07:40, 15.12it/s]\n",
      "Loss: 1.676 :  30%|███       | 3043/10000 [03:18<07:29, 15.48it/s]\n",
      "Loss: 1.676 :  30%|███       | 3045/10000 [03:18<07:25, 15.62it/s]\n",
      "Loss: 1.676 :  30%|███       | 3047/10000 [03:18<07:21, 15.76it/s]\n",
      "Loss: 1.676 :  30%|███       | 3049/10000 [03:18<07:05, 16.33it/s]\n",
      "Loss: 1.553 :  30%|███       | 3049/10000 [03:18<07:05, 16.33it/s]\n",
      "Loss: 1.553 :  31%|███       | 3051/10000 [03:18<07:41, 15.06it/s]\n",
      "Loss: 1.553 :  31%|███       | 3053/10000 [03:18<07:23, 15.66it/s]\n",
      "Loss: 1.553 :  31%|███       | 3055/10000 [03:18<07:16, 15.92it/s]\n",
      "Loss: 1.553 :  31%|███       | 3057/10000 [03:19<07:07, 16.24it/s]\n",
      "Loss: 1.553 :  31%|███       | 3059/10000 [03:19<06:56, 16.65it/s]\n",
      "Loss: 1.553 :  31%|███       | 3061/10000 [03:19<06:49, 16.94it/s]\n",
      "Loss: 1.553 :  31%|███       | 3063/10000 [03:19<06:52, 16.81it/s]\n",
      "Loss: 1.553 :  31%|███       | 3065/10000 [03:19<06:54, 16.73it/s]\n",
      "Loss: 1.553 :  31%|███       | 3067/10000 [03:19<06:54, 16.74it/s]\n",
      "Loss: 1.553 :  31%|███       | 3069/10000 [03:19<06:45, 17.07it/s]\n",
      "Loss: 1.553 :  31%|███       | 3071/10000 [03:19<06:42, 17.21it/s]\n",
      "Loss: 1.553 :  31%|███       | 3073/10000 [03:20<06:46, 17.06it/s]\n",
      "Loss: 1.553 :  31%|███       | 3075/10000 [03:20<06:56, 16.62it/s]\n",
      "Loss: 1.553 :  31%|███       | 3077/10000 [03:20<07:01, 16.41it/s]\n",
      "Loss: 1.553 :  31%|███       | 3079/10000 [03:20<07:10, 16.07it/s]\n",
      "Loss: 1.553 :  31%|███       | 3081/10000 [03:20<07:12, 16.01it/s]\n",
      "Loss: 1.553 :  31%|███       | 3083/10000 [03:20<07:18, 15.78it/s]\n",
      "Loss: 1.553 :  31%|███       | 3085/10000 [03:20<07:16, 15.83it/s]\n",
      "Loss: 1.553 :  31%|███       | 3087/10000 [03:20<07:17, 15.80it/s]\n",
      "Loss: 1.553 :  31%|███       | 3089/10000 [03:21<07:15, 15.88it/s]\n",
      "Loss: 1.553 :  31%|███       | 3091/10000 [03:21<07:16, 15.84it/s]\n",
      "Loss: 1.553 :  31%|███       | 3093/10000 [03:21<07:14, 15.91it/s]\n",
      "Loss: 1.553 :  31%|███       | 3095/10000 [03:21<07:08, 16.10it/s]\n",
      "Loss: 1.553 :  31%|███       | 3097/10000 [03:21<07:08, 16.11it/s]\n",
      "Loss: 1.553 :  31%|███       | 3099/10000 [03:21<07:04, 16.24it/s]\n",
      "Loss: 1.520 :  31%|███       | 3099/10000 [03:21<07:04, 16.24it/s]\n",
      "Loss: 1.520 :  31%|███       | 3101/10000 [03:21<07:46, 14.78it/s]\n",
      "Loss: 1.520 :  31%|███       | 3103/10000 [03:21<07:40, 14.97it/s]\n",
      "Loss: 1.520 :  31%|███       | 3105/10000 [03:22<08:24, 13.67it/s]\n",
      "Loss: 1.520 :  31%|███       | 3107/10000 [03:22<08:49, 13.01it/s]\n",
      "Loss: 1.520 :  31%|███       | 3109/10000 [03:22<08:27, 13.58it/s]\n",
      "Loss: 1.520 :  31%|███       | 3111/10000 [03:22<08:25, 13.63it/s]\n",
      "Loss: 1.520 :  31%|███       | 3113/10000 [03:22<11:05, 10.35it/s]\n",
      "Loss: 1.520 :  31%|███       | 3115/10000 [03:23<10:53, 10.53it/s]\n",
      "Loss: 1.520 :  31%|███       | 3117/10000 [03:23<10:11, 11.25it/s]\n",
      "Loss: 1.520 :  31%|███       | 3119/10000 [03:23<09:37, 11.92it/s]\n",
      "Loss: 1.520 :  31%|███       | 3121/10000 [03:23<09:12, 12.46it/s]\n",
      "Loss: 1.520 :  31%|███       | 3123/10000 [03:23<09:02, 12.67it/s]\n",
      "Loss: 1.520 :  31%|███▏      | 3125/10000 [03:23<08:44, 13.11it/s]\n",
      "Loss: 1.520 :  31%|███▏      | 3127/10000 [03:23<08:26, 13.57it/s]\n",
      "Loss: 1.520 :  31%|███▏      | 3129/10000 [03:24<08:09, 14.03it/s]\n",
      "Loss: 1.520 :  31%|███▏      | 3131/10000 [03:24<07:42, 14.84it/s]\n",
      "Loss: 1.520 :  31%|███▏      | 3133/10000 [03:24<07:31, 15.20it/s]\n",
      "Loss: 1.520 :  31%|███▏      | 3135/10000 [03:24<07:29, 15.29it/s]\n",
      "Loss: 1.520 :  31%|███▏      | 3137/10000 [03:24<07:26, 15.36it/s]\n",
      "Loss: 1.520 :  31%|███▏      | 3139/10000 [03:24<07:25, 15.40it/s]\n",
      "Loss: 1.520 :  31%|███▏      | 3141/10000 [03:24<07:20, 15.56it/s]\n",
      "Loss: 1.520 :  31%|███▏      | 3143/10000 [03:24<07:11, 15.89it/s]\n",
      "Loss: 1.520 :  31%|███▏      | 3145/10000 [03:25<07:00, 16.29it/s]\n",
      "Loss: 1.520 :  31%|███▏      | 3147/10000 [03:25<06:53, 16.56it/s]\n",
      "Loss: 1.520 :  31%|███▏      | 3149/10000 [03:25<06:44, 16.94it/s]\n",
      "Loss: 1.530 :  31%|███▏      | 3149/10000 [03:25<06:44, 16.94it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3151/10000 [03:25<07:21, 15.53it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3153/10000 [03:25<07:07, 16.03it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3155/10000 [03:25<06:47, 16.81it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3158/10000 [03:25<06:22, 17.87it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3160/10000 [03:25<06:14, 18.26it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3162/10000 [03:26<06:15, 18.20it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3164/10000 [03:26<06:10, 18.47it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3166/10000 [03:26<06:02, 18.85it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3168/10000 [03:26<05:58, 19.06it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3171/10000 [03:26<05:49, 19.53it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3173/10000 [03:26<05:49, 19.53it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3175/10000 [03:26<05:50, 19.48it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3177/10000 [03:26<05:53, 19.33it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3179/10000 [03:26<06:13, 18.27it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3181/10000 [03:27<06:47, 16.75it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3183/10000 [03:27<07:28, 15.19it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3185/10000 [03:27<08:56, 12.70it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3187/10000 [03:27<10:13, 11.10it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3189/10000 [03:27<10:11, 11.15it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3191/10000 [03:28<10:12, 11.12it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3193/10000 [03:28<10:26, 10.87it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3195/10000 [03:28<10:15, 11.06it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3197/10000 [03:28<10:05, 11.24it/s]\n",
      "Loss: 1.530 :  32%|███▏      | 3199/10000 [03:28<09:57, 11.39it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3199/10000 [03:28<09:57, 11.39it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3201/10000 [03:28<10:37, 10.67it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3203/10000 [03:29<10:16, 11.03it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3205/10000 [03:29<09:49, 11.52it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3207/10000 [03:29<09:42, 11.65it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3209/10000 [03:29<09:22, 12.08it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3211/10000 [03:29<09:14, 12.24it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3213/10000 [03:29<08:57, 12.64it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3215/10000 [03:30<09:05, 12.44it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3217/10000 [03:30<09:23, 12.04it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3219/10000 [03:30<11:55,  9.47it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3221/10000 [03:30<11:05, 10.18it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3223/10000 [03:30<09:57, 11.34it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3225/10000 [03:30<09:07, 12.37it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3227/10000 [03:31<08:51, 12.74it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3229/10000 [03:31<08:43, 12.93it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3231/10000 [03:31<08:38, 13.04it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3233/10000 [03:31<08:30, 13.24it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3235/10000 [03:31<08:34, 13.15it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3237/10000 [03:31<08:33, 13.17it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3239/10000 [03:32<08:32, 13.19it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3241/10000 [03:32<08:34, 13.13it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3243/10000 [03:32<08:39, 13.01it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3245/10000 [03:32<10:54, 10.32it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3247/10000 [03:32<11:09, 10.08it/s]\n",
      "Loss: 1.268 :  32%|███▏      | 3249/10000 [03:33<11:01, 10.20it/s]\n",
      "Loss: 1.569 :  32%|███▏      | 3249/10000 [03:33<11:01, 10.20it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3251/10000 [03:33<11:20,  9.92it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3253/10000 [03:33<11:05, 10.14it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3255/10000 [03:33<10:25, 10.79it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3257/10000 [03:33<10:06, 11.12it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3259/10000 [03:33<09:41, 11.59it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3261/10000 [03:34<09:04, 12.37it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3263/10000 [03:34<09:02, 12.41it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3265/10000 [03:34<08:48, 12.74it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3267/10000 [03:34<08:38, 12.97it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3269/10000 [03:34<08:34, 13.08it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3271/10000 [03:34<08:07, 13.80it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3273/10000 [03:34<07:41, 14.57it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3275/10000 [03:34<07:17, 15.36it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3277/10000 [03:35<07:00, 15.97it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3279/10000 [03:35<06:44, 16.62it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3281/10000 [03:35<06:40, 16.79it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3283/10000 [03:35<06:40, 16.76it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3285/10000 [03:35<07:02, 15.89it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3287/10000 [03:35<07:19, 15.29it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3289/10000 [03:35<07:28, 14.95it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3291/10000 [03:36<07:42, 14.51it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3293/10000 [03:36<10:22, 10.77it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3295/10000 [03:36<09:47, 11.42it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3297/10000 [03:36<09:14, 12.08it/s]\n",
      "Loss: 1.569 :  33%|███▎      | 3299/10000 [03:36<09:03, 12.33it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3299/10000 [03:36<09:03, 12.33it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3301/10000 [03:36<09:29, 11.76it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3303/10000 [03:37<09:06, 12.25it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3305/10000 [03:37<08:43, 12.78it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3307/10000 [03:37<08:35, 12.98it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3309/10000 [03:37<08:34, 13.00it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3311/10000 [03:37<08:17, 13.45it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3313/10000 [03:37<08:10, 13.62it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3315/10000 [03:37<07:55, 14.04it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3317/10000 [03:38<07:54, 14.09it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3319/10000 [03:38<08:25, 13.21it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3321/10000 [03:38<08:40, 12.84it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3323/10000 [03:38<08:18, 13.40it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3325/10000 [03:38<08:01, 13.85it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3327/10000 [03:38<08:01, 13.85it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3329/10000 [03:38<07:54, 14.07it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3331/10000 [03:39<07:57, 13.97it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3333/10000 [03:39<08:08, 13.66it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3335/10000 [03:39<08:13, 13.49it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3337/10000 [03:39<08:14, 13.47it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3339/10000 [03:39<07:56, 13.99it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3341/10000 [03:39<07:27, 14.88it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3343/10000 [03:39<07:06, 15.62it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3345/10000 [03:40<06:51, 16.18it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3347/10000 [03:40<07:29, 14.80it/s]\n",
      "Loss: 1.321 :  33%|███▎      | 3349/10000 [03:40<06:56, 15.99it/s]\n",
      "Loss: 1.282 :  33%|███▎      | 3349/10000 [03:40<06:56, 15.99it/s]\n",
      "Loss: 1.282 :  34%|███▎      | 3351/10000 [03:40<07:14, 15.30it/s]\n",
      "Loss: 1.282 :  34%|███▎      | 3353/10000 [03:40<06:48, 16.26it/s]\n",
      "Loss: 1.282 :  34%|███▎      | 3355/10000 [03:40<06:33, 16.90it/s]\n",
      "Loss: 1.282 :  34%|███▎      | 3358/10000 [03:40<06:11, 17.90it/s]\n",
      "Loss: 1.282 :  34%|███▎      | 3360/10000 [03:40<06:03, 18.28it/s]\n",
      "Loss: 1.282 :  34%|███▎      | 3362/10000 [03:41<05:55, 18.69it/s]\n",
      "Loss: 1.282 :  34%|███▎      | 3364/10000 [03:41<05:51, 18.86it/s]\n",
      "Loss: 1.282 :  34%|███▎      | 3366/10000 [03:41<05:59, 18.45it/s]\n",
      "Loss: 1.282 :  34%|███▎      | 3368/10000 [03:41<06:20, 17.41it/s]\n",
      "Loss: 1.282 :  34%|███▎      | 3370/10000 [03:41<06:29, 17.03it/s]\n",
      "Loss: 1.282 :  34%|███▎      | 3372/10000 [03:41<06:38, 16.63it/s]\n",
      "Loss: 1.282 :  34%|███▎      | 3374/10000 [03:41<06:47, 16.26it/s]\n",
      "Loss: 1.282 :  34%|███▍      | 3376/10000 [03:41<06:56, 15.91it/s]\n",
      "Loss: 1.282 :  34%|███▍      | 3378/10000 [03:42<07:04, 15.60it/s]\n",
      "Loss: 1.282 :  34%|███▍      | 3380/10000 [03:42<07:45, 14.23it/s]\n",
      "Loss: 1.282 :  34%|███▍      | 3382/10000 [03:42<07:55, 13.90it/s]\n",
      "Loss: 1.282 :  34%|███▍      | 3384/10000 [03:42<08:03, 13.67it/s]\n",
      "Loss: 1.282 :  34%|███▍      | 3386/10000 [03:42<08:09, 13.51it/s]\n",
      "Loss: 1.282 :  34%|███▍      | 3388/10000 [03:42<08:10, 13.49it/s]\n",
      "Loss: 1.282 :  34%|███▍      | 3390/10000 [03:42<08:02, 13.70it/s]\n",
      "Loss: 1.282 :  34%|███▍      | 3392/10000 [03:43<07:58, 13.80it/s]\n",
      "Loss: 1.282 :  34%|███▍      | 3394/10000 [03:43<09:49, 11.21it/s]\n",
      "Loss: 1.282 :  34%|███▍      | 3396/10000 [03:43<09:54, 11.10it/s]\n",
      "Loss: 1.282 :  34%|███▍      | 3398/10000 [03:43<10:07, 10.87it/s]\n",
      "Loss: 1.282 :  34%|███▍      | 3400/10000 [03:43<10:03, 10.94it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3400/10000 [03:43<10:03, 10.94it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3402/10000 [03:44<10:36, 10.37it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3404/10000 [03:44<10:01, 10.96it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3406/10000 [03:44<09:31, 11.53it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3408/10000 [03:44<09:00, 12.19it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3410/10000 [03:44<08:24, 13.06it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3412/10000 [03:44<07:54, 13.88it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3414/10000 [03:44<07:30, 14.61it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3416/10000 [03:45<07:05, 15.46it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3418/10000 [03:45<06:49, 16.08it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3420/10000 [03:45<06:45, 16.21it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3422/10000 [03:45<06:35, 16.62it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3424/10000 [03:45<06:29, 16.90it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3426/10000 [03:45<06:17, 17.43it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3428/10000 [03:45<06:28, 16.93it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3430/10000 [03:45<06:28, 16.89it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3432/10000 [03:45<06:33, 16.69it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3434/10000 [03:46<06:42, 16.30it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3436/10000 [03:46<06:52, 15.91it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3438/10000 [03:46<07:04, 15.46it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3440/10000 [03:46<07:08, 15.31it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3442/10000 [03:46<07:05, 15.40it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3444/10000 [03:46<06:50, 15.97it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3446/10000 [03:46<06:37, 16.48it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3448/10000 [03:46<06:23, 17.10it/s]\n",
      "Loss: 1.300 :  34%|███▍      | 3450/10000 [03:47<06:12, 17.59it/s]\n",
      "Loss: 1.442 :  34%|███▍      | 3450/10000 [03:47<06:12, 17.59it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3452/10000 [03:47<06:38, 16.45it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3454/10000 [03:47<06:29, 16.81it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3456/10000 [03:47<06:14, 17.46it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3458/10000 [03:47<06:03, 18.02it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3460/10000 [03:47<05:56, 18.33it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3462/10000 [03:47<05:48, 18.74it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3464/10000 [03:47<05:47, 18.84it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3466/10000 [03:47<05:59, 18.20it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3468/10000 [03:48<05:56, 18.31it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3470/10000 [03:48<05:55, 18.35it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3472/10000 [03:48<05:50, 18.62it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3474/10000 [03:48<05:49, 18.65it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3476/10000 [03:48<05:48, 18.73it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3479/10000 [03:48<05:35, 19.42it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3481/10000 [03:48<05:37, 19.33it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3484/10000 [03:48<05:37, 19.28it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3486/10000 [03:49<06:15, 17.35it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3488/10000 [03:49<06:29, 16.72it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3490/10000 [03:49<06:25, 16.87it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3492/10000 [03:49<06:38, 16.33it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3494/10000 [03:49<06:45, 16.06it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3496/10000 [03:49<06:57, 15.57it/s]\n",
      "Loss: 1.442 :  35%|███▍      | 3498/10000 [03:49<07:22, 14.71it/s]\n",
      "Loss: 1.442 :  35%|███▌      | 3500/10000 [03:50<07:35, 14.28it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3500/10000 [03:50<07:35, 14.28it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3502/10000 [03:50<08:24, 12.87it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3504/10000 [03:50<08:12, 13.20it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3506/10000 [03:50<07:56, 13.64it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3508/10000 [03:50<07:37, 14.19it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3510/10000 [03:50<07:16, 14.88it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3512/10000 [03:50<07:04, 15.30it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3514/10000 [03:50<06:54, 15.65it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3516/10000 [03:51<06:37, 16.30it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3518/10000 [03:51<06:31, 16.54it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3520/10000 [03:51<06:22, 16.95it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3522/10000 [03:51<06:18, 17.10it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3524/10000 [03:51<06:16, 17.19it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3526/10000 [03:51<06:27, 16.72it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3528/10000 [03:51<06:37, 16.29it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3530/10000 [03:51<06:48, 15.84it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3532/10000 [03:52<06:36, 16.33it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3534/10000 [03:52<06:24, 16.83it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3536/10000 [03:52<06:22, 16.89it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3538/10000 [03:52<06:13, 17.30it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3540/10000 [03:52<06:02, 17.83it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3542/10000 [03:52<05:52, 18.34it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3544/10000 [03:52<05:52, 18.33it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3546/10000 [03:52<05:49, 18.46it/s]\n",
      "Loss: 1.224 :  35%|███▌      | 3549/10000 [03:52<05:42, 18.86it/s]\n",
      "Loss: 1.345 :  35%|███▌      | 3549/10000 [03:53<05:42, 18.86it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3551/10000 [03:53<06:09, 17.44it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3553/10000 [03:53<05:57, 18.04it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3555/10000 [03:53<05:49, 18.46it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3558/10000 [03:53<05:37, 19.07it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3560/10000 [03:53<05:45, 18.62it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3562/10000 [03:53<05:51, 18.30it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3564/10000 [03:53<05:54, 18.16it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3566/10000 [03:53<05:51, 18.33it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3569/10000 [03:54<05:39, 18.92it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3571/10000 [03:54<05:38, 19.00it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3573/10000 [03:54<06:01, 17.78it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3575/10000 [03:54<06:24, 16.73it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3577/10000 [03:54<06:23, 16.73it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3579/10000 [03:54<06:19, 16.90it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3581/10000 [03:54<06:15, 17.10it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3583/10000 [03:54<06:23, 16.72it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3585/10000 [03:55<06:48, 15.72it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3587/10000 [03:55<06:58, 15.34it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3589/10000 [03:55<07:06, 15.04it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3591/10000 [03:55<06:55, 15.43it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3593/10000 [03:55<06:48, 15.68it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3595/10000 [03:55<06:30, 16.40it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3597/10000 [03:55<06:20, 16.84it/s]\n",
      "Loss: 1.345 :  36%|███▌      | 3599/10000 [03:55<06:04, 17.58it/s]\n",
      "Loss: 1.245 :  36%|███▌      | 3599/10000 [03:56<06:04, 17.58it/s]\n",
      "Loss: 1.245 :  36%|███▌      | 3601/10000 [03:56<06:40, 15.99it/s]\n",
      "Loss: 1.245 :  36%|███▌      | 3603/10000 [03:56<06:22, 16.73it/s]\n",
      "Loss: 1.245 :  36%|███▌      | 3605/10000 [03:56<06:09, 17.30it/s]\n",
      "Loss: 1.245 :  36%|███▌      | 3607/10000 [03:56<05:57, 17.88it/s]\n",
      "Loss: 1.245 :  36%|███▌      | 3609/10000 [03:56<05:55, 17.96it/s]\n",
      "Loss: 1.245 :  36%|███▌      | 3611/10000 [03:56<05:49, 18.28it/s]\n",
      "Loss: 1.245 :  36%|███▌      | 3613/10000 [03:56<05:53, 18.05it/s]\n",
      "Loss: 1.245 :  36%|███▌      | 3615/10000 [03:56<08:18, 12.81it/s]\n",
      "Loss: 1.245 :  36%|███▌      | 3617/10000 [03:57<08:06, 13.13it/s]\n",
      "Loss: 1.245 :  36%|███▌      | 3619/10000 [03:57<07:45, 13.69it/s]\n",
      "Loss: 1.245 :  36%|███▌      | 3621/10000 [03:57<07:31, 14.13it/s]\n",
      "Loss: 1.245 :  36%|███▌      | 3623/10000 [03:57<07:22, 14.40it/s]\n",
      "Loss: 1.245 :  36%|███▋      | 3625/10000 [03:57<07:20, 14.47it/s]\n",
      "Loss: 1.245 :  36%|███▋      | 3627/10000 [03:57<07:09, 14.84it/s]\n",
      "Loss: 1.245 :  36%|███▋      | 3629/10000 [03:57<06:58, 15.23it/s]\n",
      "Loss: 1.245 :  36%|███▋      | 3631/10000 [03:57<06:46, 15.66it/s]\n",
      "Loss: 1.245 :  36%|███▋      | 3633/10000 [03:58<06:34, 16.12it/s]\n",
      "Loss: 1.245 :  36%|███▋      | 3635/10000 [03:58<06:31, 16.24it/s]\n",
      "Loss: 1.245 :  36%|███▋      | 3637/10000 [03:58<06:23, 16.58it/s]\n",
      "Loss: 1.245 :  36%|███▋      | 3639/10000 [03:58<06:09, 17.23it/s]\n",
      "Loss: 1.245 :  36%|███▋      | 3641/10000 [03:58<06:05, 17.38it/s]\n",
      "Loss: 1.245 :  36%|███▋      | 3643/10000 [03:58<06:03, 17.49it/s]\n",
      "Loss: 1.245 :  36%|███▋      | 3645/10000 [03:58<05:57, 17.79it/s]\n",
      "Loss: 1.245 :  36%|███▋      | 3647/10000 [03:58<05:47, 18.27it/s]\n",
      "Loss: 1.245 :  36%|███▋      | 3649/10000 [03:58<05:46, 18.32it/s]\n",
      "Loss: 1.181 :  36%|███▋      | 3649/10000 [03:59<05:46, 18.32it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3651/10000 [03:59<06:36, 16.02it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3653/10000 [03:59<06:35, 16.05it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3655/10000 [03:59<06:34, 16.07it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3657/10000 [03:59<06:57, 15.18it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3659/10000 [03:59<07:38, 13.83it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3661/10000 [03:59<07:54, 13.36it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3663/10000 [04:00<07:32, 14.00it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3665/10000 [04:00<07:13, 14.62it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3667/10000 [04:00<07:03, 14.96it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3669/10000 [04:00<06:47, 15.55it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3671/10000 [04:00<06:36, 15.96it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3673/10000 [04:00<06:21, 16.59it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3675/10000 [04:00<06:11, 17.03it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3677/10000 [04:00<06:00, 17.56it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3679/10000 [04:00<05:58, 17.62it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3681/10000 [04:01<05:50, 18.03it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3683/10000 [04:01<05:50, 18.00it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3685/10000 [04:01<05:48, 18.13it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3687/10000 [04:01<05:46, 18.21it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3689/10000 [04:01<05:43, 18.37it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3691/10000 [04:01<05:39, 18.56it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3693/10000 [04:01<05:37, 18.71it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3695/10000 [04:01<05:33, 18.92it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3697/10000 [04:01<05:31, 19.01it/s]\n",
      "Loss: 1.181 :  37%|███▋      | 3699/10000 [04:01<05:33, 18.92it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3699/10000 [04:02<05:33, 18.92it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3701/10000 [04:02<06:04, 17.30it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3704/10000 [04:02<05:40, 18.47it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3706/10000 [04:02<05:35, 18.76it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3708/10000 [04:02<05:32, 18.91it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3710/10000 [04:02<05:43, 18.31it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3712/10000 [04:02<05:40, 18.47it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3714/10000 [04:02<05:33, 18.84it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3716/10000 [04:02<05:32, 18.92it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3719/10000 [04:03<05:23, 19.41it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3721/10000 [04:03<05:23, 19.42it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3723/10000 [04:03<05:30, 19.01it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3725/10000 [04:03<05:30, 18.97it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3727/10000 [04:03<05:26, 19.24it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3729/10000 [04:03<05:23, 19.36it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3731/10000 [04:03<05:31, 18.92it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3733/10000 [04:03<05:27, 19.13it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3735/10000 [04:03<05:26, 19.20it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3737/10000 [04:04<05:47, 18.01it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3739/10000 [04:04<05:56, 17.58it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3741/10000 [04:04<06:08, 16.98it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3743/10000 [04:04<06:05, 17.14it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3745/10000 [04:04<05:52, 17.72it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3747/10000 [04:04<05:47, 18.01it/s]\n",
      "Loss: 1.337 :  37%|███▋      | 3749/10000 [04:04<05:41, 18.29it/s]\n",
      "Loss: 1.258 :  37%|███▋      | 3749/10000 [04:04<05:41, 18.29it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3751/10000 [04:04<06:14, 16.70it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3753/10000 [04:04<06:31, 15.94it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3755/10000 [04:05<06:35, 15.79it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3757/10000 [04:05<06:25, 16.19it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3759/10000 [04:05<06:12, 16.74it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3761/10000 [04:05<06:14, 16.65it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3763/10000 [04:05<06:08, 16.93it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3765/10000 [04:05<05:57, 17.43it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3767/10000 [04:05<05:46, 18.00it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3769/10000 [04:05<05:47, 17.92it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3771/10000 [04:06<05:44, 18.11it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3773/10000 [04:06<05:45, 18.02it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3775/10000 [04:06<05:45, 18.00it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3777/10000 [04:06<06:08, 16.90it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3779/10000 [04:06<06:18, 16.42it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3781/10000 [04:06<06:07, 16.90it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3783/10000 [04:06<06:03, 17.10it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3785/10000 [04:06<06:12, 16.69it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3787/10000 [04:06<06:19, 16.36it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3789/10000 [04:07<06:09, 16.82it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3791/10000 [04:07<05:58, 17.31it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3793/10000 [04:07<05:53, 17.56it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3795/10000 [04:07<05:50, 17.69it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3797/10000 [04:07<05:51, 17.63it/s]\n",
      "Loss: 1.258 :  38%|███▊      | 3799/10000 [04:07<05:43, 18.07it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3799/10000 [04:07<05:43, 18.07it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3801/10000 [04:07<06:15, 16.51it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3803/10000 [04:07<06:04, 16.98it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3805/10000 [04:08<05:59, 17.25it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3807/10000 [04:08<05:46, 17.89it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3809/10000 [04:08<05:38, 18.32it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3811/10000 [04:08<05:35, 18.46it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3813/10000 [04:08<05:31, 18.65it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3815/10000 [04:08<05:31, 18.64it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3817/10000 [04:08<05:27, 18.87it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3820/10000 [04:08<05:17, 19.47it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3823/10000 [04:08<05:15, 19.58it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3825/10000 [04:09<05:15, 19.56it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3827/10000 [04:09<05:18, 19.35it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3829/10000 [04:09<05:23, 19.06it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3831/10000 [04:09<05:25, 18.94it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3833/10000 [04:09<05:29, 18.71it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3835/10000 [04:09<05:26, 18.90it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3837/10000 [04:09<05:32, 18.55it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3839/10000 [04:09<05:45, 17.83it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3841/10000 [04:09<05:52, 17.49it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3843/10000 [04:10<05:55, 17.31it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3845/10000 [04:10<06:14, 16.43it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3847/10000 [04:10<06:59, 14.66it/s]\n",
      "Loss: 1.163 :  38%|███▊      | 3849/10000 [04:10<07:16, 14.08it/s]\n",
      "Loss: 1.126 :  38%|███▊      | 3849/10000 [04:10<07:16, 14.08it/s]\n",
      "Loss: 1.126 :  39%|███▊      | 3851/10000 [04:10<07:52, 13.02it/s]\n",
      "Loss: 1.126 :  39%|███▊      | 3853/10000 [04:10<08:00, 12.80it/s]\n",
      "Loss: 1.126 :  39%|███▊      | 3855/10000 [04:10<07:34, 13.52it/s]\n",
      "Loss: 1.126 :  39%|███▊      | 3857/10000 [04:11<07:08, 14.35it/s]\n",
      "Loss: 1.126 :  39%|███▊      | 3859/10000 [04:11<06:44, 15.16it/s]\n",
      "Loss: 1.126 :  39%|███▊      | 3861/10000 [04:11<06:28, 15.79it/s]\n",
      "Loss: 1.126 :  39%|███▊      | 3863/10000 [04:11<06:19, 16.18it/s]\n",
      "Loss: 1.126 :  39%|███▊      | 3865/10000 [04:11<06:11, 16.50it/s]\n",
      "Loss: 1.126 :  39%|███▊      | 3867/10000 [04:11<06:20, 16.10it/s]\n",
      "Loss: 1.126 :  39%|███▊      | 3869/10000 [04:11<06:18, 16.18it/s]\n",
      "Loss: 1.126 :  39%|███▊      | 3871/10000 [04:11<06:16, 16.29it/s]\n",
      "Loss: 1.126 :  39%|███▊      | 3873/10000 [04:12<06:03, 16.83it/s]\n",
      "Loss: 1.126 :  39%|███▉      | 3875/10000 [04:12<05:52, 17.39it/s]\n",
      "Loss: 1.126 :  39%|███▉      | 3877/10000 [04:12<05:50, 17.48it/s]\n",
      "Loss: 1.126 :  39%|███▉      | 3879/10000 [04:12<05:47, 17.62it/s]\n",
      "Loss: 1.126 :  39%|███▉      | 3881/10000 [04:12<05:48, 17.55it/s]\n",
      "Loss: 1.126 :  39%|███▉      | 3883/10000 [04:12<06:03, 16.81it/s]\n",
      "Loss: 1.126 :  39%|███▉      | 3885/10000 [04:12<06:02, 16.87it/s]\n",
      "Loss: 1.126 :  39%|███▉      | 3887/10000 [04:12<06:00, 16.98it/s]\n",
      "Loss: 1.126 :  39%|███▉      | 3889/10000 [04:12<05:53, 17.27it/s]\n",
      "Loss: 1.126 :  39%|███▉      | 3891/10000 [04:13<05:45, 17.70it/s]\n",
      "Loss: 1.126 :  39%|███▉      | 3893/10000 [04:13<05:33, 18.28it/s]\n",
      "Loss: 1.126 :  39%|███▉      | 3895/10000 [04:13<05:29, 18.54it/s]\n",
      "Loss: 1.126 :  39%|███▉      | 3897/10000 [04:13<05:30, 18.49it/s]\n",
      "Loss: 1.126 :  39%|███▉      | 3899/10000 [04:13<05:34, 18.26it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3899/10000 [04:13<05:34, 18.26it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3901/10000 [04:13<06:08, 16.55it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3903/10000 [04:13<05:54, 17.20it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3905/10000 [04:13<05:46, 17.61it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3907/10000 [04:13<05:47, 17.52it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3909/10000 [04:14<05:49, 17.45it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3911/10000 [04:14<05:51, 17.34it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3913/10000 [04:14<05:50, 17.35it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3915/10000 [04:14<05:45, 17.60it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3917/10000 [04:14<05:38, 17.95it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3919/10000 [04:14<05:29, 18.46it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3921/10000 [04:14<05:28, 18.50it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3923/10000 [04:14<05:25, 18.67it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3925/10000 [04:14<05:22, 18.86it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3927/10000 [04:15<05:16, 19.17it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3929/10000 [04:15<05:17, 19.14it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3931/10000 [04:15<05:21, 18.90it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3933/10000 [04:15<05:22, 18.80it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3935/10000 [04:15<05:33, 18.16it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3937/10000 [04:15<06:03, 16.68it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3939/10000 [04:15<06:13, 16.21it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3941/10000 [04:15<06:04, 16.64it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3943/10000 [04:15<05:50, 17.29it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3945/10000 [04:16<05:44, 17.59it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3947/10000 [04:16<05:40, 17.80it/s]\n",
      "Loss: 1.162 :  39%|███▉      | 3949/10000 [04:16<05:38, 17.87it/s]\n",
      "Loss: 1.289 :  39%|███▉      | 3949/10000 [04:16<05:38, 17.87it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3951/10000 [04:16<06:15, 16.11it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3953/10000 [04:16<05:57, 16.92it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3955/10000 [04:16<05:46, 17.44it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3957/10000 [04:16<05:38, 17.86it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3959/10000 [04:16<05:32, 18.19it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3961/10000 [04:16<05:27, 18.43it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3963/10000 [04:17<05:20, 18.84it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3965/10000 [04:17<05:19, 18.87it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3967/10000 [04:17<05:27, 18.44it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3969/10000 [04:17<05:28, 18.39it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3971/10000 [04:17<05:24, 18.59it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3974/10000 [04:17<05:17, 18.97it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3976/10000 [04:17<05:21, 18.72it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3978/10000 [04:17<05:19, 18.87it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3980/10000 [04:18<05:23, 18.62it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3982/10000 [04:18<05:20, 18.77it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3984/10000 [04:18<05:18, 18.89it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3986/10000 [04:18<05:21, 18.71it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3988/10000 [04:18<06:00, 16.68it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3991/10000 [04:18<05:33, 18.00it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3993/10000 [04:18<05:29, 18.23it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3995/10000 [04:18<05:24, 18.53it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3997/10000 [04:18<05:31, 18.09it/s]\n",
      "Loss: 1.289 :  40%|███▉      | 3999/10000 [04:19<05:32, 18.06it/s]\n",
      "Loss: 1.246 :  40%|███▉      | 3999/10000 [04:19<05:32, 18.06it/s]\n",
      "Loss: 1.246 :  40%|████      | 4001/10000 [04:19<05:57, 16.76it/s]\n",
      "Loss: 1.246 :  40%|████      | 4003/10000 [04:19<05:54, 16.92it/s]\n",
      "Loss: 1.246 :  40%|████      | 4005/10000 [04:19<05:47, 17.27it/s]\n",
      "Loss: 1.246 :  40%|████      | 4007/10000 [04:19<05:33, 17.96it/s]\n",
      "Loss: 1.246 :  40%|████      | 4009/10000 [04:19<05:25, 18.42it/s]\n",
      "Loss: 1.246 :  40%|████      | 4011/10000 [04:19<05:25, 18.42it/s]\n",
      "Loss: 1.246 :  40%|████      | 4013/10000 [04:19<05:47, 17.23it/s]\n",
      "Loss: 1.246 :  40%|████      | 4015/10000 [04:20<06:05, 16.38it/s]\n",
      "Loss: 1.246 :  40%|████      | 4017/10000 [04:20<06:15, 15.92it/s]\n",
      "Loss: 1.246 :  40%|████      | 4019/10000 [04:20<06:18, 15.78it/s]\n",
      "Loss: 1.246 :  40%|████      | 4021/10000 [04:20<06:34, 15.14it/s]\n",
      "Loss: 1.246 :  40%|████      | 4023/10000 [04:20<06:38, 15.01it/s]\n",
      "Loss: 1.246 :  40%|████      | 4025/10000 [04:20<06:44, 14.78it/s]\n",
      "Loss: 1.246 :  40%|████      | 4027/10000 [04:20<07:13, 13.76it/s]\n",
      "Loss: 1.246 :  40%|████      | 4029/10000 [04:21<07:55, 12.54it/s]\n",
      "Loss: 1.246 :  40%|████      | 4031/10000 [04:21<08:02, 12.38it/s]\n",
      "Loss: 1.246 :  40%|████      | 4033/10000 [04:21<07:55, 12.55it/s]\n",
      "Loss: 1.246 :  40%|████      | 4035/10000 [04:21<07:46, 12.80it/s]\n",
      "Loss: 1.246 :  40%|████      | 4037/10000 [04:21<07:35, 13.10it/s]\n",
      "Loss: 1.246 :  40%|████      | 4039/10000 [04:21<07:27, 13.32it/s]\n",
      "Loss: 1.246 :  40%|████      | 4041/10000 [04:21<07:11, 13.82it/s]\n",
      "Loss: 1.246 :  40%|████      | 4043/10000 [04:22<06:51, 14.48it/s]\n",
      "Loss: 1.246 :  40%|████      | 4045/10000 [04:22<06:39, 14.92it/s]\n",
      "Loss: 1.246 :  40%|████      | 4047/10000 [04:22<06:29, 15.29it/s]\n",
      "Loss: 1.246 :  40%|████      | 4049/10000 [04:22<06:22, 15.55it/s]\n",
      "Loss: 1.057 :  40%|████      | 4049/10000 [04:22<06:22, 15.55it/s]\n",
      "Loss: 1.057 :  41%|████      | 4051/10000 [04:22<06:50, 14.47it/s]\n",
      "Loss: 1.057 :  41%|████      | 4053/10000 [04:22<06:29, 15.28it/s]\n",
      "Loss: 1.057 :  41%|████      | 4055/10000 [04:22<06:15, 15.85it/s]\n",
      "Loss: 1.057 :  41%|████      | 4057/10000 [04:22<06:06, 16.20it/s]\n",
      "Loss: 1.057 :  41%|████      | 4059/10000 [04:23<05:59, 16.51it/s]\n",
      "Loss: 1.057 :  41%|████      | 4061/10000 [04:23<05:59, 16.54it/s]\n",
      "Loss: 1.057 :  41%|████      | 4063/10000 [04:23<05:51, 16.90it/s]\n",
      "Loss: 1.057 :  41%|████      | 4065/10000 [04:23<05:48, 17.03it/s]\n",
      "Loss: 1.057 :  41%|████      | 4067/10000 [04:23<05:45, 17.19it/s]\n",
      "Loss: 1.057 :  41%|████      | 4069/10000 [04:23<05:41, 17.39it/s]\n",
      "Loss: 1.057 :  41%|████      | 4071/10000 [04:23<05:48, 17.03it/s]\n",
      "Loss: 1.057 :  41%|████      | 4073/10000 [04:23<05:47, 17.04it/s]\n",
      "Loss: 1.057 :  41%|████      | 4075/10000 [04:23<05:37, 17.54it/s]\n",
      "Loss: 1.057 :  41%|████      | 4077/10000 [04:24<05:44, 17.22it/s]\n",
      "Loss: 1.057 :  41%|████      | 4079/10000 [04:24<05:38, 17.51it/s]\n",
      "Loss: 1.057 :  41%|████      | 4081/10000 [04:24<05:44, 17.17it/s]\n",
      "Loss: 1.057 :  41%|████      | 4083/10000 [04:24<05:48, 16.96it/s]\n",
      "Loss: 1.057 :  41%|████      | 4085/10000 [04:24<05:48, 16.97it/s]\n",
      "Loss: 1.057 :  41%|████      | 4087/10000 [04:24<05:43, 17.21it/s]\n",
      "Loss: 1.057 :  41%|████      | 4089/10000 [04:24<05:40, 17.34it/s]\n",
      "Loss: 1.057 :  41%|████      | 4091/10000 [04:24<05:40, 17.37it/s]\n",
      "Loss: 1.057 :  41%|████      | 4093/10000 [04:25<05:38, 17.47it/s]\n",
      "Loss: 1.057 :  41%|████      | 4095/10000 [04:25<05:31, 17.83it/s]\n",
      "Loss: 1.057 :  41%|████      | 4097/10000 [04:25<05:30, 17.84it/s]\n",
      "Loss: 1.057 :  41%|████      | 4099/10000 [04:25<05:29, 17.88it/s]\n",
      "Loss: 1.329 :  41%|████      | 4099/10000 [04:25<05:29, 17.88it/s]\n",
      "Loss: 1.329 :  41%|████      | 4101/10000 [04:25<06:06, 16.11it/s]\n",
      "Loss: 1.329 :  41%|████      | 4103/10000 [04:25<05:56, 16.52it/s]\n",
      "Loss: 1.329 :  41%|████      | 4105/10000 [04:25<05:51, 16.79it/s]\n",
      "Loss: 1.329 :  41%|████      | 4107/10000 [04:25<05:53, 16.65it/s]\n",
      "Loss: 1.329 :  41%|████      | 4109/10000 [04:25<05:48, 16.90it/s]\n",
      "Loss: 1.329 :  41%|████      | 4111/10000 [04:26<05:45, 17.05it/s]\n",
      "Loss: 1.329 :  41%|████      | 4113/10000 [04:26<05:59, 16.37it/s]\n",
      "Loss: 1.329 :  41%|████      | 4115/10000 [04:26<06:29, 15.11it/s]\n",
      "Loss: 1.329 :  41%|████      | 4117/10000 [04:26<06:39, 14.73it/s]\n",
      "Loss: 1.329 :  41%|████      | 4119/10000 [04:26<06:29, 15.08it/s]\n",
      "Loss: 1.329 :  41%|████      | 4121/10000 [04:26<06:21, 15.41it/s]\n",
      "Loss: 1.329 :  41%|████      | 4123/10000 [04:26<06:22, 15.38it/s]\n",
      "Loss: 1.329 :  41%|████▏     | 4125/10000 [04:27<06:18, 15.51it/s]\n",
      "Loss: 1.329 :  41%|████▏     | 4127/10000 [04:27<06:15, 15.63it/s]\n",
      "Loss: 1.329 :  41%|████▏     | 4129/10000 [04:27<06:09, 15.88it/s]\n",
      "Loss: 1.329 :  41%|████▏     | 4131/10000 [04:27<06:02, 16.18it/s]\n",
      "Loss: 1.329 :  41%|████▏     | 4133/10000 [04:27<06:02, 16.20it/s]\n",
      "Loss: 1.329 :  41%|████▏     | 4135/10000 [04:27<05:59, 16.31it/s]\n",
      "Loss: 1.329 :  41%|████▏     | 4137/10000 [04:27<05:53, 16.56it/s]\n",
      "Loss: 1.329 :  41%|████▏     | 4139/10000 [04:27<05:50, 16.72it/s]\n",
      "Loss: 1.329 :  41%|████▏     | 4141/10000 [04:27<05:43, 17.04it/s]\n",
      "Loss: 1.329 :  41%|████▏     | 4143/10000 [04:28<05:45, 16.96it/s]\n",
      "Loss: 1.329 :  41%|████▏     | 4145/10000 [04:28<05:41, 17.12it/s]\n",
      "Loss: 1.329 :  41%|████▏     | 4147/10000 [04:28<05:39, 17.23it/s]\n",
      "Loss: 1.329 :  41%|████▏     | 4149/10000 [04:28<05:47, 16.84it/s]\n",
      "Loss: 1.309 :  41%|████▏     | 4149/10000 [04:28<05:47, 16.84it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4151/10000 [04:28<06:26, 15.14it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4153/10000 [04:28<06:03, 16.06it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4155/10000 [04:28<05:53, 16.52it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4157/10000 [04:28<05:44, 16.94it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4159/10000 [04:29<05:46, 16.85it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4161/10000 [04:29<05:36, 17.34it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4163/10000 [04:29<05:33, 17.49it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4165/10000 [04:29<05:35, 17.41it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4167/10000 [04:29<05:37, 17.27it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4169/10000 [04:29<05:35, 17.38it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4171/10000 [04:29<05:33, 17.50it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4173/10000 [04:29<05:27, 17.77it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4175/10000 [04:29<05:33, 17.48it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4177/10000 [04:30<05:29, 17.66it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4179/10000 [04:30<05:33, 17.46it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4181/10000 [04:30<05:29, 17.68it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4183/10000 [04:30<05:24, 17.91it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4185/10000 [04:30<05:27, 17.77it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4187/10000 [04:30<05:29, 17.65it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4189/10000 [04:30<05:28, 17.69it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4191/10000 [04:30<05:27, 17.76it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4193/10000 [04:30<05:29, 17.62it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4195/10000 [04:31<05:30, 17.58it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4197/10000 [04:31<05:29, 17.61it/s]\n",
      "Loss: 1.309 :  42%|████▏     | 4199/10000 [04:31<05:26, 17.78it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4199/10000 [04:31<05:26, 17.78it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4201/10000 [04:31<06:05, 15.87it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4203/10000 [04:31<06:36, 14.61it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4205/10000 [04:31<06:44, 14.33it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4207/10000 [04:31<06:41, 14.42it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4209/10000 [04:32<06:25, 15.02it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4211/10000 [04:32<06:18, 15.28it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4213/10000 [04:32<06:12, 15.52it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4215/10000 [04:32<06:07, 15.73it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4217/10000 [04:32<06:03, 15.91it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4219/10000 [04:32<06:04, 15.85it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4221/10000 [04:32<05:53, 16.35it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4223/10000 [04:32<05:46, 16.68it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4225/10000 [04:33<05:40, 16.95it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4227/10000 [04:33<05:36, 17.17it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4229/10000 [04:33<05:37, 17.09it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4231/10000 [04:33<05:34, 17.27it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4233/10000 [04:33<05:32, 17.34it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4235/10000 [04:33<05:29, 17.51it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4237/10000 [04:33<05:36, 17.12it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4239/10000 [04:33<05:34, 17.24it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4241/10000 [04:33<05:27, 17.58it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4243/10000 [04:34<05:30, 17.44it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4245/10000 [04:34<05:31, 17.36it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4247/10000 [04:34<05:29, 17.45it/s]\n",
      "Loss: 1.053 :  42%|████▏     | 4249/10000 [04:34<05:33, 17.24it/s]\n",
      "Loss: 1.291 :  42%|████▏     | 4249/10000 [04:34<05:33, 17.24it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4251/10000 [04:34<06:06, 15.70it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4253/10000 [04:34<05:59, 15.98it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4255/10000 [04:34<05:43, 16.75it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4257/10000 [04:34<05:34, 17.19it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4259/10000 [04:35<05:37, 17.01it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4261/10000 [04:35<05:36, 17.07it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4263/10000 [04:35<05:36, 17.03it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4265/10000 [04:35<05:30, 17.33it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4267/10000 [04:35<05:29, 17.37it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4269/10000 [04:35<05:35, 17.07it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4271/10000 [04:35<05:37, 16.96it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4273/10000 [04:35<05:36, 17.01it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4275/10000 [04:35<05:33, 17.15it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4277/10000 [04:36<05:33, 17.17it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4279/10000 [04:36<05:36, 16.98it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4281/10000 [04:36<05:37, 16.97it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4283/10000 [04:36<05:35, 17.05it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4285/10000 [04:36<05:33, 17.13it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4287/10000 [04:36<05:36, 16.99it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4289/10000 [04:36<05:33, 17.14it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4291/10000 [04:36<05:52, 16.18it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4293/10000 [04:37<06:22, 14.92it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4295/10000 [04:37<06:29, 14.64it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4297/10000 [04:37<06:18, 15.05it/s]\n",
      "Loss: 1.291 :  43%|████▎     | 4299/10000 [04:37<06:13, 15.28it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4299/10000 [04:37<06:13, 15.28it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4301/10000 [04:37<06:44, 14.09it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4303/10000 [04:37<06:21, 14.93it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4305/10000 [04:37<06:02, 15.70it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4307/10000 [04:37<05:54, 16.06it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4309/10000 [04:38<05:50, 16.22it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4311/10000 [04:38<05:47, 16.37it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4313/10000 [04:38<05:47, 16.37it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4315/10000 [04:38<05:42, 16.61it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4317/10000 [04:38<05:39, 16.74it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4319/10000 [04:38<05:31, 17.15it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4321/10000 [04:38<05:31, 17.12it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4323/10000 [04:38<05:28, 17.30it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4325/10000 [04:39<05:27, 17.31it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4327/10000 [04:39<05:24, 17.47it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4329/10000 [04:39<05:30, 17.15it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4331/10000 [04:39<05:37, 16.79it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4333/10000 [04:39<05:37, 16.78it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4335/10000 [04:39<05:37, 16.78it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4337/10000 [04:39<05:34, 16.91it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4339/10000 [04:39<05:32, 17.01it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4341/10000 [04:39<05:34, 16.93it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4343/10000 [04:40<05:28, 17.24it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4345/10000 [04:40<05:22, 17.54it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4347/10000 [04:40<05:26, 17.31it/s]\n",
      "Loss: 1.218 :  43%|████▎     | 4349/10000 [04:40<05:24, 17.43it/s]\n",
      "Loss: 1.063 :  43%|████▎     | 4349/10000 [04:40<05:24, 17.43it/s]\n",
      "Loss: 1.063 :  44%|████▎     | 4351/10000 [04:40<06:09, 15.28it/s]\n",
      "Loss: 1.063 :  44%|████▎     | 4353/10000 [04:40<05:55, 15.88it/s]\n",
      "Loss: 1.063 :  44%|████▎     | 4355/10000 [04:40<05:42, 16.47it/s]\n",
      "Loss: 1.063 :  44%|████▎     | 4357/10000 [04:40<05:34, 16.85it/s]\n",
      "Loss: 1.063 :  44%|████▎     | 4359/10000 [04:41<05:29, 17.11it/s]\n",
      "Loss: 1.063 :  44%|████▎     | 4361/10000 [04:41<05:28, 17.19it/s]\n",
      "Loss: 1.063 :  44%|████▎     | 4363/10000 [04:41<05:26, 17.25it/s]\n",
      "Loss: 1.063 :  44%|████▎     | 4365/10000 [04:41<05:25, 17.31it/s]\n",
      "Loss: 1.063 :  44%|████▎     | 4367/10000 [04:41<05:24, 17.37it/s]\n",
      "Loss: 1.063 :  44%|████▎     | 4369/10000 [04:41<05:21, 17.51it/s]\n",
      "Loss: 1.063 :  44%|████▎     | 4371/10000 [04:41<05:23, 17.38it/s]\n",
      "Loss: 1.063 :  44%|████▎     | 4373/10000 [04:41<05:19, 17.63it/s]\n",
      "Loss: 1.063 :  44%|████▍     | 4375/10000 [04:41<05:22, 17.47it/s]\n",
      "Loss: 1.063 :  44%|████▍     | 4377/10000 [04:42<05:22, 17.44it/s]\n",
      "Loss: 1.063 :  44%|████▍     | 4379/10000 [04:42<05:34, 16.82it/s]\n",
      "Loss: 1.063 :  44%|████▍     | 4381/10000 [04:42<06:07, 15.29it/s]\n",
      "Loss: 1.063 :  44%|████▍     | 4383/10000 [04:42<06:17, 14.90it/s]\n",
      "Loss: 1.063 :  44%|████▍     | 4385/10000 [04:42<06:07, 15.27it/s]\n",
      "Loss: 1.063 :  44%|████▍     | 4387/10000 [04:42<06:03, 15.46it/s]\n",
      "Loss: 1.063 :  44%|████▍     | 4389/10000 [04:42<06:06, 15.32it/s]\n",
      "Loss: 1.063 :  44%|████▍     | 4391/10000 [04:42<05:55, 15.78it/s]\n",
      "Loss: 1.063 :  44%|████▍     | 4393/10000 [04:43<05:55, 15.75it/s]\n",
      "Loss: 1.063 :  44%|████▍     | 4395/10000 [04:43<05:49, 16.06it/s]\n",
      "Loss: 1.063 :  44%|████▍     | 4397/10000 [04:43<05:46, 16.16it/s]\n",
      "Loss: 1.063 :  44%|████▍     | 4399/10000 [04:43<05:44, 16.26it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4399/10000 [04:43<05:44, 16.26it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4401/10000 [04:43<06:15, 14.92it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4403/10000 [04:43<06:01, 15.50it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4405/10000 [04:43<05:52, 15.87it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4407/10000 [04:43<05:42, 16.35it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4409/10000 [04:44<05:33, 16.74it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4411/10000 [04:44<05:30, 16.91it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4413/10000 [04:44<05:28, 17.02it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4415/10000 [04:44<05:28, 17.00it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4417/10000 [04:44<05:23, 17.27it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4419/10000 [04:44<05:24, 17.18it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4421/10000 [04:44<05:19, 17.46it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4423/10000 [04:44<05:18, 17.51it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4425/10000 [04:45<05:13, 17.78it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4427/10000 [04:45<05:12, 17.81it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4429/10000 [04:45<05:17, 17.56it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4431/10000 [04:45<05:16, 17.59it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4433/10000 [04:45<05:18, 17.47it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4435/10000 [04:45<05:19, 17.44it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4437/10000 [04:45<05:18, 17.48it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4439/10000 [04:45<05:15, 17.61it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4441/10000 [04:45<05:15, 17.63it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4443/10000 [04:46<05:14, 17.69it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4445/10000 [04:46<05:13, 17.71it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4447/10000 [04:46<05:14, 17.66it/s]\n",
      "Loss: 1.283 :  44%|████▍     | 4449/10000 [04:46<05:15, 17.59it/s]\n",
      "Loss: 0.910 :  44%|████▍     | 4449/10000 [04:46<05:15, 17.59it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4451/10000 [04:46<05:48, 15.91it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4453/10000 [04:46<05:41, 16.26it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4455/10000 [04:46<05:35, 16.54it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4457/10000 [04:46<05:30, 16.79it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4459/10000 [04:46<05:28, 16.87it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4461/10000 [04:47<05:32, 16.66it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4463/10000 [04:47<05:25, 16.98it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4465/10000 [04:47<05:30, 16.73it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4467/10000 [04:47<05:33, 16.60it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4469/10000 [04:47<06:08, 15.00it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4471/10000 [04:47<06:22, 14.45it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4473/10000 [04:47<06:10, 14.92it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4475/10000 [04:48<06:02, 15.22it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4477/10000 [04:48<05:55, 15.53it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4479/10000 [04:48<05:55, 15.51it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4481/10000 [04:48<05:48, 15.85it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4483/10000 [04:48<05:48, 15.82it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4485/10000 [04:48<05:39, 16.27it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4487/10000 [04:48<05:37, 16.36it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4489/10000 [04:48<05:33, 16.51it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4491/10000 [04:49<05:33, 16.51it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4493/10000 [04:49<05:26, 16.85it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4495/10000 [04:49<05:22, 17.09it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4497/10000 [04:49<05:23, 17.01it/s]\n",
      "Loss: 0.910 :  45%|████▍     | 4499/10000 [04:49<05:20, 17.15it/s]\n",
      "Loss: 1.209 :  45%|████▍     | 4499/10000 [04:49<05:20, 17.15it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4501/10000 [04:49<05:59, 15.31it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4503/10000 [04:49<05:50, 15.68it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4505/10000 [04:49<05:41, 16.10it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4507/10000 [04:49<05:29, 16.67it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4509/10000 [04:50<05:26, 16.83it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4511/10000 [04:50<05:19, 17.16it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4513/10000 [04:50<05:19, 17.20it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4515/10000 [04:50<05:19, 17.18it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4517/10000 [04:50<05:18, 17.19it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4519/10000 [04:50<05:16, 17.33it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4521/10000 [04:50<05:14, 17.40it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4523/10000 [04:50<05:15, 17.37it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4525/10000 [04:51<05:11, 17.58it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4527/10000 [04:51<05:14, 17.42it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4529/10000 [04:51<05:12, 17.54it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4531/10000 [04:51<05:12, 17.51it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4533/10000 [04:51<05:13, 17.44it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4535/10000 [04:51<05:17, 17.20it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4537/10000 [04:51<05:18, 17.16it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4539/10000 [04:51<05:20, 17.02it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4541/10000 [04:51<05:23, 16.86it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4543/10000 [04:52<05:17, 17.19it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4545/10000 [04:52<05:16, 17.21it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4547/10000 [04:52<05:19, 17.07it/s]\n",
      "Loss: 1.209 :  45%|████▌     | 4549/10000 [04:52<05:21, 16.96it/s]\n",
      "Loss: 1.187 :  45%|████▌     | 4549/10000 [04:52<05:21, 16.96it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4551/10000 [04:52<05:47, 15.66it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4553/10000 [04:52<05:39, 16.06it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4555/10000 [04:52<05:26, 16.70it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4557/10000 [04:52<05:59, 15.13it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4559/10000 [04:53<06:12, 14.60it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4561/10000 [04:53<06:10, 14.66it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4563/10000 [04:53<06:04, 14.90it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4565/10000 [04:53<05:56, 15.23it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4567/10000 [04:53<05:46, 15.68it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4569/10000 [04:53<05:41, 15.88it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4571/10000 [04:53<05:34, 16.23it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4573/10000 [04:53<05:26, 16.60it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4575/10000 [04:54<05:23, 16.79it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4577/10000 [04:54<05:25, 16.66it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4579/10000 [04:54<05:30, 16.40it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4581/10000 [04:54<05:22, 16.78it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4583/10000 [04:54<05:22, 16.80it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4585/10000 [04:54<05:16, 17.11it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4587/10000 [04:54<05:15, 17.17it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4589/10000 [04:54<05:13, 17.25it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4591/10000 [04:55<05:15, 17.12it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4593/10000 [04:55<05:10, 17.42it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4595/10000 [04:55<05:07, 17.56it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4597/10000 [04:55<05:11, 17.37it/s]\n",
      "Loss: 1.187 :  46%|████▌     | 4599/10000 [04:55<05:12, 17.29it/s]\n",
      "Loss: 1.051 :  46%|████▌     | 4599/10000 [04:55<05:12, 17.29it/s]\n",
      "Loss: 1.051 :  46%|████▌     | 4601/10000 [04:55<05:50, 15.42it/s]\n",
      "Loss: 1.051 :  46%|████▌     | 4603/10000 [04:55<05:38, 15.94it/s]\n",
      "Loss: 1.051 :  46%|████▌     | 4605/10000 [04:55<05:28, 16.43it/s]\n",
      "Loss: 1.051 :  46%|████▌     | 4607/10000 [04:55<05:21, 16.76it/s]\n",
      "Loss: 1.051 :  46%|████▌     | 4609/10000 [04:56<05:22, 16.72it/s]\n",
      "Loss: 1.051 :  46%|████▌     | 4611/10000 [04:56<05:16, 17.01it/s]\n",
      "Loss: 1.051 :  46%|████▌     | 4613/10000 [04:56<05:12, 17.24it/s]\n",
      "Loss: 1.051 :  46%|████▌     | 4615/10000 [04:56<05:16, 17.02it/s]\n",
      "Loss: 1.051 :  46%|████▌     | 4617/10000 [04:56<05:37, 15.96it/s]\n",
      "Loss: 1.051 :  46%|████▌     | 4619/10000 [04:56<05:53, 15.23it/s]\n",
      "Loss: 1.051 :  46%|████▌     | 4621/10000 [04:56<05:49, 15.39it/s]\n",
      "Loss: 1.051 :  46%|████▌     | 4623/10000 [04:56<05:44, 15.62it/s]\n",
      "Loss: 1.051 :  46%|████▋     | 4625/10000 [04:57<05:37, 15.93it/s]\n",
      "Loss: 1.051 :  46%|████▋     | 4627/10000 [04:57<05:33, 16.11it/s]\n",
      "Loss: 1.051 :  46%|████▋     | 4629/10000 [04:57<06:25, 13.93it/s]\n",
      "Loss: 1.051 :  46%|████▋     | 4631/10000 [04:57<06:41, 13.38it/s]\n",
      "Loss: 1.051 :  46%|████▋     | 4633/10000 [04:57<06:50, 13.07it/s]\n",
      "Loss: 1.051 :  46%|████▋     | 4635/10000 [04:57<06:48, 13.13it/s]\n",
      "Loss: 1.051 :  46%|████▋     | 4637/10000 [04:58<06:25, 13.90it/s]\n",
      "Loss: 1.051 :  46%|████▋     | 4639/10000 [04:58<06:10, 14.46it/s]\n",
      "Loss: 1.051 :  46%|████▋     | 4641/10000 [04:58<06:36, 13.52it/s]\n",
      "Loss: 1.051 :  46%|████▋     | 4643/10000 [04:58<06:44, 13.23it/s]\n",
      "Loss: 1.051 :  46%|████▋     | 4645/10000 [04:58<06:31, 13.68it/s]\n",
      "Loss: 1.051 :  46%|████▋     | 4647/10000 [04:58<06:37, 13.46it/s]\n",
      "Loss: 1.051 :  46%|████▋     | 4649/10000 [04:58<07:14, 12.31it/s]\n",
      "Loss: 1.075 :  46%|████▋     | 4649/10000 [04:59<07:14, 12.31it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4651/10000 [04:59<08:27, 10.54it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4653/10000 [04:59<07:48, 11.41it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4655/10000 [04:59<07:17, 12.22it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4657/10000 [04:59<06:45, 13.17it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4659/10000 [04:59<06:31, 13.64it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4661/10000 [04:59<06:17, 14.13it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4663/10000 [05:00<06:21, 14.00it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4665/10000 [05:00<06:26, 13.80it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4667/10000 [05:00<06:31, 13.63it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4669/10000 [05:00<06:31, 13.61it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4671/10000 [05:00<06:34, 13.51it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4673/10000 [05:00<06:40, 13.30it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4675/10000 [05:00<06:26, 13.79it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4677/10000 [05:01<06:12, 14.29it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4679/10000 [05:01<06:01, 14.73it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4681/10000 [05:01<05:53, 15.04it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4683/10000 [05:01<05:45, 15.41it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4685/10000 [05:01<05:39, 15.67it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4687/10000 [05:01<05:28, 16.15it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4689/10000 [05:01<05:30, 16.09it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4691/10000 [05:01<05:35, 15.83it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4693/10000 [05:02<05:28, 16.14it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4695/10000 [05:02<05:22, 16.45it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4697/10000 [05:02<05:13, 16.92it/s]\n",
      "Loss: 1.075 :  47%|████▋     | 4699/10000 [05:02<05:09, 17.14it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4699/10000 [05:02<05:09, 17.14it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4701/10000 [05:02<05:45, 15.36it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4703/10000 [05:02<05:35, 15.77it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4705/10000 [05:02<05:25, 16.28it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4707/10000 [05:02<05:23, 16.35it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4709/10000 [05:02<05:19, 16.57it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4711/10000 [05:03<05:11, 16.98it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4713/10000 [05:03<05:09, 17.06it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4715/10000 [05:03<05:08, 17.15it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4717/10000 [05:03<05:05, 17.27it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4719/10000 [05:03<05:36, 15.70it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4721/10000 [05:03<05:50, 15.07it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4723/10000 [05:03<05:50, 15.04it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4725/10000 [05:04<05:46, 15.22it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4727/10000 [05:04<05:41, 15.44it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4729/10000 [05:04<05:42, 15.40it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4731/10000 [05:04<05:41, 15.42it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4733/10000 [05:04<05:40, 15.49it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4735/10000 [05:04<05:31, 15.89it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4737/10000 [05:04<05:30, 15.95it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4739/10000 [05:04<05:27, 16.06it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4741/10000 [05:05<05:23, 16.24it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4743/10000 [05:05<05:18, 16.48it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4745/10000 [05:05<05:12, 16.79it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4747/10000 [05:05<05:08, 17.01it/s]\n",
      "Loss: 1.055 :  47%|████▋     | 4749/10000 [05:05<05:08, 17.04it/s]\n",
      "Loss: 1.010 :  47%|████▋     | 4749/10000 [05:05<05:08, 17.04it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4751/10000 [05:05<05:37, 15.55it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4753/10000 [05:05<05:32, 15.80it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4755/10000 [05:05<05:29, 15.94it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4757/10000 [05:05<05:24, 16.16it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4759/10000 [05:06<05:18, 16.43it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4761/10000 [05:06<05:20, 16.33it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4763/10000 [05:06<05:16, 16.53it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4765/10000 [05:06<05:09, 16.89it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4767/10000 [05:06<05:13, 16.67it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4769/10000 [05:06<05:10, 16.83it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4771/10000 [05:06<05:09, 16.89it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4773/10000 [05:06<05:08, 16.96it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4775/10000 [05:07<05:08, 16.93it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4777/10000 [05:07<05:03, 17.19it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4779/10000 [05:07<05:06, 17.06it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4781/10000 [05:07<05:08, 16.90it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4783/10000 [05:07<05:12, 16.68it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4785/10000 [05:07<05:12, 16.68it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4787/10000 [05:07<05:06, 16.99it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4789/10000 [05:07<05:09, 16.85it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4791/10000 [05:08<05:08, 16.86it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4793/10000 [05:08<05:15, 16.52it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4795/10000 [05:08<05:19, 16.32it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4797/10000 [05:08<05:15, 16.49it/s]\n",
      "Loss: 1.010 :  48%|████▊     | 4799/10000 [05:08<05:10, 16.77it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4799/10000 [05:08<05:10, 16.77it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4801/10000 [05:08<05:35, 15.52it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4803/10000 [05:08<05:24, 16.01it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4805/10000 [05:08<05:40, 15.24it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4807/10000 [05:09<05:55, 14.62it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4809/10000 [05:09<05:53, 14.69it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4811/10000 [05:09<05:51, 14.77it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4813/10000 [05:09<05:42, 15.16it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4815/10000 [05:09<05:32, 15.62it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4817/10000 [05:09<05:32, 15.60it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4819/10000 [05:09<05:23, 15.99it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4821/10000 [05:09<05:20, 16.14it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4823/10000 [05:10<05:23, 16.00it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4825/10000 [05:10<05:22, 16.03it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4827/10000 [05:10<05:20, 16.14it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4829/10000 [05:10<05:14, 16.45it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4831/10000 [05:10<05:13, 16.51it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4833/10000 [05:10<05:07, 16.79it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4835/10000 [05:10<05:02, 17.06it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4837/10000 [05:10<05:04, 16.95it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4839/10000 [05:11<05:02, 17.09it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4841/10000 [05:11<04:57, 17.36it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4843/10000 [05:11<04:54, 17.50it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4845/10000 [05:11<05:02, 17.06it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4847/10000 [05:11<05:01, 17.07it/s]\n",
      "Loss: 1.046 :  48%|████▊     | 4849/10000 [05:11<05:00, 17.17it/s]\n",
      "Loss: 1.024 :  48%|████▊     | 4849/10000 [05:11<05:00, 17.17it/s]\n",
      "Loss: 1.024 :  49%|████▊     | 4851/10000 [05:11<05:27, 15.74it/s]\n",
      "Loss: 1.024 :  49%|████▊     | 4853/10000 [05:11<05:17, 16.20it/s]\n",
      "Loss: 1.024 :  49%|████▊     | 4855/10000 [05:11<05:16, 16.26it/s]\n",
      "Loss: 1.024 :  49%|████▊     | 4857/10000 [05:12<05:10, 16.58it/s]\n",
      "Loss: 1.024 :  49%|████▊     | 4859/10000 [05:12<05:05, 16.80it/s]\n",
      "Loss: 1.024 :  49%|████▊     | 4861/10000 [05:12<05:10, 16.53it/s]\n",
      "Loss: 1.024 :  49%|████▊     | 4863/10000 [05:12<05:06, 16.75it/s]\n",
      "Loss: 1.024 :  49%|████▊     | 4865/10000 [05:12<05:05, 16.78it/s]\n",
      "Loss: 1.024 :  49%|████▊     | 4867/10000 [05:12<04:57, 17.24it/s]\n",
      "Loss: 1.024 :  49%|████▊     | 4869/10000 [05:12<04:53, 17.51it/s]\n",
      "Loss: 1.024 :  49%|████▊     | 4871/10000 [05:12<04:56, 17.31it/s]\n",
      "Loss: 1.024 :  49%|████▊     | 4873/10000 [05:13<04:55, 17.35it/s]\n",
      "Loss: 1.024 :  49%|████▉     | 4875/10000 [05:13<04:58, 17.19it/s]\n",
      "Loss: 1.024 :  49%|████▉     | 4877/10000 [05:13<05:04, 16.84it/s]\n",
      "Loss: 1.024 :  49%|████▉     | 4879/10000 [05:13<04:57, 17.20it/s]\n",
      "Loss: 1.024 :  49%|████▉     | 4881/10000 [05:13<04:53, 17.43it/s]\n",
      "Loss: 1.024 :  49%|████▉     | 4883/10000 [05:13<04:52, 17.47it/s]\n",
      "Loss: 1.024 :  49%|████▉     | 4885/10000 [05:13<04:55, 17.31it/s]\n",
      "Loss: 1.024 :  49%|████▉     | 4887/10000 [05:13<05:00, 17.01it/s]\n",
      "Loss: 1.024 :  49%|████▉     | 4889/10000 [05:13<05:01, 16.95it/s]\n",
      "Loss: 1.024 :  49%|████▉     | 4891/10000 [05:14<04:59, 17.09it/s]\n",
      "Loss: 1.024 :  49%|████▉     | 4893/10000 [05:14<05:12, 16.32it/s]\n",
      "Loss: 1.024 :  49%|████▉     | 4895/10000 [05:14<05:34, 15.24it/s]\n",
      "Loss: 1.024 :  49%|████▉     | 4897/10000 [05:14<05:41, 14.96it/s]\n",
      "Loss: 1.024 :  49%|████▉     | 4899/10000 [05:14<05:37, 15.13it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4899/10000 [05:14<05:37, 15.13it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4901/10000 [05:14<06:01, 14.10it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4903/10000 [05:14<05:47, 14.66it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4905/10000 [05:15<05:35, 15.17it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4907/10000 [05:15<05:26, 15.58it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4909/10000 [05:15<05:24, 15.66it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4911/10000 [05:15<05:19, 15.93it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4913/10000 [05:15<05:19, 15.93it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4915/10000 [05:15<05:15, 16.12it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4917/10000 [05:15<05:10, 16.36it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4919/10000 [05:15<05:04, 16.71it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4921/10000 [05:15<05:06, 16.59it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4923/10000 [05:16<05:01, 16.84it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4925/10000 [05:16<05:03, 16.74it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4927/10000 [05:16<05:01, 16.85it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4929/10000 [05:16<04:56, 17.13it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4931/10000 [05:16<04:59, 16.93it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4933/10000 [05:16<05:04, 16.62it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4935/10000 [05:16<05:03, 16.71it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4937/10000 [05:16<05:02, 16.71it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4939/10000 [05:17<04:58, 16.96it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4941/10000 [05:17<04:59, 16.92it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4943/10000 [05:17<04:57, 17.02it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4945/10000 [05:17<04:53, 17.20it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4947/10000 [05:17<04:56, 17.03it/s]\n",
      "Loss: 1.070 :  49%|████▉     | 4949/10000 [05:17<04:55, 17.11it/s]\n",
      "Loss: 0.921 :  49%|████▉     | 4949/10000 [05:17<04:55, 17.11it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4951/10000 [05:17<05:20, 15.78it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4953/10000 [05:17<05:13, 16.08it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4955/10000 [05:18<04:59, 16.87it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4957/10000 [05:18<04:59, 16.84it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4959/10000 [05:18<04:56, 16.98it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4961/10000 [05:18<04:58, 16.86it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4963/10000 [05:18<05:06, 16.41it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4965/10000 [05:18<05:16, 15.90it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4967/10000 [05:18<05:36, 14.98it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4969/10000 [05:18<05:43, 14.63it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4971/10000 [05:19<05:46, 14.49it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4973/10000 [05:19<05:35, 14.97it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4975/10000 [05:19<05:34, 15.03it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4977/10000 [05:19<05:36, 14.94it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4979/10000 [05:19<06:13, 13.45it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4981/10000 [05:19<06:17, 13.30it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4983/10000 [05:19<06:11, 13.51it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4985/10000 [05:20<06:04, 13.75it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4987/10000 [05:20<05:59, 13.93it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4989/10000 [05:20<05:45, 14.50it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4991/10000 [05:20<05:38, 14.78it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4993/10000 [05:20<05:29, 15.19it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4995/10000 [05:20<05:24, 15.42it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4997/10000 [05:20<05:22, 15.52it/s]\n",
      "Loss: 0.921 :  50%|████▉     | 4999/10000 [05:20<05:17, 15.76it/s]\n",
      "Loss: 1.032 :  50%|████▉     | 4999/10000 [05:21<05:17, 15.76it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5001/10000 [05:21<05:35, 14.88it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5003/10000 [05:21<05:20, 15.61it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5005/10000 [05:21<05:12, 15.99it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5007/10000 [05:21<05:08, 16.19it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5009/10000 [05:21<05:00, 16.58it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5011/10000 [05:21<04:55, 16.89it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5013/10000 [05:21<04:56, 16.83it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5015/10000 [05:21<04:56, 16.79it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5017/10000 [05:22<04:57, 16.77it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5019/10000 [05:22<04:49, 17.19it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5021/10000 [05:22<04:51, 17.10it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5023/10000 [05:22<04:50, 17.14it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5025/10000 [05:22<04:51, 17.05it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5027/10000 [05:22<04:51, 17.05it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5029/10000 [05:22<04:49, 17.15it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5031/10000 [05:22<04:47, 17.30it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5033/10000 [05:22<04:47, 17.26it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5035/10000 [05:23<04:42, 17.56it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5037/10000 [05:23<04:44, 17.47it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5039/10000 [05:23<04:50, 17.10it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5041/10000 [05:23<04:50, 17.05it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5043/10000 [05:23<04:50, 17.04it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5045/10000 [05:23<04:52, 16.94it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5047/10000 [05:23<04:52, 16.91it/s]\n",
      "Loss: 1.032 :  50%|█████     | 5049/10000 [05:23<04:57, 16.65it/s]\n",
      "Loss: 1.062 :  50%|█████     | 5049/10000 [05:24<04:57, 16.65it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5051/10000 [05:24<05:22, 15.33it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5053/10000 [05:24<05:11, 15.86it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5055/10000 [05:24<05:05, 16.20it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5057/10000 [05:24<04:55, 16.72it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5059/10000 [05:24<04:49, 17.09it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5061/10000 [05:24<04:51, 16.96it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5063/10000 [05:24<04:54, 16.75it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5065/10000 [05:24<05:20, 15.41it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5067/10000 [05:25<05:31, 14.89it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5069/10000 [05:25<05:29, 14.98it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5071/10000 [05:25<05:21, 15.35it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5073/10000 [05:25<05:23, 15.25it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5075/10000 [05:25<05:20, 15.36it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5077/10000 [05:25<05:14, 15.66it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5079/10000 [05:25<05:12, 15.76it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5081/10000 [05:25<05:10, 15.84it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5083/10000 [05:26<05:09, 15.89it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5085/10000 [05:26<05:01, 16.31it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5087/10000 [05:26<05:04, 16.16it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5089/10000 [05:26<05:00, 16.34it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5091/10000 [05:26<05:03, 16.18it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5093/10000 [05:26<05:06, 16.00it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5095/10000 [05:26<05:05, 16.04it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5097/10000 [05:26<05:05, 16.04it/s]\n",
      "Loss: 1.062 :  51%|█████     | 5099/10000 [05:27<04:59, 16.37it/s]\n",
      "Loss: 0.994 :  51%|█████     | 5099/10000 [05:27<04:59, 16.37it/s]\n",
      "Loss: 0.994 :  51%|█████     | 5101/10000 [05:27<05:30, 14.82it/s]\n",
      "Loss: 0.994 :  51%|█████     | 5103/10000 [05:27<05:17, 15.40it/s]\n",
      "Loss: 0.994 :  51%|█████     | 5105/10000 [05:27<05:05, 16.01it/s]\n",
      "Loss: 0.994 :  51%|█████     | 5107/10000 [05:27<04:59, 16.33it/s]\n",
      "Loss: 0.994 :  51%|█████     | 5109/10000 [05:27<05:02, 16.19it/s]\n",
      "Loss: 0.994 :  51%|█████     | 5111/10000 [05:27<04:59, 16.30it/s]\n",
      "Loss: 0.994 :  51%|█████     | 5113/10000 [05:27<04:58, 16.36it/s]\n",
      "Loss: 0.994 :  51%|█████     | 5115/10000 [05:28<04:47, 16.99it/s]\n",
      "Loss: 0.994 :  51%|█████     | 5117/10000 [05:28<04:49, 16.89it/s]\n",
      "Loss: 0.994 :  51%|█████     | 5119/10000 [05:28<04:52, 16.71it/s]\n",
      "Loss: 0.994 :  51%|█████     | 5121/10000 [05:28<04:54, 16.59it/s]\n",
      "Loss: 0.994 :  51%|█████     | 5123/10000 [05:28<04:55, 16.53it/s]\n",
      "Loss: 0.994 :  51%|█████▏    | 5125/10000 [05:28<04:52, 16.67it/s]\n",
      "Loss: 0.994 :  51%|█████▏    | 5127/10000 [05:28<04:49, 16.84it/s]\n",
      "Loss: 0.994 :  51%|█████▏    | 5129/10000 [05:28<04:46, 17.00it/s]\n",
      "Loss: 0.994 :  51%|█████▏    | 5131/10000 [05:29<04:46, 16.99it/s]\n",
      "Loss: 0.994 :  51%|█████▏    | 5133/10000 [05:29<04:44, 17.14it/s]\n",
      "Loss: 0.994 :  51%|█████▏    | 5135/10000 [05:29<04:45, 17.07it/s]\n",
      "Loss: 0.994 :  51%|█████▏    | 5137/10000 [05:29<04:46, 16.99it/s]\n",
      "Loss: 0.994 :  51%|█████▏    | 5139/10000 [05:29<04:48, 16.87it/s]\n",
      "Loss: 0.994 :  51%|█████▏    | 5141/10000 [05:29<04:46, 16.94it/s]\n",
      "Loss: 0.994 :  51%|█████▏    | 5143/10000 [05:29<04:45, 17.00it/s]\n",
      "Loss: 0.994 :  51%|█████▏    | 5145/10000 [05:29<04:41, 17.24it/s]\n",
      "Loss: 0.994 :  51%|█████▏    | 5147/10000 [05:29<04:44, 17.06it/s]\n",
      "Loss: 0.994 :  51%|█████▏    | 5149/10000 [05:30<04:45, 16.98it/s]\n",
      "Loss: 0.977 :  51%|█████▏    | 5149/10000 [05:30<04:45, 16.98it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5151/10000 [05:30<05:45, 14.03it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5153/10000 [05:30<05:49, 13.87it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5155/10000 [05:30<05:40, 14.25it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5157/10000 [05:30<05:26, 14.81it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5159/10000 [05:30<05:21, 15.06it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5161/10000 [05:30<05:18, 15.17it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5163/10000 [05:31<05:07, 15.72it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5165/10000 [05:31<05:05, 15.84it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5167/10000 [05:31<05:06, 15.78it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5169/10000 [05:31<05:04, 15.88it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5171/10000 [05:31<05:04, 15.84it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5173/10000 [05:31<05:00, 16.08it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5175/10000 [05:31<04:58, 16.19it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5177/10000 [05:31<04:55, 16.35it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5179/10000 [05:32<04:53, 16.44it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5181/10000 [05:32<04:53, 16.44it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5183/10000 [05:32<04:49, 16.64it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5185/10000 [05:32<04:45, 16.86it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5187/10000 [05:32<04:49, 16.63it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5189/10000 [05:32<04:48, 16.65it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5191/10000 [05:32<04:46, 16.76it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5193/10000 [05:32<04:44, 16.90it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5195/10000 [05:32<04:47, 16.72it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5197/10000 [05:33<04:45, 16.80it/s]\n",
      "Loss: 0.977 :  52%|█████▏    | 5199/10000 [05:33<04:46, 16.74it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5199/10000 [05:33<04:46, 16.74it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5201/10000 [05:33<05:15, 15.22it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5203/10000 [05:33<05:04, 15.73it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5205/10000 [05:33<04:52, 16.40it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5207/10000 [05:33<04:52, 16.36it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5209/10000 [05:33<04:51, 16.44it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5211/10000 [05:33<04:54, 16.27it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5213/10000 [05:34<04:47, 16.64it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5215/10000 [05:34<04:41, 16.97it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5217/10000 [05:34<04:40, 17.08it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5219/10000 [05:34<04:39, 17.10it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5221/10000 [05:34<04:40, 17.04it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5223/10000 [05:34<04:39, 17.10it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5225/10000 [05:34<04:37, 17.23it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5227/10000 [05:34<04:38, 17.13it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5229/10000 [05:35<04:37, 17.22it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5231/10000 [05:35<04:37, 17.16it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5233/10000 [05:35<04:41, 16.95it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5235/10000 [05:35<04:34, 17.33it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5237/10000 [05:35<04:58, 15.98it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5239/10000 [05:35<05:15, 15.11it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5241/10000 [05:35<07:01, 11.29it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5243/10000 [05:36<06:38, 11.93it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5245/10000 [05:36<06:11, 12.79it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5247/10000 [05:36<05:58, 13.27it/s]\n",
      "Loss: 1.056 :  52%|█████▏    | 5249/10000 [05:36<05:44, 13.79it/s]\n",
      "Loss: 0.925 :  52%|█████▏    | 5249/10000 [05:36<05:44, 13.79it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5251/10000 [05:36<05:58, 13.23it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5253/10000 [05:36<05:35, 14.17it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5255/10000 [05:36<05:21, 14.77it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5257/10000 [05:37<05:14, 15.08it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5259/10000 [05:37<05:06, 15.48it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5261/10000 [05:37<04:59, 15.80it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5263/10000 [05:37<04:52, 16.21it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5265/10000 [05:37<04:52, 16.17it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5267/10000 [05:37<04:51, 16.25it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5269/10000 [05:37<05:19, 14.83it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5271/10000 [05:37<05:02, 15.62it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5273/10000 [05:38<04:54, 16.06it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5275/10000 [05:38<04:52, 16.17it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5277/10000 [05:38<04:45, 16.55it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5279/10000 [05:38<04:41, 16.74it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5281/10000 [05:38<04:46, 16.44it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5283/10000 [05:38<04:44, 16.57it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5285/10000 [05:38<04:46, 16.47it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5287/10000 [05:38<04:41, 16.74it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5289/10000 [05:38<04:41, 16.76it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5291/10000 [05:39<04:37, 16.97it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5293/10000 [05:39<04:37, 16.97it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5295/10000 [05:39<04:34, 17.15it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5297/10000 [05:39<04:30, 17.39it/s]\n",
      "Loss: 0.925 :  53%|█████▎    | 5299/10000 [05:39<04:29, 17.44it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5299/10000 [05:39<04:29, 17.44it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5301/10000 [05:39<04:59, 15.67it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5303/10000 [05:39<04:56, 15.84it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5305/10000 [05:39<04:51, 16.12it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5307/10000 [05:40<04:43, 16.53it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5309/10000 [05:40<04:39, 16.81it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5311/10000 [05:40<04:37, 16.88it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5313/10000 [05:40<04:39, 16.77it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5315/10000 [05:40<04:34, 17.06it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5317/10000 [05:40<04:33, 17.10it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5319/10000 [05:40<04:35, 16.98it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5321/10000 [05:40<05:02, 15.45it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5323/10000 [05:41<05:10, 15.07it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5325/10000 [05:41<05:08, 15.15it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5327/10000 [05:41<05:00, 15.54it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5329/10000 [05:41<05:02, 15.43it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5331/10000 [05:41<04:59, 15.61it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5333/10000 [05:41<04:47, 16.22it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5335/10000 [05:41<04:49, 16.14it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5337/10000 [05:41<04:44, 16.41it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5339/10000 [05:42<04:43, 16.44it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5341/10000 [05:42<04:45, 16.29it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5343/10000 [05:42<04:44, 16.37it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5345/10000 [05:42<04:38, 16.69it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5347/10000 [05:42<04:38, 16.71it/s]\n",
      "Loss: 0.947 :  53%|█████▎    | 5349/10000 [05:42<04:37, 16.77it/s]\n",
      "Loss: 0.985 :  53%|█████▎    | 5349/10000 [05:42<04:37, 16.77it/s]\n",
      "Loss: 0.985 :  54%|█████▎    | 5351/10000 [05:42<05:06, 15.16it/s]\n",
      "Loss: 0.985 :  54%|█████▎    | 5353/10000 [05:42<04:56, 15.67it/s]\n",
      "Loss: 0.985 :  54%|█████▎    | 5355/10000 [05:43<04:52, 15.89it/s]\n",
      "Loss: 0.985 :  54%|█████▎    | 5357/10000 [05:43<04:51, 15.93it/s]\n",
      "Loss: 0.985 :  54%|█████▎    | 5359/10000 [05:43<04:41, 16.46it/s]\n",
      "Loss: 0.985 :  54%|█████▎    | 5361/10000 [05:43<04:41, 16.48it/s]\n",
      "Loss: 0.985 :  54%|█████▎    | 5363/10000 [05:43<04:36, 16.74it/s]\n",
      "Loss: 0.985 :  54%|█████▎    | 5365/10000 [05:43<04:40, 16.51it/s]\n",
      "Loss: 0.985 :  54%|█████▎    | 5367/10000 [05:43<04:38, 16.66it/s]\n",
      "Loss: 0.985 :  54%|█████▎    | 5369/10000 [05:43<04:33, 16.91it/s]\n",
      "Loss: 0.985 :  54%|█████▎    | 5371/10000 [05:43<04:34, 16.86it/s]\n",
      "Loss: 0.985 :  54%|█████▎    | 5373/10000 [05:44<04:35, 16.78it/s]\n",
      "Loss: 0.985 :  54%|█████▍    | 5375/10000 [05:44<04:38, 16.59it/s]\n",
      "Loss: 0.985 :  54%|█████▍    | 5377/10000 [05:44<04:39, 16.52it/s]\n",
      "Loss: 0.985 :  54%|█████▍    | 5379/10000 [05:44<04:39, 16.56it/s]\n",
      "Loss: 0.985 :  54%|█████▍    | 5381/10000 [05:44<04:38, 16.56it/s]\n",
      "Loss: 0.985 :  54%|█████▍    | 5383/10000 [05:44<04:40, 16.48it/s]\n",
      "Loss: 0.985 :  54%|█████▍    | 5385/10000 [05:44<04:37, 16.61it/s]\n",
      "Loss: 0.985 :  54%|█████▍    | 5387/10000 [05:44<04:38, 16.57it/s]\n",
      "Loss: 0.985 :  54%|█████▍    | 5389/10000 [05:45<04:36, 16.66it/s]\n",
      "Loss: 0.985 :  54%|█████▍    | 5391/10000 [05:45<04:35, 16.75it/s]\n",
      "Loss: 0.985 :  54%|█████▍    | 5393/10000 [05:45<04:30, 17.04it/s]\n",
      "Loss: 0.985 :  54%|█████▍    | 5395/10000 [05:45<04:30, 17.05it/s]\n",
      "Loss: 0.985 :  54%|█████▍    | 5397/10000 [05:45<04:22, 17.53it/s]\n",
      "Loss: 0.985 :  54%|█████▍    | 5399/10000 [05:45<04:22, 17.51it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5399/10000 [05:45<04:22, 17.51it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5401/10000 [05:45<04:52, 15.72it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5403/10000 [05:45<04:49, 15.88it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5405/10000 [05:46<04:49, 15.87it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5407/10000 [05:46<05:10, 14.79it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5409/10000 [05:46<05:12, 14.69it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5411/10000 [05:46<05:06, 14.99it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5413/10000 [05:46<05:01, 15.24it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5415/10000 [05:46<04:56, 15.48it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5417/10000 [05:46<04:52, 15.64it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5419/10000 [05:46<04:47, 15.96it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5421/10000 [05:47<04:45, 16.06it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5423/10000 [05:47<04:45, 16.05it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5425/10000 [05:47<04:41, 16.28it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5427/10000 [05:47<04:40, 16.28it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5429/10000 [05:47<04:38, 16.39it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5431/10000 [05:47<04:34, 16.64it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5433/10000 [05:47<04:35, 16.57it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5435/10000 [05:47<04:36, 16.49it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5437/10000 [05:48<04:38, 16.39it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5439/10000 [05:48<04:37, 16.44it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5441/10000 [05:48<04:39, 16.30it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5443/10000 [05:48<04:34, 16.59it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5445/10000 [05:48<04:30, 16.82it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5447/10000 [05:48<04:31, 16.75it/s]\n",
      "Loss: 0.870 :  54%|█████▍    | 5449/10000 [05:48<04:28, 16.95it/s]\n",
      "Loss: 0.952 :  54%|█████▍    | 5449/10000 [05:48<04:28, 16.95it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5451/10000 [05:48<04:54, 15.45it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5453/10000 [05:49<04:42, 16.11it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5455/10000 [05:49<04:35, 16.50it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5457/10000 [05:49<04:32, 16.69it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5459/10000 [05:49<04:30, 16.78it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5461/10000 [05:49<04:30, 16.76it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5463/10000 [05:49<04:30, 16.76it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5465/10000 [05:49<04:30, 16.74it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5467/10000 [05:49<04:26, 17.03it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5469/10000 [05:49<04:25, 17.08it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5471/10000 [05:50<04:27, 16.91it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5473/10000 [05:50<04:29, 16.77it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5475/10000 [05:50<04:28, 16.86it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5477/10000 [05:50<04:28, 16.84it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5479/10000 [05:50<04:24, 17.08it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5481/10000 [05:50<04:23, 17.17it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5483/10000 [05:50<04:21, 17.25it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5485/10000 [05:50<04:23, 17.12it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5487/10000 [05:51<04:23, 17.14it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5489/10000 [05:51<04:24, 17.05it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5491/10000 [05:51<04:23, 17.10it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5493/10000 [05:51<04:36, 16.31it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5495/10000 [05:51<05:01, 14.92it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5497/10000 [05:51<05:10, 14.50it/s]\n",
      "Loss: 0.952 :  55%|█████▍    | 5499/10000 [05:51<05:09, 14.55it/s]\n",
      "Loss: 0.930 :  55%|█████▍    | 5499/10000 [05:51<05:09, 14.55it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5501/10000 [05:52<05:38, 13.30it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5503/10000 [05:52<05:22, 13.93it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5505/10000 [05:52<05:09, 14.52it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5507/10000 [05:52<05:04, 14.76it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5509/10000 [05:52<04:54, 15.23it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5511/10000 [05:52<04:49, 15.53it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5513/10000 [05:52<04:39, 16.04it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5515/10000 [05:52<04:33, 16.39it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5517/10000 [05:53<04:33, 16.40it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5519/10000 [05:53<04:26, 16.84it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5521/10000 [05:53<04:24, 16.91it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5523/10000 [05:53<04:21, 17.15it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5525/10000 [05:53<04:25, 16.85it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5527/10000 [05:53<04:26, 16.81it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5529/10000 [05:53<04:28, 16.64it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5531/10000 [05:53<04:29, 16.58it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5533/10000 [05:53<04:25, 16.81it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5535/10000 [05:54<04:25, 16.81it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5537/10000 [05:54<04:25, 16.84it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5539/10000 [05:54<04:26, 16.74it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5541/10000 [05:54<04:20, 17.12it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5543/10000 [05:54<04:21, 17.05it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5545/10000 [05:54<04:22, 16.94it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5547/10000 [05:54<04:18, 17.22it/s]\n",
      "Loss: 0.930 :  55%|█████▌    | 5549/10000 [05:54<04:11, 17.69it/s]\n",
      "Loss: 0.849 :  55%|█████▌    | 5549/10000 [05:55<04:11, 17.69it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5551/10000 [05:55<04:42, 15.77it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5553/10000 [05:55<04:37, 16.02it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5555/10000 [05:55<04:35, 16.13it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5557/10000 [05:55<04:34, 16.20it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5559/10000 [05:55<04:29, 16.45it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5561/10000 [05:55<04:28, 16.56it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5563/10000 [05:55<04:29, 16.47it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5565/10000 [05:55<04:26, 16.67it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5567/10000 [05:56<04:24, 16.75it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5569/10000 [05:56<04:20, 17.03it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5571/10000 [05:56<04:17, 17.21it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5573/10000 [05:56<04:18, 17.13it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5575/10000 [05:56<04:17, 17.17it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5577/10000 [05:56<04:16, 17.21it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5579/10000 [05:56<04:16, 17.27it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5581/10000 [05:56<04:42, 15.65it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5583/10000 [05:56<04:56, 14.91it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5585/10000 [05:57<04:55, 14.94it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5587/10000 [05:57<04:47, 15.33it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5589/10000 [05:57<04:44, 15.48it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5591/10000 [05:57<04:42, 15.61it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5593/10000 [05:57<04:39, 15.75it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5595/10000 [05:57<04:33, 16.13it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5597/10000 [05:57<04:28, 16.38it/s]\n",
      "Loss: 0.849 :  56%|█████▌    | 5599/10000 [05:57<04:30, 16.29it/s]\n",
      "Loss: 1.010 :  56%|█████▌    | 5599/10000 [05:58<04:30, 16.29it/s]\n",
      "Loss: 1.010 :  56%|█████▌    | 5601/10000 [05:58<04:50, 15.16it/s]\n",
      "Loss: 1.010 :  56%|█████▌    | 5603/10000 [05:58<04:39, 15.74it/s]\n",
      "Loss: 1.010 :  56%|█████▌    | 5605/10000 [05:58<04:38, 15.80it/s]\n",
      "Loss: 1.010 :  56%|█████▌    | 5607/10000 [05:58<04:34, 16.03it/s]\n",
      "Loss: 1.010 :  56%|█████▌    | 5609/10000 [05:58<04:29, 16.29it/s]\n",
      "Loss: 1.010 :  56%|█████▌    | 5611/10000 [05:58<04:23, 16.65it/s]\n",
      "Loss: 1.010 :  56%|█████▌    | 5613/10000 [05:58<04:21, 16.79it/s]\n",
      "Loss: 1.010 :  56%|█████▌    | 5615/10000 [05:58<04:24, 16.59it/s]\n",
      "Loss: 1.010 :  56%|█████▌    | 5617/10000 [05:59<04:23, 16.61it/s]\n",
      "Loss: 1.010 :  56%|█████▌    | 5619/10000 [05:59<04:20, 16.80it/s]\n",
      "Loss: 1.010 :  56%|█████▌    | 5621/10000 [05:59<04:21, 16.76it/s]\n",
      "Loss: 1.010 :  56%|█████▌    | 5623/10000 [05:59<04:20, 16.81it/s]\n",
      "Loss: 1.010 :  56%|█████▋    | 5625/10000 [05:59<04:17, 17.02it/s]\n",
      "Loss: 1.010 :  56%|█████▋    | 5627/10000 [05:59<04:17, 17.01it/s]\n",
      "Loss: 1.010 :  56%|█████▋    | 5629/10000 [05:59<04:12, 17.31it/s]\n",
      "Loss: 1.010 :  56%|█████▋    | 5631/10000 [05:59<04:13, 17.22it/s]\n",
      "Loss: 1.010 :  56%|█████▋    | 5633/10000 [06:00<04:15, 17.09it/s]\n",
      "Loss: 1.010 :  56%|█████▋    | 5635/10000 [06:00<04:16, 17.01it/s]\n",
      "Loss: 1.010 :  56%|█████▋    | 5637/10000 [06:00<04:18, 16.88it/s]\n",
      "Loss: 1.010 :  56%|█████▋    | 5639/10000 [06:00<04:22, 16.64it/s]\n",
      "Loss: 1.010 :  56%|█████▋    | 5641/10000 [06:00<04:22, 16.60it/s]\n",
      "Loss: 1.010 :  56%|█████▋    | 5643/10000 [06:00<04:20, 16.72it/s]\n",
      "Loss: 1.010 :  56%|█████▋    | 5645/10000 [06:00<04:16, 16.96it/s]\n",
      "Loss: 1.010 :  56%|█████▋    | 5647/10000 [06:00<04:29, 16.18it/s]\n",
      "Loss: 1.010 :  56%|█████▋    | 5649/10000 [06:01<04:37, 15.70it/s]\n",
      "Loss: 0.838 :  56%|█████▋    | 5649/10000 [06:01<04:37, 15.70it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5651/10000 [06:01<05:18, 13.65it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5653/10000 [06:01<05:17, 13.68it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5655/10000 [06:01<05:19, 13.60it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5657/10000 [06:01<05:15, 13.76it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5659/10000 [06:01<05:03, 14.31it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5661/10000 [06:01<04:51, 14.86it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5663/10000 [06:02<04:49, 14.99it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5665/10000 [06:02<05:09, 13.99it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5667/10000 [06:02<05:17, 13.66it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5669/10000 [06:02<05:11, 13.91it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5671/10000 [06:02<05:04, 14.21it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5673/10000 [06:02<04:58, 14.49it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5675/10000 [06:02<04:55, 14.65it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5677/10000 [06:03<04:47, 15.05it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5679/10000 [06:03<04:40, 15.38it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5681/10000 [06:03<04:38, 15.48it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5683/10000 [06:03<04:34, 15.72it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5685/10000 [06:03<04:29, 16.02it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5687/10000 [06:03<04:20, 16.53it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5689/10000 [06:03<04:13, 17.00it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5691/10000 [06:03<04:14, 16.95it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5693/10000 [06:03<04:17, 16.71it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5695/10000 [06:04<04:19, 16.59it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5697/10000 [06:04<04:17, 16.68it/s]\n",
      "Loss: 0.838 :  57%|█████▋    | 5699/10000 [06:04<04:17, 16.68it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5699/10000 [06:04<04:17, 16.68it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5701/10000 [06:04<04:38, 15.44it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5703/10000 [06:04<04:32, 15.74it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5705/10000 [06:04<04:28, 16.02it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5707/10000 [06:04<04:27, 16.03it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5709/10000 [06:04<04:27, 16.05it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5711/10000 [06:05<04:24, 16.24it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5713/10000 [06:05<04:23, 16.24it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5715/10000 [06:05<04:21, 16.38it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5717/10000 [06:05<04:19, 16.51it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5719/10000 [06:05<04:21, 16.35it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5721/10000 [06:05<04:20, 16.43it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5723/10000 [06:05<04:14, 16.79it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5725/10000 [06:05<04:08, 17.18it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5727/10000 [06:06<04:11, 17.00it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5729/10000 [06:06<04:12, 16.91it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5731/10000 [06:06<04:07, 17.24it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5733/10000 [06:06<04:07, 17.26it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5735/10000 [06:06<04:07, 17.22it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5737/10000 [06:06<04:06, 17.30it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5739/10000 [06:06<04:09, 17.11it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5741/10000 [06:06<04:11, 16.96it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5743/10000 [06:06<04:12, 16.84it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5745/10000 [06:07<04:09, 17.07it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5747/10000 [06:07<04:05, 17.30it/s]\n",
      "Loss: 0.706 :  57%|█████▋    | 5749/10000 [06:07<04:02, 17.52it/s]\n",
      "Loss: 0.700 :  57%|█████▋    | 5749/10000 [06:07<04:02, 17.52it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5751/10000 [06:07<04:55, 14.39it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5753/10000 [06:07<04:56, 14.31it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5755/10000 [06:07<04:51, 14.59it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5757/10000 [06:07<04:49, 14.67it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5759/10000 [06:08<04:43, 14.95it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5761/10000 [06:08<04:41, 15.06it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5763/10000 [06:08<04:35, 15.38it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5765/10000 [06:08<04:32, 15.56it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5767/10000 [06:08<04:29, 15.69it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5769/10000 [06:08<04:23, 16.05it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5771/10000 [06:08<04:19, 16.31it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5773/10000 [06:08<04:15, 16.53it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5775/10000 [06:09<04:10, 16.85it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5777/10000 [06:09<04:13, 16.68it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5779/10000 [06:09<04:08, 16.96it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5781/10000 [06:09<04:11, 16.77it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5783/10000 [06:09<04:13, 16.65it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5785/10000 [06:09<04:12, 16.69it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5787/10000 [06:09<04:09, 16.87it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5789/10000 [06:09<04:02, 17.36it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5791/10000 [06:09<04:02, 17.38it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5793/10000 [06:10<04:08, 16.90it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5795/10000 [06:10<04:07, 16.96it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5797/10000 [06:10<04:05, 17.15it/s]\n",
      "Loss: 0.700 :  58%|█████▊    | 5799/10000 [06:10<04:06, 17.04it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5799/10000 [06:10<04:06, 17.04it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5801/10000 [06:10<04:36, 15.17it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5803/10000 [06:10<04:25, 15.78it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5805/10000 [06:10<04:16, 16.38it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5807/10000 [06:10<04:12, 16.61it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5809/10000 [06:11<04:11, 16.69it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5811/10000 [06:11<04:06, 17.03it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5813/10000 [06:11<04:07, 16.92it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5815/10000 [06:11<04:08, 16.86it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5817/10000 [06:11<04:06, 16.96it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5819/10000 [06:11<04:09, 16.76it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5821/10000 [06:11<04:06, 16.95it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5823/10000 [06:11<04:08, 16.82it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5825/10000 [06:12<04:13, 16.48it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5827/10000 [06:12<04:16, 16.28it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5829/10000 [06:12<04:10, 16.67it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5831/10000 [06:12<04:06, 16.88it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5833/10000 [06:12<04:03, 17.10it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5835/10000 [06:12<04:05, 16.99it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5837/10000 [06:12<04:09, 16.71it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5839/10000 [06:12<04:34, 15.15it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5841/10000 [06:13<04:43, 14.67it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5843/10000 [06:13<04:37, 14.96it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5845/10000 [06:13<04:32, 15.28it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5847/10000 [06:13<04:31, 15.32it/s]\n",
      "Loss: 0.847 :  58%|█████▊    | 5849/10000 [06:13<04:29, 15.39it/s]\n",
      "Loss: 0.893 :  58%|█████▊    | 5849/10000 [06:13<04:29, 15.39it/s]\n",
      "Loss: 0.893 :  59%|█████▊    | 5851/10000 [06:13<04:52, 14.19it/s]\n",
      "Loss: 0.893 :  59%|█████▊    | 5853/10000 [06:13<04:43, 14.65it/s]\n",
      "Loss: 0.893 :  59%|█████▊    | 5855/10000 [06:13<04:31, 15.26it/s]\n",
      "Loss: 0.893 :  59%|█████▊    | 5857/10000 [06:14<04:21, 15.84it/s]\n",
      "Loss: 0.893 :  59%|█████▊    | 5859/10000 [06:14<04:20, 15.90it/s]\n",
      "Loss: 0.893 :  59%|█████▊    | 5861/10000 [06:14<04:16, 16.11it/s]\n",
      "Loss: 0.893 :  59%|█████▊    | 5863/10000 [06:14<04:09, 16.58it/s]\n",
      "Loss: 0.893 :  59%|█████▊    | 5865/10000 [06:14<04:07, 16.67it/s]\n",
      "Loss: 0.893 :  59%|█████▊    | 5867/10000 [06:14<04:01, 17.13it/s]\n",
      "Loss: 0.893 :  59%|█████▊    | 5869/10000 [06:14<04:02, 17.02it/s]\n",
      "Loss: 0.893 :  59%|█████▊    | 5871/10000 [06:14<04:02, 17.00it/s]\n",
      "Loss: 0.893 :  59%|█████▊    | 5873/10000 [06:15<04:06, 16.73it/s]\n",
      "Loss: 0.893 :  59%|█████▉    | 5875/10000 [06:15<04:07, 16.67it/s]\n",
      "Loss: 0.893 :  59%|█████▉    | 5877/10000 [06:15<04:06, 16.72it/s]\n",
      "Loss: 0.893 :  59%|█████▉    | 5879/10000 [06:15<04:07, 16.68it/s]\n",
      "Loss: 0.893 :  59%|█████▉    | 5881/10000 [06:15<03:59, 17.23it/s]\n",
      "Loss: 0.893 :  59%|█████▉    | 5883/10000 [06:15<04:01, 17.04it/s]\n",
      "Loss: 0.893 :  59%|█████▉    | 5885/10000 [06:15<04:02, 16.95it/s]\n",
      "Loss: 0.893 :  59%|█████▉    | 5887/10000 [06:15<04:03, 16.89it/s]\n",
      "Loss: 0.893 :  59%|█████▉    | 5889/10000 [06:15<04:03, 16.90it/s]\n",
      "Loss: 0.893 :  59%|█████▉    | 5891/10000 [06:16<04:01, 16.99it/s]\n",
      "Loss: 0.893 :  59%|█████▉    | 5893/10000 [06:16<03:57, 17.27it/s]\n",
      "Loss: 0.893 :  59%|█████▉    | 5895/10000 [06:16<04:00, 17.08it/s]\n",
      "Loss: 0.893 :  59%|█████▉    | 5897/10000 [06:16<04:02, 16.91it/s]\n",
      "Loss: 0.893 :  59%|█████▉    | 5899/10000 [06:16<03:59, 17.15it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5899/10000 [06:16<03:59, 17.15it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5901/10000 [06:16<04:28, 15.29it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5903/10000 [06:16<04:17, 15.92it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5905/10000 [06:16<04:10, 16.34it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5907/10000 [06:17<04:06, 16.60it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5909/10000 [06:17<04:44, 14.36it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5911/10000 [06:17<04:29, 15.16it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5913/10000 [06:17<04:19, 15.77it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5915/10000 [06:17<04:08, 16.41it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5917/10000 [06:17<04:08, 16.41it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5919/10000 [06:17<04:07, 16.46it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5921/10000 [06:17<04:07, 16.50it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5923/10000 [06:18<04:09, 16.35it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5925/10000 [06:18<04:27, 15.22it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5927/10000 [06:18<04:36, 14.72it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5929/10000 [06:18<04:34, 14.82it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5931/10000 [06:18<04:28, 15.14it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5933/10000 [06:18<04:27, 15.22it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5935/10000 [06:18<04:24, 15.38it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5937/10000 [06:18<04:20, 15.62it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5939/10000 [06:19<04:13, 16.01it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5941/10000 [06:19<04:09, 16.24it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5943/10000 [06:19<04:07, 16.41it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5945/10000 [06:19<04:03, 16.62it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5947/10000 [06:19<04:01, 16.76it/s]\n",
      "Loss: 1.065 :  59%|█████▉    | 5949/10000 [06:19<04:06, 16.44it/s]\n",
      "Loss: 0.770 :  59%|█████▉    | 5949/10000 [06:19<04:06, 16.44it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5951/10000 [06:19<04:30, 14.99it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5953/10000 [06:19<04:20, 15.52it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5955/10000 [06:20<04:16, 15.80it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5957/10000 [06:20<04:14, 15.87it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5959/10000 [06:20<04:11, 16.05it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5961/10000 [06:20<04:11, 16.05it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5963/10000 [06:20<04:05, 16.44it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5965/10000 [06:20<04:05, 16.45it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5967/10000 [06:20<04:03, 16.56it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5969/10000 [06:20<04:00, 16.73it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5971/10000 [06:21<04:04, 16.49it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5973/10000 [06:21<03:56, 17.03it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5975/10000 [06:21<03:54, 17.14it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5977/10000 [06:21<03:54, 17.18it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5979/10000 [06:21<03:53, 17.25it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5981/10000 [06:21<03:56, 17.01it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5983/10000 [06:21<03:58, 16.85it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5985/10000 [06:21<03:59, 16.75it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5987/10000 [06:22<03:58, 16.84it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5989/10000 [06:22<03:58, 16.83it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5991/10000 [06:22<03:57, 16.85it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5993/10000 [06:22<03:56, 16.97it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5995/10000 [06:22<03:56, 16.96it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5997/10000 [06:22<03:57, 16.89it/s]\n",
      "Loss: 0.770 :  60%|█████▉    | 5999/10000 [06:22<03:59, 16.70it/s]\n",
      "Loss: 0.935 :  60%|█████▉    | 5999/10000 [06:22<03:59, 16.70it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6001/10000 [06:22<04:23, 15.16it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6003/10000 [06:23<04:18, 15.45it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6005/10000 [06:23<04:11, 15.90it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6007/10000 [06:23<04:07, 16.13it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6009/10000 [06:23<04:13, 15.75it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6011/10000 [06:23<04:29, 14.79it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6013/10000 [06:23<04:38, 14.29it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6015/10000 [06:23<04:32, 14.62it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6017/10000 [06:23<04:27, 14.90it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6019/10000 [06:24<04:24, 15.05it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6021/10000 [06:24<04:24, 15.02it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6023/10000 [06:24<04:20, 15.29it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6025/10000 [06:24<04:20, 15.24it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6027/10000 [06:24<04:13, 15.68it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6029/10000 [06:24<04:10, 15.86it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6031/10000 [06:24<04:08, 15.99it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6033/10000 [06:24<04:05, 16.13it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6035/10000 [06:25<04:07, 16.05it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6037/10000 [06:25<04:03, 16.25it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6039/10000 [06:25<04:00, 16.44it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6041/10000 [06:25<03:58, 16.57it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6043/10000 [06:25<03:59, 16.52it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6045/10000 [06:25<03:57, 16.65it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6047/10000 [06:25<03:59, 16.49it/s]\n",
      "Loss: 0.935 :  60%|██████    | 6049/10000 [06:25<03:58, 16.54it/s]\n",
      "Loss: 0.724 :  60%|██████    | 6049/10000 [06:26<03:58, 16.54it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6051/10000 [06:26<04:23, 15.00it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6053/10000 [06:26<04:15, 15.46it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6055/10000 [06:26<04:07, 15.94it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6057/10000 [06:26<04:03, 16.19it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6059/10000 [06:26<04:03, 16.20it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6061/10000 [06:26<03:59, 16.46it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6063/10000 [06:26<03:52, 16.95it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6065/10000 [06:26<03:46, 17.38it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6067/10000 [06:27<03:45, 17.41it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6069/10000 [06:27<03:46, 17.39it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6071/10000 [06:27<03:50, 17.01it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6073/10000 [06:27<03:52, 16.89it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6075/10000 [06:27<03:50, 17.04it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6077/10000 [06:27<03:50, 16.99it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6079/10000 [06:27<03:56, 16.61it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6081/10000 [06:27<03:57, 16.50it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6083/10000 [06:27<03:54, 16.67it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6085/10000 [06:28<04:03, 16.08it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6087/10000 [06:28<04:08, 15.74it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6089/10000 [06:28<04:05, 15.92it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6091/10000 [06:28<04:06, 15.85it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6093/10000 [06:28<04:06, 15.88it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6095/10000 [06:28<04:19, 15.04it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6097/10000 [06:28<04:27, 14.61it/s]\n",
      "Loss: 0.724 :  61%|██████    | 6099/10000 [06:29<04:27, 14.58it/s]\n",
      "Loss: 0.781 :  61%|██████    | 6099/10000 [06:29<04:27, 14.58it/s]\n",
      "Loss: 0.781 :  61%|██████    | 6101/10000 [06:29<04:52, 13.32it/s]\n",
      "Loss: 0.781 :  61%|██████    | 6103/10000 [06:29<04:41, 13.85it/s]\n",
      "Loss: 0.781 :  61%|██████    | 6105/10000 [06:29<04:35, 14.14it/s]\n",
      "Loss: 0.781 :  61%|██████    | 6107/10000 [06:29<04:21, 14.86it/s]\n",
      "Loss: 0.781 :  61%|██████    | 6109/10000 [06:29<04:17, 15.10it/s]\n",
      "Loss: 0.781 :  61%|██████    | 6111/10000 [06:29<04:12, 15.41it/s]\n",
      "Loss: 0.781 :  61%|██████    | 6113/10000 [06:29<04:06, 15.78it/s]\n",
      "Loss: 0.781 :  61%|██████    | 6115/10000 [06:30<04:03, 15.93it/s]\n",
      "Loss: 0.781 :  61%|██████    | 6117/10000 [06:30<03:57, 16.37it/s]\n",
      "Loss: 0.781 :  61%|██████    | 6119/10000 [06:30<03:55, 16.51it/s]\n",
      "Loss: 0.781 :  61%|██████    | 6121/10000 [06:30<03:52, 16.70it/s]\n",
      "Loss: 0.781 :  61%|██████    | 6123/10000 [06:30<03:53, 16.59it/s]\n",
      "Loss: 0.781 :  61%|██████▏   | 6125/10000 [06:30<03:53, 16.58it/s]\n",
      "Loss: 0.781 :  61%|██████▏   | 6127/10000 [06:30<03:52, 16.63it/s]\n",
      "Loss: 0.781 :  61%|██████▏   | 6129/10000 [06:30<03:49, 16.85it/s]\n",
      "Loss: 0.781 :  61%|██████▏   | 6131/10000 [06:31<03:48, 16.95it/s]\n",
      "Loss: 0.781 :  61%|██████▏   | 6133/10000 [06:31<03:48, 16.90it/s]\n",
      "Loss: 0.781 :  61%|██████▏   | 6135/10000 [06:31<03:48, 16.89it/s]\n",
      "Loss: 0.781 :  61%|██████▏   | 6137/10000 [06:31<03:48, 16.92it/s]\n",
      "Loss: 0.781 :  61%|██████▏   | 6139/10000 [06:31<03:48, 16.91it/s]\n",
      "Loss: 0.781 :  61%|██████▏   | 6141/10000 [06:31<03:49, 16.82it/s]\n",
      "Loss: 0.781 :  61%|██████▏   | 6143/10000 [06:31<03:50, 16.75it/s]\n",
      "Loss: 0.781 :  61%|██████▏   | 6145/10000 [06:31<03:46, 17.02it/s]\n",
      "Loss: 0.781 :  61%|██████▏   | 6147/10000 [06:31<03:47, 16.91it/s]\n",
      "Loss: 0.781 :  61%|██████▏   | 6149/10000 [06:32<04:20, 14.77it/s]\n",
      "Loss: 0.916 :  61%|██████▏   | 6149/10000 [06:32<04:20, 14.77it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6151/10000 [06:32<04:50, 13.27it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6153/10000 [06:32<04:42, 13.63it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6155/10000 [06:32<04:34, 13.99it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6157/10000 [06:32<04:37, 13.83it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6159/10000 [06:32<04:26, 14.39it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6161/10000 [06:33<04:18, 14.83it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6163/10000 [06:33<04:16, 14.99it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6165/10000 [06:33<04:10, 15.33it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6167/10000 [06:33<04:02, 15.81it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6169/10000 [06:33<03:56, 16.21it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6171/10000 [06:33<03:53, 16.38it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6173/10000 [06:33<03:47, 16.83it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6175/10000 [06:33<03:44, 17.02it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6177/10000 [06:33<03:41, 17.23it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6179/10000 [06:34<04:05, 15.56it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6181/10000 [06:34<04:15, 14.93it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6183/10000 [06:34<04:15, 14.92it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6185/10000 [06:34<04:13, 15.04it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6187/10000 [06:34<04:07, 15.40it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6189/10000 [06:34<04:02, 15.73it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6191/10000 [06:34<04:00, 15.81it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6193/10000 [06:35<03:56, 16.08it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6195/10000 [06:35<03:51, 16.41it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6197/10000 [06:35<03:48, 16.66it/s]\n",
      "Loss: 0.916 :  62%|██████▏   | 6199/10000 [06:35<03:46, 16.76it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6199/10000 [06:35<03:46, 16.76it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6201/10000 [06:35<04:16, 14.78it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6203/10000 [06:35<04:13, 14.97it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6205/10000 [06:35<04:01, 15.73it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6207/10000 [06:35<03:56, 16.04it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6209/10000 [06:36<03:50, 16.46it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6211/10000 [06:36<03:47, 16.64it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6213/10000 [06:36<03:47, 16.64it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6215/10000 [06:36<03:45, 16.75it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6217/10000 [06:36<03:47, 16.66it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6219/10000 [06:36<03:45, 16.76it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6221/10000 [06:36<03:41, 17.03it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6223/10000 [06:36<03:42, 16.98it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6225/10000 [06:36<03:44, 16.80it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6227/10000 [06:37<03:45, 16.75it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6229/10000 [06:37<03:39, 17.18it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6231/10000 [06:37<03:37, 17.30it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6233/10000 [06:37<03:36, 17.38it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6235/10000 [06:37<03:35, 17.49it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6237/10000 [06:37<03:38, 17.22it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6239/10000 [06:37<03:39, 17.15it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6241/10000 [06:37<03:42, 16.92it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6243/10000 [06:38<03:40, 17.01it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6245/10000 [06:38<03:40, 17.06it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6247/10000 [06:38<03:38, 17.16it/s]\n",
      "Loss: 0.816 :  62%|██████▏   | 6249/10000 [06:38<03:39, 17.09it/s]\n",
      "Loss: 0.848 :  62%|██████▏   | 6249/10000 [06:38<03:39, 17.09it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6251/10000 [06:38<04:00, 15.56it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6253/10000 [06:38<03:54, 15.95it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6255/10000 [06:38<03:45, 16.59it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6257/10000 [06:38<03:44, 16.64it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6259/10000 [06:38<03:47, 16.45it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6261/10000 [06:39<03:48, 16.39it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6263/10000 [06:39<03:43, 16.73it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6265/10000 [06:39<03:52, 16.07it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6267/10000 [06:39<04:09, 14.93it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6269/10000 [06:39<04:15, 14.60it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6271/10000 [06:39<04:10, 14.91it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6273/10000 [06:39<04:07, 15.03it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6275/10000 [06:40<04:01, 15.41it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6277/10000 [06:40<03:59, 15.51it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6279/10000 [06:40<03:55, 15.77it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6281/10000 [06:40<03:53, 15.91it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6283/10000 [06:40<03:52, 15.97it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6285/10000 [06:40<03:56, 15.71it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6287/10000 [06:40<03:54, 15.86it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6289/10000 [06:40<03:49, 16.15it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6291/10000 [06:41<03:46, 16.36it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6293/10000 [06:41<03:43, 16.62it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6295/10000 [06:41<03:38, 16.96it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6297/10000 [06:41<03:41, 16.69it/s]\n",
      "Loss: 0.848 :  63%|██████▎   | 6299/10000 [06:41<03:39, 16.86it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6299/10000 [06:41<03:39, 16.86it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6301/10000 [06:41<04:04, 15.13it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6303/10000 [06:41<03:59, 15.47it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6305/10000 [06:41<03:54, 15.77it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6307/10000 [06:42<03:45, 16.35it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6309/10000 [06:42<03:38, 16.87it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6311/10000 [06:42<03:39, 16.82it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6313/10000 [06:42<03:34, 17.17it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6315/10000 [06:42<03:34, 17.16it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6317/10000 [06:42<03:38, 16.86it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6319/10000 [06:42<03:38, 16.85it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6321/10000 [06:42<03:39, 16.73it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6323/10000 [06:42<03:35, 17.07it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6325/10000 [06:43<03:33, 17.22it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6327/10000 [06:43<03:34, 17.10it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6329/10000 [06:43<03:35, 17.04it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6331/10000 [06:43<03:37, 16.88it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6333/10000 [06:43<03:41, 16.52it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6335/10000 [06:43<03:41, 16.57it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6337/10000 [06:43<03:39, 16.70it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6339/10000 [06:43<03:38, 16.76it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6341/10000 [06:44<03:36, 16.92it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6343/10000 [06:44<03:36, 16.92it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6345/10000 [06:44<03:38, 16.75it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6347/10000 [06:44<03:42, 16.42it/s]\n",
      "Loss: 0.838 :  63%|██████▎   | 6349/10000 [06:44<03:41, 16.47it/s]\n",
      "Loss: 0.776 :  63%|██████▎   | 6349/10000 [06:44<03:41, 16.47it/s]\n",
      "Loss: 0.776 :  64%|██████▎   | 6351/10000 [06:44<04:11, 14.53it/s]\n",
      "Loss: 0.776 :  64%|██████▎   | 6353/10000 [06:44<04:21, 13.97it/s]\n",
      "Loss: 0.776 :  64%|██████▎   | 6355/10000 [06:44<04:21, 13.96it/s]\n",
      "Loss: 0.776 :  64%|██████▎   | 6357/10000 [06:45<04:13, 14.37it/s]\n",
      "Loss: 0.776 :  64%|██████▎   | 6359/10000 [06:45<04:04, 14.92it/s]\n",
      "Loss: 0.776 :  64%|██████▎   | 6361/10000 [06:45<04:02, 14.99it/s]\n",
      "Loss: 0.776 :  64%|██████▎   | 6363/10000 [06:45<04:02, 15.01it/s]\n",
      "Loss: 0.776 :  64%|██████▎   | 6365/10000 [06:45<04:00, 15.10it/s]\n",
      "Loss: 0.776 :  64%|██████▎   | 6367/10000 [06:45<04:01, 15.05it/s]\n",
      "Loss: 0.776 :  64%|██████▎   | 6369/10000 [06:45<04:04, 14.83it/s]\n",
      "Loss: 0.776 :  64%|██████▎   | 6371/10000 [06:46<04:04, 14.87it/s]\n",
      "Loss: 0.776 :  64%|██████▎   | 6373/10000 [06:46<04:03, 14.87it/s]\n",
      "Loss: 0.776 :  64%|██████▍   | 6375/10000 [06:46<04:05, 14.79it/s]\n",
      "Loss: 0.776 :  64%|██████▍   | 6377/10000 [06:46<04:01, 14.99it/s]\n",
      "Loss: 0.776 :  64%|██████▍   | 6379/10000 [06:46<03:53, 15.52it/s]\n",
      "Loss: 0.776 :  64%|██████▍   | 6381/10000 [06:46<03:54, 15.45it/s]\n",
      "Loss: 0.776 :  64%|██████▍   | 6383/10000 [06:46<03:48, 15.84it/s]\n",
      "Loss: 0.776 :  64%|██████▍   | 6385/10000 [06:46<03:50, 15.68it/s]\n",
      "Loss: 0.776 :  64%|██████▍   | 6387/10000 [06:47<03:47, 15.88it/s]\n",
      "Loss: 0.776 :  64%|██████▍   | 6389/10000 [06:47<03:41, 16.32it/s]\n",
      "Loss: 0.776 :  64%|██████▍   | 6391/10000 [06:47<03:37, 16.58it/s]\n",
      "Loss: 0.776 :  64%|██████▍   | 6393/10000 [06:47<03:45, 16.01it/s]\n",
      "Loss: 0.776 :  64%|██████▍   | 6395/10000 [06:47<03:53, 15.47it/s]\n",
      "Loss: 0.776 :  64%|██████▍   | 6397/10000 [06:47<03:57, 15.19it/s]\n",
      "Loss: 0.776 :  64%|██████▍   | 6399/10000 [06:47<04:02, 14.86it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6399/10000 [06:47<04:02, 14.86it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6401/10000 [06:48<04:35, 13.07it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6403/10000 [06:48<04:32, 13.21it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6405/10000 [06:48<04:33, 13.16it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6407/10000 [06:48<04:32, 13.20it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6409/10000 [06:48<04:30, 13.28it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6411/10000 [06:48<04:21, 13.70it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6413/10000 [06:48<04:12, 14.20it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6415/10000 [06:49<04:05, 14.61it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6417/10000 [06:49<04:01, 14.86it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6419/10000 [06:49<03:53, 15.32it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6421/10000 [06:49<03:49, 15.57it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6423/10000 [06:49<03:41, 16.15it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6425/10000 [06:49<03:38, 16.34it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6427/10000 [06:49<03:36, 16.50it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6429/10000 [06:49<03:34, 16.64it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6431/10000 [06:50<03:43, 15.94it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6433/10000 [06:50<03:59, 14.90it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6435/10000 [06:50<04:06, 14.48it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6437/10000 [06:50<04:04, 14.59it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6439/10000 [06:50<03:57, 15.02it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6441/10000 [06:50<03:54, 15.15it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6443/10000 [06:50<03:53, 15.23it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6445/10000 [06:50<03:50, 15.43it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6447/10000 [06:51<03:49, 15.48it/s]\n",
      "Loss: 0.855 :  64%|██████▍   | 6449/10000 [06:51<03:46, 15.67it/s]\n",
      "Loss: 0.781 :  64%|██████▍   | 6449/10000 [06:51<03:46, 15.67it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6451/10000 [06:51<04:04, 14.49it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6453/10000 [06:51<03:56, 14.99it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6455/10000 [06:51<03:49, 15.47it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6457/10000 [06:51<03:39, 16.11it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6459/10000 [06:51<03:37, 16.25it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6461/10000 [06:51<03:34, 16.52it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6463/10000 [06:52<03:31, 16.70it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6465/10000 [06:52<03:27, 17.04it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6467/10000 [06:52<03:30, 16.79it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6469/10000 [06:52<03:33, 16.57it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6471/10000 [06:52<03:30, 16.73it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6473/10000 [06:52<03:28, 16.92it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6475/10000 [06:52<03:30, 16.79it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6477/10000 [06:52<03:28, 16.90it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6479/10000 [06:53<03:28, 16.91it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6481/10000 [06:53<03:28, 16.85it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6483/10000 [06:53<03:31, 16.66it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6485/10000 [06:53<03:30, 16.70it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6487/10000 [06:53<03:31, 16.65it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6489/10000 [06:53<03:22, 17.35it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6491/10000 [06:53<03:19, 17.55it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6493/10000 [06:53<03:20, 17.49it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6495/10000 [06:53<03:22, 17.30it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6497/10000 [06:54<03:25, 17.05it/s]\n",
      "Loss: 0.781 :  65%|██████▍   | 6499/10000 [06:54<03:26, 16.98it/s]\n",
      "Loss: 0.722 :  65%|██████▍   | 6499/10000 [06:54<03:26, 16.98it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6501/10000 [06:54<03:50, 15.16it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6503/10000 [06:54<03:45, 15.48it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6505/10000 [06:54<03:38, 15.97it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6507/10000 [06:54<03:28, 16.72it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6509/10000 [06:54<03:23, 17.12it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6511/10000 [06:54<03:23, 17.16it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6513/10000 [06:55<03:27, 16.80it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6515/10000 [06:55<03:25, 16.94it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6517/10000 [06:55<03:32, 16.35it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6519/10000 [06:55<03:48, 15.24it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6521/10000 [06:55<03:57, 14.67it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6523/10000 [06:55<03:55, 14.76it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6525/10000 [06:55<03:51, 14.99it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6527/10000 [06:56<03:50, 15.06it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6529/10000 [06:56<03:48, 15.22it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6531/10000 [06:56<03:47, 15.24it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6533/10000 [06:56<03:43, 15.51it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6535/10000 [06:56<03:40, 15.75it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6537/10000 [06:56<03:35, 16.07it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6539/10000 [06:56<03:34, 16.17it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6541/10000 [06:56<03:30, 16.43it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6543/10000 [06:56<03:28, 16.60it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6545/10000 [06:57<03:29, 16.45it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6547/10000 [06:57<03:31, 16.36it/s]\n",
      "Loss: 0.722 :  65%|██████▌   | 6549/10000 [06:57<03:29, 16.51it/s]\n",
      "Loss: 0.738 :  65%|██████▌   | 6549/10000 [06:57<03:29, 16.51it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6551/10000 [06:57<04:21, 13.21it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6553/10000 [06:57<04:04, 14.10it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6555/10000 [06:57<03:50, 14.93it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6557/10000 [06:57<03:42, 15.46it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6559/10000 [06:58<03:33, 16.11it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6561/10000 [06:58<03:26, 16.62it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6563/10000 [06:58<03:28, 16.49it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6565/10000 [06:58<03:25, 16.75it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6567/10000 [06:58<03:26, 16.60it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6569/10000 [06:58<03:22, 16.95it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6571/10000 [06:58<03:18, 17.26it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6573/10000 [06:58<03:18, 17.28it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6575/10000 [06:58<03:22, 16.93it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6577/10000 [06:59<03:16, 17.46it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6579/10000 [06:59<03:16, 17.41it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6581/10000 [06:59<03:18, 17.24it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6583/10000 [06:59<03:17, 17.30it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6585/10000 [06:59<03:18, 17.17it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6587/10000 [06:59<03:19, 17.13it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6589/10000 [06:59<03:19, 17.11it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6591/10000 [06:59<03:17, 17.26it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6593/10000 [07:00<03:21, 16.92it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6595/10000 [07:00<03:20, 17.00it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6597/10000 [07:00<03:21, 16.89it/s]\n",
      "Loss: 0.738 :  66%|██████▌   | 6599/10000 [07:00<03:18, 17.13it/s]\n",
      "Loss: 0.913 :  66%|██████▌   | 6599/10000 [07:00<03:18, 17.13it/s]\n",
      "Loss: 0.913 :  66%|██████▌   | 6601/10000 [07:00<03:37, 15.59it/s]\n",
      "Loss: 0.913 :  66%|██████▌   | 6603/10000 [07:00<03:39, 15.46it/s]\n",
      "Loss: 0.913 :  66%|██████▌   | 6605/10000 [07:00<03:56, 14.38it/s]\n",
      "Loss: 0.913 :  66%|██████▌   | 6607/10000 [07:00<04:00, 14.12it/s]\n",
      "Loss: 0.913 :  66%|██████▌   | 6609/10000 [07:01<03:52, 14.58it/s]\n",
      "Loss: 0.913 :  66%|██████▌   | 6611/10000 [07:01<03:47, 14.93it/s]\n",
      "Loss: 0.913 :  66%|██████▌   | 6613/10000 [07:01<03:43, 15.14it/s]\n",
      "Loss: 0.913 :  66%|██████▌   | 6615/10000 [07:01<03:36, 15.67it/s]\n",
      "Loss: 0.913 :  66%|██████▌   | 6617/10000 [07:01<03:34, 15.79it/s]\n",
      "Loss: 0.913 :  66%|██████▌   | 6619/10000 [07:01<03:41, 15.25it/s]\n",
      "Loss: 0.913 :  66%|██████▌   | 6621/10000 [07:01<03:45, 14.97it/s]\n",
      "Loss: 0.913 :  66%|██████▌   | 6623/10000 [07:02<03:53, 14.44it/s]\n",
      "Loss: 0.913 :  66%|██████▋   | 6625/10000 [07:02<03:59, 14.12it/s]\n",
      "Loss: 0.913 :  66%|██████▋   | 6627/10000 [07:02<04:00, 14.01it/s]\n",
      "Loss: 0.913 :  66%|██████▋   | 6629/10000 [07:02<04:02, 13.92it/s]\n",
      "Loss: 0.913 :  66%|██████▋   | 6631/10000 [07:02<03:55, 14.31it/s]\n",
      "Loss: 0.913 :  66%|██████▋   | 6633/10000 [07:02<03:53, 14.42it/s]\n",
      "Loss: 0.913 :  66%|██████▋   | 6635/10000 [07:02<03:45, 14.95it/s]\n",
      "Loss: 0.913 :  66%|██████▋   | 6637/10000 [07:02<03:42, 15.15it/s]\n",
      "Loss: 0.913 :  66%|██████▋   | 6639/10000 [07:03<03:40, 15.22it/s]\n",
      "Loss: 0.913 :  66%|██████▋   | 6641/10000 [07:03<03:35, 15.61it/s]\n",
      "Loss: 0.913 :  66%|██████▋   | 6643/10000 [07:03<03:28, 16.07it/s]\n",
      "Loss: 0.913 :  66%|██████▋   | 6645/10000 [07:03<03:31, 15.85it/s]\n",
      "Loss: 0.913 :  66%|██████▋   | 6647/10000 [07:03<03:28, 16.08it/s]\n",
      "Loss: 0.913 :  66%|██████▋   | 6649/10000 [07:03<03:23, 16.43it/s]\n",
      "Loss: 0.707 :  66%|██████▋   | 6649/10000 [07:03<03:23, 16.43it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6651/10000 [07:03<03:42, 15.06it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6653/10000 [07:03<03:35, 15.57it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6655/10000 [07:04<03:31, 15.81it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6657/10000 [07:04<03:29, 15.95it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6659/10000 [07:04<03:22, 16.50it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6661/10000 [07:04<03:17, 16.89it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6663/10000 [07:04<03:16, 16.95it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6665/10000 [07:04<03:15, 17.10it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6667/10000 [07:04<03:12, 17.30it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6669/10000 [07:04<03:14, 17.10it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6671/10000 [07:05<03:12, 17.32it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6673/10000 [07:05<03:15, 17.01it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6675/10000 [07:05<03:13, 17.15it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6677/10000 [07:05<03:15, 17.03it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6679/10000 [07:05<03:15, 16.97it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6681/10000 [07:05<03:15, 16.98it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6683/10000 [07:05<03:13, 17.13it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6685/10000 [07:05<03:12, 17.24it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6687/10000 [07:05<03:23, 16.26it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6689/10000 [07:06<03:37, 15.24it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6691/10000 [07:06<03:45, 14.66it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6693/10000 [07:06<03:46, 14.58it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6695/10000 [07:06<04:03, 13.55it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6697/10000 [07:06<04:28, 12.29it/s]\n",
      "Loss: 0.707 :  67%|██████▋   | 6699/10000 [07:06<04:15, 12.90it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6699/10000 [07:07<04:15, 12.90it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6701/10000 [07:07<04:24, 12.47it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6703/10000 [07:07<04:08, 13.26it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6705/10000 [07:07<03:53, 14.14it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6707/10000 [07:07<03:42, 14.79it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6709/10000 [07:07<03:37, 15.15it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6711/10000 [07:07<03:31, 15.52it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6713/10000 [07:07<03:28, 15.77it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6715/10000 [07:07<03:23, 16.15it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6717/10000 [07:08<03:20, 16.38it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6719/10000 [07:08<03:18, 16.50it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6721/10000 [07:08<03:17, 16.59it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6723/10000 [07:08<03:18, 16.54it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6725/10000 [07:08<03:16, 16.66it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6727/10000 [07:08<03:17, 16.54it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6729/10000 [07:08<03:14, 16.78it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6731/10000 [07:08<03:14, 16.83it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6733/10000 [07:09<03:09, 17.23it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6735/10000 [07:09<03:10, 17.10it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6737/10000 [07:09<03:10, 17.13it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6739/10000 [07:09<03:16, 16.62it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6741/10000 [07:09<03:23, 16.02it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6743/10000 [07:09<03:28, 15.59it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6745/10000 [07:09<03:29, 15.56it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6747/10000 [07:09<03:25, 15.84it/s]\n",
      "Loss: 0.814 :  67%|██████▋   | 6749/10000 [07:10<03:23, 15.94it/s]\n",
      "Loss: 0.799 :  67%|██████▋   | 6749/10000 [07:10<03:23, 15.94it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6751/10000 [07:10<03:42, 14.63it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6753/10000 [07:10<03:35, 15.09it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6755/10000 [07:10<03:30, 15.38it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6757/10000 [07:10<03:29, 15.49it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6759/10000 [07:10<03:25, 15.78it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6761/10000 [07:10<03:19, 16.26it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6763/10000 [07:10<03:14, 16.67it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6765/10000 [07:11<03:12, 16.82it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6767/10000 [07:11<03:11, 16.92it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6769/10000 [07:11<03:16, 16.48it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6771/10000 [07:11<03:33, 15.12it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6773/10000 [07:11<03:38, 14.74it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6775/10000 [07:11<03:34, 15.06it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6777/10000 [07:11<03:34, 15.03it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6779/10000 [07:11<03:33, 15.07it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6781/10000 [07:12<03:29, 15.35it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6783/10000 [07:12<03:22, 15.88it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6785/10000 [07:12<03:21, 15.93it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6787/10000 [07:12<03:18, 16.15it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6789/10000 [07:12<03:17, 16.26it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6791/10000 [07:12<03:18, 16.18it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6793/10000 [07:12<03:13, 16.60it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6795/10000 [07:12<03:13, 16.60it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6797/10000 [07:13<03:09, 16.95it/s]\n",
      "Loss: 0.799 :  68%|██████▊   | 6799/10000 [07:13<03:09, 16.92it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6799/10000 [07:13<03:09, 16.92it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6801/10000 [07:13<03:32, 15.03it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6803/10000 [07:13<03:22, 15.76it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6805/10000 [07:13<03:18, 16.06it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6807/10000 [07:13<03:14, 16.40it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6809/10000 [07:13<03:09, 16.84it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6811/10000 [07:13<03:07, 16.97it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6813/10000 [07:14<03:04, 17.24it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6815/10000 [07:14<03:03, 17.35it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6817/10000 [07:14<03:01, 17.56it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6819/10000 [07:14<03:03, 17.38it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6821/10000 [07:14<03:05, 17.18it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6823/10000 [07:14<03:01, 17.53it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6825/10000 [07:14<03:01, 17.45it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6827/10000 [07:14<03:04, 17.20it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6829/10000 [07:14<03:06, 17.04it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6831/10000 [07:15<03:09, 16.76it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6833/10000 [07:15<03:07, 16.88it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6835/10000 [07:15<03:07, 16.90it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6837/10000 [07:15<03:06, 16.93it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6839/10000 [07:15<03:05, 17.07it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6841/10000 [07:15<03:06, 16.97it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6843/10000 [07:15<03:09, 16.66it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6845/10000 [07:15<03:06, 16.93it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6847/10000 [07:16<03:04, 17.06it/s]\n",
      "Loss: 0.637 :  68%|██████▊   | 6849/10000 [07:16<03:03, 17.13it/s]\n",
      "Loss: 0.702 :  68%|██████▊   | 6849/10000 [07:16<03:03, 17.13it/s]\n",
      "Loss: 0.702 :  69%|██████▊   | 6851/10000 [07:16<03:22, 15.52it/s]\n",
      "Loss: 0.702 :  69%|██████▊   | 6853/10000 [07:16<03:17, 15.96it/s]\n",
      "Loss: 0.702 :  69%|██████▊   | 6855/10000 [07:16<03:14, 16.13it/s]\n",
      "Loss: 0.702 :  69%|██████▊   | 6857/10000 [07:16<03:24, 15.39it/s]\n",
      "Loss: 0.702 :  69%|██████▊   | 6859/10000 [07:16<03:33, 14.75it/s]\n",
      "Loss: 0.702 :  69%|██████▊   | 6861/10000 [07:16<03:36, 14.49it/s]\n",
      "Loss: 0.702 :  69%|██████▊   | 6863/10000 [07:17<03:30, 14.88it/s]\n",
      "Loss: 0.702 :  69%|██████▊   | 6865/10000 [07:17<03:23, 15.37it/s]\n",
      "Loss: 0.702 :  69%|██████▊   | 6867/10000 [07:17<03:21, 15.52it/s]\n",
      "Loss: 0.702 :  69%|██████▊   | 6869/10000 [07:17<03:19, 15.70it/s]\n",
      "Loss: 0.702 :  69%|██████▊   | 6871/10000 [07:17<03:18, 15.77it/s]\n",
      "Loss: 0.702 :  69%|██████▊   | 6873/10000 [07:17<03:17, 15.86it/s]\n",
      "Loss: 0.702 :  69%|██████▉   | 6875/10000 [07:17<03:14, 16.04it/s]\n",
      "Loss: 0.702 :  69%|██████▉   | 6877/10000 [07:17<03:14, 16.08it/s]\n",
      "Loss: 0.702 :  69%|██████▉   | 6879/10000 [07:18<03:12, 16.21it/s]\n",
      "Loss: 0.702 :  69%|██████▉   | 6881/10000 [07:18<03:10, 16.34it/s]\n",
      "Loss: 0.702 :  69%|██████▉   | 6883/10000 [07:18<03:07, 16.66it/s]\n",
      "Loss: 0.702 :  69%|██████▉   | 6885/10000 [07:18<03:08, 16.54it/s]\n",
      "Loss: 0.702 :  69%|██████▉   | 6887/10000 [07:18<03:07, 16.62it/s]\n",
      "Loss: 0.702 :  69%|██████▉   | 6889/10000 [07:18<03:08, 16.52it/s]\n",
      "Loss: 0.702 :  69%|██████▉   | 6891/10000 [07:18<03:07, 16.56it/s]\n",
      "Loss: 0.702 :  69%|██████▉   | 6893/10000 [07:18<03:03, 16.91it/s]\n",
      "Loss: 0.702 :  69%|██████▉   | 6895/10000 [07:19<03:02, 17.02it/s]\n",
      "Loss: 0.702 :  69%|██████▉   | 6897/10000 [07:19<03:03, 16.94it/s]\n",
      "Loss: 0.702 :  69%|██████▉   | 6899/10000 [07:19<03:00, 17.22it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6899/10000 [07:19<03:00, 17.22it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6901/10000 [07:19<03:15, 15.82it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6903/10000 [07:19<03:14, 15.93it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6905/10000 [07:19<03:08, 16.39it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6907/10000 [07:19<03:05, 16.70it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6909/10000 [07:19<03:03, 16.87it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6911/10000 [07:19<03:04, 16.77it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6913/10000 [07:20<03:04, 16.78it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6915/10000 [07:20<03:01, 16.97it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6917/10000 [07:20<03:02, 16.85it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6919/10000 [07:20<03:02, 16.86it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6921/10000 [07:20<03:02, 16.85it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6923/10000 [07:20<03:04, 16.69it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6925/10000 [07:20<03:01, 16.90it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6927/10000 [07:20<03:03, 16.79it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6929/10000 [07:21<03:02, 16.83it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6931/10000 [07:21<03:01, 16.90it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6933/10000 [07:21<02:56, 17.36it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6935/10000 [07:21<03:00, 17.01it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6937/10000 [07:21<03:01, 16.92it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6939/10000 [07:21<03:02, 16.74it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6941/10000 [07:21<03:03, 16.65it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6943/10000 [07:21<03:02, 16.71it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6945/10000 [07:22<03:18, 15.37it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6947/10000 [07:22<03:24, 14.93it/s]\n",
      "Loss: 0.834 :  69%|██████▉   | 6949/10000 [07:22<03:23, 14.96it/s]\n",
      "Loss: 0.889 :  69%|██████▉   | 6949/10000 [07:22<03:23, 14.96it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6951/10000 [07:22<03:43, 13.65it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6953/10000 [07:22<03:36, 14.05it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6955/10000 [07:22<03:28, 14.59it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6957/10000 [07:22<03:21, 15.14it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6959/10000 [07:23<03:19, 15.22it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6961/10000 [07:23<03:18, 15.32it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6963/10000 [07:23<03:14, 15.62it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6965/10000 [07:23<03:13, 15.65it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6967/10000 [07:23<03:12, 15.72it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6969/10000 [07:23<03:09, 16.00it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6971/10000 [07:23<03:06, 16.26it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6973/10000 [07:23<03:03, 16.47it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6975/10000 [07:23<03:00, 16.78it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6977/10000 [07:24<03:00, 16.75it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6979/10000 [07:24<03:02, 16.56it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6981/10000 [07:24<03:04, 16.33it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6983/10000 [07:24<03:05, 16.29it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6985/10000 [07:24<03:03, 16.39it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6987/10000 [07:24<03:01, 16.56it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6989/10000 [07:24<03:01, 16.59it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6991/10000 [07:24<03:02, 16.53it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6993/10000 [07:25<02:56, 17.03it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6995/10000 [07:25<02:55, 17.16it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6997/10000 [07:25<02:54, 17.25it/s]\n",
      "Loss: 0.889 :  70%|██████▉   | 6999/10000 [07:25<02:56, 17.02it/s]\n",
      "Loss: 0.650 :  70%|██████▉   | 6999/10000 [07:25<02:56, 17.02it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7001/10000 [07:25<03:17, 15.20it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7003/10000 [07:25<03:11, 15.65it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7005/10000 [07:25<03:06, 16.09it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7007/10000 [07:25<03:02, 16.39it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7009/10000 [07:26<02:57, 16.81it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7011/10000 [07:26<02:57, 16.80it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7013/10000 [07:26<02:59, 16.63it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7015/10000 [07:26<02:59, 16.63it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7017/10000 [07:26<02:56, 16.87it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7019/10000 [07:26<02:58, 16.73it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7021/10000 [07:26<02:53, 17.13it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7023/10000 [07:26<02:52, 17.23it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7025/10000 [07:26<02:54, 17.08it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7027/10000 [07:27<02:55, 16.94it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7029/10000 [07:27<02:51, 17.29it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7031/10000 [07:27<03:08, 15.72it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7033/10000 [07:27<03:14, 15.23it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7035/10000 [07:27<03:17, 14.98it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7037/10000 [07:27<03:16, 15.06it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7039/10000 [07:27<03:12, 15.36it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7041/10000 [07:28<03:10, 15.57it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7043/10000 [07:28<03:09, 15.61it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7045/10000 [07:28<03:09, 15.60it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7047/10000 [07:28<03:07, 15.73it/s]\n",
      "Loss: 0.650 :  70%|███████   | 7049/10000 [07:28<03:07, 15.78it/s]\n",
      "Loss: 0.664 :  70%|███████   | 7049/10000 [07:28<03:07, 15.78it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7051/10000 [07:28<03:26, 14.27it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7053/10000 [07:28<03:17, 14.89it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7055/10000 [07:28<03:11, 15.40it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7057/10000 [07:29<03:05, 15.85it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7059/10000 [07:29<02:59, 16.34it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7061/10000 [07:29<03:02, 16.14it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7063/10000 [07:29<02:58, 16.50it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7065/10000 [07:29<02:58, 16.40it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7067/10000 [07:29<03:00, 16.26it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7069/10000 [07:29<02:55, 16.68it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7071/10000 [07:29<02:55, 16.71it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7073/10000 [07:30<02:57, 16.49it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7075/10000 [07:30<02:57, 16.49it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7077/10000 [07:30<02:55, 16.62it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7079/10000 [07:30<02:54, 16.74it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7081/10000 [07:30<02:54, 16.71it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7083/10000 [07:30<02:55, 16.66it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7085/10000 [07:30<02:53, 16.75it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7087/10000 [07:30<02:53, 16.83it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7089/10000 [07:30<02:56, 16.49it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7091/10000 [07:31<02:57, 16.39it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7093/10000 [07:31<02:55, 16.59it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7095/10000 [07:31<02:54, 16.64it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7097/10000 [07:31<02:54, 16.66it/s]\n",
      "Loss: 0.664 :  71%|███████   | 7099/10000 [07:31<02:53, 16.71it/s]\n",
      "Loss: 0.675 :  71%|███████   | 7099/10000 [07:31<02:53, 16.71it/s]\n",
      "Loss: 0.675 :  71%|███████   | 7101/10000 [07:31<03:11, 15.16it/s]\n",
      "Loss: 0.675 :  71%|███████   | 7103/10000 [07:31<03:06, 15.56it/s]\n",
      "Loss: 0.675 :  71%|███████   | 7105/10000 [07:31<03:00, 16.07it/s]\n",
      "Loss: 0.675 :  71%|███████   | 7107/10000 [07:32<03:00, 16.05it/s]\n",
      "Loss: 0.675 :  71%|███████   | 7109/10000 [07:32<03:01, 15.95it/s]\n",
      "Loss: 0.675 :  71%|███████   | 7111/10000 [07:32<02:59, 16.13it/s]\n",
      "Loss: 0.675 :  71%|███████   | 7113/10000 [07:32<03:00, 15.99it/s]\n",
      "Loss: 0.675 :  71%|███████   | 7115/10000 [07:32<03:12, 15.01it/s]\n",
      "Loss: 0.675 :  71%|███████   | 7117/10000 [07:32<03:20, 14.38it/s]\n",
      "Loss: 0.675 :  71%|███████   | 7119/10000 [07:32<03:18, 14.53it/s]\n",
      "Loss: 0.675 :  71%|███████   | 7121/10000 [07:33<03:15, 14.74it/s]\n",
      "Loss: 0.675 :  71%|███████   | 7123/10000 [07:33<03:13, 14.86it/s]\n",
      "Loss: 0.675 :  71%|███████▏  | 7125/10000 [07:33<03:09, 15.18it/s]\n",
      "Loss: 0.675 :  71%|███████▏  | 7127/10000 [07:33<03:04, 15.60it/s]\n",
      "Loss: 0.675 :  71%|███████▏  | 7129/10000 [07:33<03:03, 15.61it/s]\n",
      "Loss: 0.675 :  71%|███████▏  | 7131/10000 [07:33<03:03, 15.66it/s]\n",
      "Loss: 0.675 :  71%|███████▏  | 7133/10000 [07:33<03:01, 15.82it/s]\n",
      "Loss: 0.675 :  71%|███████▏  | 7135/10000 [07:33<02:58, 16.02it/s]\n",
      "Loss: 0.675 :  71%|███████▏  | 7137/10000 [07:34<02:59, 15.93it/s]\n",
      "Loss: 0.675 :  71%|███████▏  | 7139/10000 [07:34<02:56, 16.24it/s]\n",
      "Loss: 0.675 :  71%|███████▏  | 7141/10000 [07:34<02:54, 16.40it/s]\n",
      "Loss: 0.675 :  71%|███████▏  | 7143/10000 [07:34<02:53, 16.49it/s]\n",
      "Loss: 0.675 :  71%|███████▏  | 7145/10000 [07:34<02:53, 16.49it/s]\n",
      "Loss: 0.675 :  71%|███████▏  | 7147/10000 [07:34<02:52, 16.50it/s]\n",
      "Loss: 0.675 :  71%|███████▏  | 7149/10000 [07:34<02:51, 16.58it/s]\n",
      "Loss: 0.738 :  71%|███████▏  | 7149/10000 [07:34<02:51, 16.58it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7151/10000 [07:34<03:10, 14.96it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7153/10000 [07:35<03:03, 15.50it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7155/10000 [07:35<02:58, 15.96it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7157/10000 [07:35<02:55, 16.21it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7159/10000 [07:35<02:52, 16.43it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7161/10000 [07:35<02:53, 16.38it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7163/10000 [07:35<02:54, 16.28it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7165/10000 [07:35<02:52, 16.47it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7167/10000 [07:35<02:51, 16.56it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7169/10000 [07:36<02:51, 16.49it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7171/10000 [07:36<02:47, 16.84it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7173/10000 [07:36<02:46, 17.02it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7175/10000 [07:36<02:49, 16.68it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7177/10000 [07:36<02:48, 16.75it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7179/10000 [07:36<03:10, 14.83it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7181/10000 [07:36<03:03, 15.34it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7183/10000 [07:36<02:56, 15.99it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7185/10000 [07:37<02:51, 16.40it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7187/10000 [07:37<02:46, 16.90it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7189/10000 [07:37<02:46, 16.91it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7191/10000 [07:37<02:46, 16.83it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7193/10000 [07:37<02:42, 17.27it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7195/10000 [07:37<02:44, 17.09it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7197/10000 [07:37<02:44, 17.06it/s]\n",
      "Loss: 0.738 :  72%|███████▏  | 7199/10000 [07:37<02:41, 17.31it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7199/10000 [07:37<02:41, 17.31it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7201/10000 [07:38<03:20, 13.98it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7203/10000 [07:38<03:19, 14.02it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7205/10000 [07:38<03:11, 14.59it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7207/10000 [07:38<03:07, 14.87it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7209/10000 [07:38<03:02, 15.27it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7211/10000 [07:38<02:57, 15.72it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7213/10000 [07:38<02:54, 15.96it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7215/10000 [07:38<02:52, 16.14it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7217/10000 [07:39<02:54, 15.94it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7219/10000 [07:39<02:53, 16.07it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7221/10000 [07:39<02:50, 16.25it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7223/10000 [07:39<02:50, 16.33it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7225/10000 [07:39<02:48, 16.46it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7227/10000 [07:39<02:46, 16.62it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7229/10000 [07:39<02:43, 16.94it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7231/10000 [07:39<02:42, 17.02it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7233/10000 [07:39<02:41, 17.14it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7235/10000 [07:40<02:39, 17.30it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7237/10000 [07:40<02:41, 17.08it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7239/10000 [07:40<02:41, 17.05it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7241/10000 [07:40<02:43, 16.88it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7243/10000 [07:40<02:42, 16.95it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7245/10000 [07:40<02:42, 17.00it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7247/10000 [07:40<02:41, 17.10it/s]\n",
      "Loss: 0.619 :  72%|███████▏  | 7249/10000 [07:40<02:40, 17.15it/s]\n",
      "Loss: 0.863 :  72%|███████▏  | 7249/10000 [07:41<02:40, 17.15it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7251/10000 [07:41<02:58, 15.40it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7253/10000 [07:41<02:55, 15.64it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7255/10000 [07:41<02:48, 16.31it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7257/10000 [07:41<02:49, 16.18it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7259/10000 [07:41<02:47, 16.34it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7261/10000 [07:41<02:45, 16.54it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7263/10000 [07:41<02:42, 16.89it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7265/10000 [07:41<02:42, 16.86it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7267/10000 [07:42<02:44, 16.65it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7269/10000 [07:42<02:40, 16.99it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7271/10000 [07:42<02:40, 17.00it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7273/10000 [07:42<02:44, 16.58it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7275/10000 [07:42<02:44, 16.54it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7277/10000 [07:42<02:42, 16.74it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7279/10000 [07:42<02:41, 16.88it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7281/10000 [07:42<02:40, 16.94it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7283/10000 [07:42<02:42, 16.72it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7285/10000 [07:43<02:43, 16.58it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7287/10000 [07:43<02:47, 16.17it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7289/10000 [07:43<02:57, 15.30it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7291/10000 [07:43<03:02, 14.84it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7293/10000 [07:43<03:00, 15.04it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7295/10000 [07:43<02:58, 15.15it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7297/10000 [07:43<02:57, 15.23it/s]\n",
      "Loss: 0.863 :  73%|███████▎  | 7299/10000 [07:44<02:54, 15.45it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7299/10000 [07:44<02:54, 15.45it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7301/10000 [07:44<03:11, 14.07it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7303/10000 [07:44<03:04, 14.63it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7305/10000 [07:44<02:58, 15.14it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7307/10000 [07:44<02:54, 15.46it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7309/10000 [07:44<02:48, 15.96it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7311/10000 [07:44<02:47, 16.08it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7313/10000 [07:44<02:46, 16.12it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7315/10000 [07:45<02:43, 16.39it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7317/10000 [07:45<02:40, 16.72it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7319/10000 [07:45<02:39, 16.80it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7321/10000 [07:45<02:39, 16.83it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7323/10000 [07:45<02:39, 16.83it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7325/10000 [07:45<02:36, 17.04it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7327/10000 [07:45<02:36, 17.04it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7329/10000 [07:45<02:39, 16.72it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7331/10000 [07:45<02:38, 16.83it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7333/10000 [07:46<02:38, 16.86it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7335/10000 [07:46<02:38, 16.86it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7337/10000 [07:46<02:37, 16.95it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7339/10000 [07:46<02:35, 17.06it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7341/10000 [07:46<02:36, 16.98it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7343/10000 [07:46<02:36, 17.01it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7345/10000 [07:46<02:33, 17.29it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7347/10000 [07:46<02:31, 17.57it/s]\n",
      "Loss: 0.697 :  73%|███████▎  | 7349/10000 [07:47<02:34, 17.16it/s]\n",
      "Loss: 0.887 :  73%|███████▎  | 7349/10000 [07:47<02:34, 17.16it/s]\n",
      "Loss: 0.887 :  74%|███████▎  | 7351/10000 [07:47<02:55, 15.12it/s]\n",
      "Loss: 0.887 :  74%|███████▎  | 7353/10000 [07:47<02:49, 15.65it/s]\n",
      "Loss: 0.887 :  74%|███████▎  | 7355/10000 [07:47<02:45, 15.94it/s]\n",
      "Loss: 0.887 :  74%|███████▎  | 7357/10000 [07:47<02:43, 16.16it/s]\n",
      "Loss: 0.887 :  74%|███████▎  | 7359/10000 [07:47<02:42, 16.22it/s]\n",
      "Loss: 0.887 :  74%|███████▎  | 7361/10000 [07:47<02:42, 16.28it/s]\n",
      "Loss: 0.887 :  74%|███████▎  | 7363/10000 [07:47<02:43, 16.11it/s]\n",
      "Loss: 0.887 :  74%|███████▎  | 7365/10000 [07:48<02:43, 16.16it/s]\n",
      "Loss: 0.887 :  74%|███████▎  | 7367/10000 [07:48<02:40, 16.45it/s]\n",
      "Loss: 0.887 :  74%|███████▎  | 7369/10000 [07:48<02:38, 16.62it/s]\n",
      "Loss: 0.887 :  74%|███████▎  | 7371/10000 [07:48<02:37, 16.66it/s]\n",
      "Loss: 0.887 :  74%|███████▎  | 7373/10000 [07:48<02:47, 15.72it/s]\n",
      "Loss: 0.887 :  74%|███████▍  | 7375/10000 [07:48<02:58, 14.72it/s]\n",
      "Loss: 0.887 :  74%|███████▍  | 7377/10000 [07:48<02:58, 14.72it/s]\n",
      "Loss: 0.887 :  74%|███████▍  | 7379/10000 [07:48<02:53, 15.11it/s]\n",
      "Loss: 0.887 :  74%|███████▍  | 7381/10000 [07:49<02:52, 15.19it/s]\n",
      "Loss: 0.887 :  74%|███████▍  | 7383/10000 [07:49<02:53, 15.09it/s]\n",
      "Loss: 0.887 :  74%|███████▍  | 7385/10000 [07:49<02:51, 15.27it/s]\n",
      "Loss: 0.887 :  74%|███████▍  | 7387/10000 [07:49<02:49, 15.42it/s]\n",
      "Loss: 0.887 :  74%|███████▍  | 7389/10000 [07:49<02:49, 15.44it/s]\n",
      "Loss: 0.887 :  74%|███████▍  | 7391/10000 [07:49<02:45, 15.80it/s]\n",
      "Loss: 0.887 :  74%|███████▍  | 7393/10000 [07:49<02:41, 16.14it/s]\n",
      "Loss: 0.887 :  74%|███████▍  | 7395/10000 [07:49<02:42, 16.07it/s]\n",
      "Loss: 0.887 :  74%|███████▍  | 7397/10000 [07:50<02:38, 16.38it/s]\n",
      "Loss: 0.887 :  74%|███████▍  | 7399/10000 [07:50<02:39, 16.33it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7399/10000 [07:50<02:39, 16.33it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7401/10000 [07:50<02:53, 15.01it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7403/10000 [07:50<02:47, 15.46it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7405/10000 [07:50<02:44, 15.77it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7407/10000 [07:50<02:39, 16.23it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7409/10000 [07:50<02:39, 16.29it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7411/10000 [07:50<02:39, 16.22it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7413/10000 [07:51<02:38, 16.34it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7415/10000 [07:51<02:33, 16.85it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7417/10000 [07:51<02:34, 16.77it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7419/10000 [07:51<02:34, 16.68it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7421/10000 [07:51<02:34, 16.70it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7423/10000 [07:51<02:32, 16.94it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7425/10000 [07:51<02:31, 16.98it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7427/10000 [07:51<02:34, 16.62it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7429/10000 [07:52<02:35, 16.59it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7431/10000 [07:52<02:34, 16.60it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7433/10000 [07:52<02:35, 16.53it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7435/10000 [07:52<02:35, 16.44it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7437/10000 [07:52<02:34, 16.57it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7439/10000 [07:52<02:34, 16.61it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7441/10000 [07:52<02:32, 16.75it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7443/10000 [07:52<02:34, 16.50it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7445/10000 [07:53<02:37, 16.26it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7447/10000 [07:53<02:38, 16.11it/s]\n",
      "Loss: 0.766 :  74%|███████▍  | 7449/10000 [07:53<02:36, 16.33it/s]\n",
      "Loss: 0.684 :  74%|███████▍  | 7449/10000 [07:53<02:36, 16.33it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7451/10000 [07:53<02:49, 15.04it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7453/10000 [07:53<02:41, 15.73it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7455/10000 [07:53<02:35, 16.38it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7457/10000 [07:53<02:30, 16.85it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7459/10000 [07:53<02:42, 15.64it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7461/10000 [07:54<02:46, 15.29it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7463/10000 [07:54<02:45, 15.37it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7465/10000 [07:54<02:42, 15.57it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7467/10000 [07:54<02:42, 15.61it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7469/10000 [07:54<02:43, 15.53it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7471/10000 [07:54<02:39, 15.90it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7473/10000 [07:54<02:37, 16.00it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7475/10000 [07:54<02:35, 16.19it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7477/10000 [07:55<02:36, 16.14it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7479/10000 [07:55<02:34, 16.27it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7481/10000 [07:55<02:34, 16.33it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7483/10000 [07:55<02:33, 16.41it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7485/10000 [07:55<02:33, 16.39it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7487/10000 [07:55<02:32, 16.46it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7489/10000 [07:55<02:31, 16.59it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7491/10000 [07:55<02:30, 16.68it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7493/10000 [07:56<02:29, 16.83it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7495/10000 [07:56<02:28, 16.85it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7497/10000 [07:56<02:27, 16.97it/s]\n",
      "Loss: 0.684 :  75%|███████▍  | 7499/10000 [07:56<02:26, 17.02it/s]\n",
      "Loss: 0.649 :  75%|███████▍  | 7499/10000 [07:56<02:26, 17.02it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7501/10000 [07:56<02:43, 15.26it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7503/10000 [07:56<02:39, 15.64it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7505/10000 [07:56<02:36, 15.91it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7507/10000 [07:56<02:36, 15.90it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7509/10000 [07:57<02:32, 16.29it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7511/10000 [07:57<02:30, 16.57it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7513/10000 [07:57<02:28, 16.75it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7515/10000 [07:57<02:26, 16.97it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7517/10000 [07:57<02:23, 17.28it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7519/10000 [07:57<02:24, 17.14it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7521/10000 [07:57<02:23, 17.24it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7523/10000 [07:57<02:22, 17.36it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7525/10000 [07:57<02:29, 16.55it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7527/10000 [07:58<02:35, 15.88it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7529/10000 [07:58<02:36, 15.82it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7531/10000 [07:58<02:35, 15.85it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7533/10000 [07:58<02:34, 15.94it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7535/10000 [07:58<02:34, 15.95it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7537/10000 [07:58<02:34, 15.91it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7539/10000 [07:58<02:40, 15.31it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7541/10000 [07:59<02:51, 14.30it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7543/10000 [07:59<02:57, 13.84it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7545/10000 [07:59<03:10, 12.92it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7547/10000 [07:59<03:11, 12.81it/s]\n",
      "Loss: 0.649 :  75%|███████▌  | 7549/10000 [07:59<03:09, 12.94it/s]\n",
      "Loss: 0.705 :  75%|███████▌  | 7549/10000 [07:59<03:09, 12.94it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7551/10000 [07:59<03:20, 12.21it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7553/10000 [07:59<03:13, 12.65it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7555/10000 [08:00<03:21, 12.11it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7557/10000 [08:00<03:30, 11.62it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7559/10000 [08:00<03:34, 11.40it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7561/10000 [08:00<03:19, 12.21it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7563/10000 [08:00<03:09, 12.86it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7565/10000 [08:00<03:04, 13.23it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7567/10000 [08:01<02:55, 13.90it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7569/10000 [08:01<02:48, 14.39it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7571/10000 [08:01<02:43, 14.85it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7573/10000 [08:01<02:39, 15.19it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7575/10000 [08:01<02:35, 15.58it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7577/10000 [08:01<02:35, 15.62it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7579/10000 [08:01<02:30, 16.13it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7581/10000 [08:01<02:27, 16.35it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7583/10000 [08:02<02:26, 16.48it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7585/10000 [08:02<02:25, 16.54it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7587/10000 [08:02<02:24, 16.71it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7589/10000 [08:02<02:22, 16.90it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7591/10000 [08:02<02:27, 16.29it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7593/10000 [08:02<02:32, 15.74it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7595/10000 [08:02<02:37, 15.28it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7597/10000 [08:02<02:43, 14.69it/s]\n",
      "Loss: 0.705 :  76%|███████▌  | 7599/10000 [08:03<02:46, 14.41it/s]\n",
      "Loss: 0.650 :  76%|███████▌  | 7599/10000 [08:03<02:46, 14.41it/s]\n",
      "Loss: 0.650 :  76%|███████▌  | 7601/10000 [08:03<03:05, 12.92it/s]\n",
      "Loss: 0.650 :  76%|███████▌  | 7603/10000 [08:03<02:56, 13.58it/s]\n",
      "Loss: 0.650 :  76%|███████▌  | 7605/10000 [08:03<02:50, 14.04it/s]\n",
      "Loss: 0.650 :  76%|███████▌  | 7607/10000 [08:03<02:45, 14.43it/s]\n",
      "Loss: 0.650 :  76%|███████▌  | 7609/10000 [08:03<02:41, 14.85it/s]\n",
      "Loss: 0.650 :  76%|███████▌  | 7611/10000 [08:03<02:37, 15.16it/s]\n",
      "Loss: 0.650 :  76%|███████▌  | 7613/10000 [08:04<02:35, 15.38it/s]\n",
      "Loss: 0.650 :  76%|███████▌  | 7615/10000 [08:04<02:31, 15.72it/s]\n",
      "Loss: 0.650 :  76%|███████▌  | 7617/10000 [08:04<02:29, 15.93it/s]\n",
      "Loss: 0.650 :  76%|███████▌  | 7619/10000 [08:04<02:28, 16.07it/s]\n",
      "Loss: 0.650 :  76%|███████▌  | 7621/10000 [08:04<02:27, 16.11it/s]\n",
      "Loss: 0.650 :  76%|███████▌  | 7623/10000 [08:04<02:39, 14.93it/s]\n",
      "Loss: 0.650 :  76%|███████▋  | 7625/10000 [08:04<02:43, 14.49it/s]\n",
      "Loss: 0.650 :  76%|███████▋  | 7627/10000 [08:04<02:40, 14.78it/s]\n",
      "Loss: 0.650 :  76%|███████▋  | 7629/10000 [08:05<02:35, 15.27it/s]\n",
      "Loss: 0.650 :  76%|███████▋  | 7631/10000 [08:05<02:35, 15.19it/s]\n",
      "Loss: 0.650 :  76%|███████▋  | 7633/10000 [08:05<02:33, 15.38it/s]\n",
      "Loss: 0.650 :  76%|███████▋  | 7635/10000 [08:05<02:32, 15.54it/s]\n",
      "Loss: 0.650 :  76%|███████▋  | 7637/10000 [08:05<02:28, 15.91it/s]\n",
      "Loss: 0.650 :  76%|███████▋  | 7639/10000 [08:05<02:28, 15.91it/s]\n",
      "Loss: 0.650 :  76%|███████▋  | 7641/10000 [08:05<02:26, 16.05it/s]\n",
      "Loss: 0.650 :  76%|███████▋  | 7643/10000 [08:05<02:23, 16.40it/s]\n",
      "Loss: 0.650 :  76%|███████▋  | 7645/10000 [08:06<02:21, 16.62it/s]\n",
      "Loss: 0.650 :  76%|███████▋  | 7647/10000 [08:06<02:18, 16.97it/s]\n",
      "Loss: 0.650 :  76%|███████▋  | 7649/10000 [08:06<02:17, 17.11it/s]\n",
      "Loss: 0.617 :  76%|███████▋  | 7649/10000 [08:06<02:17, 17.11it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7651/10000 [08:06<02:35, 15.14it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7653/10000 [08:06<02:29, 15.74it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7655/10000 [08:06<02:25, 16.08it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7657/10000 [08:06<02:24, 16.18it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7659/10000 [08:06<02:23, 16.32it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7661/10000 [08:07<02:23, 16.26it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7663/10000 [08:07<02:21, 16.50it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7665/10000 [08:07<02:21, 16.56it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7667/10000 [08:07<02:18, 16.86it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7669/10000 [08:07<02:20, 16.64it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7671/10000 [08:07<02:18, 16.80it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7673/10000 [08:07<02:17, 16.87it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7675/10000 [08:07<02:18, 16.77it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7677/10000 [08:08<02:20, 16.58it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7679/10000 [08:08<02:19, 16.61it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7681/10000 [08:08<02:19, 16.65it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7683/10000 [08:08<02:18, 16.78it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7685/10000 [08:08<02:19, 16.57it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7687/10000 [08:08<02:20, 16.44it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7689/10000 [08:08<02:21, 16.36it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7691/10000 [08:08<02:18, 16.65it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7693/10000 [08:09<02:17, 16.77it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7695/10000 [08:09<02:17, 16.75it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7697/10000 [08:09<02:17, 16.80it/s]\n",
      "Loss: 0.617 :  77%|███████▋  | 7699/10000 [08:09<02:15, 16.98it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7699/10000 [08:09<02:15, 16.98it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7701/10000 [08:09<02:26, 15.69it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7703/10000 [08:09<02:22, 16.15it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7705/10000 [08:09<02:17, 16.65it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7707/10000 [08:09<02:19, 16.42it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7709/10000 [08:10<02:30, 15.26it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7711/10000 [08:10<02:34, 14.78it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7713/10000 [08:10<02:34, 14.78it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7715/10000 [08:10<02:30, 15.18it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7717/10000 [08:10<02:28, 15.37it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7719/10000 [08:10<02:24, 15.75it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7721/10000 [08:10<02:24, 15.74it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7723/10000 [08:10<02:22, 16.01it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7725/10000 [08:11<02:22, 15.97it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7727/10000 [08:11<02:22, 15.99it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7729/10000 [08:11<02:21, 16.04it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7731/10000 [08:11<02:21, 16.01it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7733/10000 [08:11<02:20, 16.11it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7735/10000 [08:11<02:16, 16.55it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7737/10000 [08:11<02:16, 16.53it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7739/10000 [08:11<02:16, 16.55it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7741/10000 [08:12<02:16, 16.55it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7743/10000 [08:12<02:17, 16.44it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7745/10000 [08:12<02:16, 16.57it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7747/10000 [08:12<02:15, 16.58it/s]\n",
      "Loss: 0.633 :  77%|███████▋  | 7749/10000 [08:12<02:16, 16.52it/s]\n",
      "Loss: 0.787 :  77%|███████▋  | 7749/10000 [08:12<02:16, 16.52it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7751/10000 [08:12<02:26, 15.33it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7753/10000 [08:12<02:22, 15.78it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7755/10000 [08:12<02:18, 16.26it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7757/10000 [08:12<02:14, 16.66it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7759/10000 [08:13<02:11, 16.98it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7761/10000 [08:13<02:12, 16.89it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7763/10000 [08:13<02:12, 16.86it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7765/10000 [08:13<02:12, 16.88it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7767/10000 [08:13<02:11, 16.99it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7769/10000 [08:13<02:10, 17.12it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7771/10000 [08:13<02:10, 17.06it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7773/10000 [08:13<02:09, 17.16it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7775/10000 [08:14<02:08, 17.30it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7777/10000 [08:14<02:08, 17.35it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7779/10000 [08:14<02:11, 16.86it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7781/10000 [08:14<02:11, 16.84it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7783/10000 [08:14<02:12, 16.68it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7785/10000 [08:14<02:12, 16.69it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7787/10000 [08:14<02:10, 16.93it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7789/10000 [08:14<02:12, 16.73it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7791/10000 [08:14<02:11, 16.81it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7793/10000 [08:15<02:10, 16.94it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7795/10000 [08:15<02:16, 16.19it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7797/10000 [08:15<02:23, 15.31it/s]\n",
      "Loss: 0.787 :  78%|███████▊  | 7799/10000 [08:15<02:26, 15.05it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7799/10000 [08:15<02:26, 15.05it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7801/10000 [08:15<02:40, 13.71it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7803/10000 [08:15<02:33, 14.28it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7805/10000 [08:15<02:28, 14.80it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7807/10000 [08:16<02:24, 15.23it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7809/10000 [08:16<02:22, 15.33it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7811/10000 [08:16<02:21, 15.45it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7813/10000 [08:16<02:19, 15.66it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7815/10000 [08:16<02:15, 16.07it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7817/10000 [08:16<02:16, 16.00it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7819/10000 [08:16<02:12, 16.42it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7821/10000 [08:16<02:30, 14.43it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7823/10000 [08:17<02:24, 15.08it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7825/10000 [08:17<02:20, 15.46it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7827/10000 [08:17<02:15, 16.09it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7829/10000 [08:17<02:11, 16.47it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7831/10000 [08:17<02:11, 16.47it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7833/10000 [08:17<02:09, 16.68it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7835/10000 [08:17<02:07, 17.01it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7837/10000 [08:17<02:03, 17.46it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7839/10000 [08:18<02:05, 17.25it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7841/10000 [08:18<02:03, 17.44it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7843/10000 [08:18<02:04, 17.34it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7845/10000 [08:18<02:03, 17.40it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7847/10000 [08:18<02:02, 17.63it/s]\n",
      "Loss: 0.740 :  78%|███████▊  | 7849/10000 [08:18<02:02, 17.58it/s]\n",
      "Loss: 0.605 :  78%|███████▊  | 7849/10000 [08:18<02:02, 17.58it/s]\n",
      "Loss: 0.605 :  79%|███████▊  | 7851/10000 [08:18<02:15, 15.81it/s]\n",
      "Loss: 0.605 :  79%|███████▊  | 7853/10000 [08:18<02:15, 15.84it/s]\n",
      "Loss: 0.605 :  79%|███████▊  | 7855/10000 [08:19<02:13, 16.06it/s]\n",
      "Loss: 0.605 :  79%|███████▊  | 7857/10000 [08:19<02:11, 16.27it/s]\n",
      "Loss: 0.605 :  79%|███████▊  | 7859/10000 [08:19<02:11, 16.24it/s]\n",
      "Loss: 0.605 :  79%|███████▊  | 7861/10000 [08:19<02:10, 16.34it/s]\n",
      "Loss: 0.605 :  79%|███████▊  | 7863/10000 [08:19<02:09, 16.54it/s]\n",
      "Loss: 0.605 :  79%|███████▊  | 7865/10000 [08:19<02:07, 16.75it/s]\n",
      "Loss: 0.605 :  79%|███████▊  | 7867/10000 [08:19<02:07, 16.79it/s]\n",
      "Loss: 0.605 :  79%|███████▊  | 7869/10000 [08:19<02:09, 16.44it/s]\n",
      "Loss: 0.605 :  79%|███████▊  | 7871/10000 [08:19<02:05, 16.97it/s]\n",
      "Loss: 0.605 :  79%|███████▊  | 7873/10000 [08:20<02:05, 16.97it/s]\n",
      "Loss: 0.605 :  79%|███████▉  | 7875/10000 [08:20<02:04, 17.05it/s]\n",
      "Loss: 0.605 :  79%|███████▉  | 7877/10000 [08:20<02:03, 17.18it/s]\n",
      "Loss: 0.605 :  79%|███████▉  | 7879/10000 [08:20<02:03, 17.18it/s]\n",
      "Loss: 0.605 :  79%|███████▉  | 7881/10000 [08:20<02:11, 16.16it/s]\n",
      "Loss: 0.605 :  79%|███████▉  | 7883/10000 [08:20<02:22, 14.85it/s]\n",
      "Loss: 0.605 :  79%|███████▉  | 7885/10000 [08:20<02:23, 14.78it/s]\n",
      "Loss: 0.605 :  79%|███████▉  | 7887/10000 [08:20<02:21, 14.91it/s]\n",
      "Loss: 0.605 :  79%|███████▉  | 7889/10000 [08:21<02:20, 14.99it/s]\n",
      "Loss: 0.605 :  79%|███████▉  | 7891/10000 [08:21<02:18, 15.27it/s]\n",
      "Loss: 0.605 :  79%|███████▉  | 7893/10000 [08:21<02:16, 15.48it/s]\n",
      "Loss: 0.605 :  79%|███████▉  | 7895/10000 [08:21<02:13, 15.75it/s]\n",
      "Loss: 0.605 :  79%|███████▉  | 7897/10000 [08:21<02:15, 15.54it/s]\n",
      "Loss: 0.605 :  79%|███████▉  | 7899/10000 [08:21<02:14, 15.61it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7899/10000 [08:21<02:14, 15.61it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7901/10000 [08:21<02:27, 14.27it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7903/10000 [08:22<02:23, 14.63it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7905/10000 [08:22<02:18, 15.12it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7907/10000 [08:22<02:13, 15.67it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7909/10000 [08:22<02:13, 15.66it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7911/10000 [08:22<02:09, 16.14it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7913/10000 [08:22<02:07, 16.38it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7915/10000 [08:22<02:06, 16.44it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7917/10000 [08:22<02:04, 16.77it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7919/10000 [08:23<02:05, 16.60it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7921/10000 [08:23<02:06, 16.40it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7923/10000 [08:23<02:05, 16.59it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7925/10000 [08:23<02:03, 16.81it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7927/10000 [08:23<02:02, 16.93it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7929/10000 [08:23<02:01, 17.04it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7931/10000 [08:23<02:02, 16.93it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7933/10000 [08:23<02:02, 16.86it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7935/10000 [08:23<02:02, 16.90it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7937/10000 [08:24<02:03, 16.76it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7939/10000 [08:24<02:02, 16.76it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7941/10000 [08:24<02:02, 16.78it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7943/10000 [08:24<02:03, 16.66it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7945/10000 [08:24<02:00, 17.09it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7947/10000 [08:24<01:59, 17.22it/s]\n",
      "Loss: 0.758 :  79%|███████▉  | 7949/10000 [08:24<01:57, 17.38it/s]\n",
      "Loss: 0.579 :  79%|███████▉  | 7949/10000 [08:24<01:57, 17.38it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7951/10000 [08:24<02:10, 15.70it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7953/10000 [08:25<02:06, 16.15it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7955/10000 [08:25<02:06, 16.11it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7957/10000 [08:25<02:06, 16.17it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7959/10000 [08:25<02:02, 16.60it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7961/10000 [08:25<02:00, 16.97it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7963/10000 [08:25<02:01, 16.78it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7965/10000 [08:25<02:03, 16.53it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7967/10000 [08:25<02:12, 15.33it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7969/10000 [08:26<02:16, 14.86it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7971/10000 [08:26<02:19, 14.50it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7973/10000 [08:26<02:17, 14.71it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7975/10000 [08:26<02:14, 15.04it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7977/10000 [08:26<02:11, 15.43it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7979/10000 [08:26<02:08, 15.72it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7981/10000 [08:26<02:05, 16.13it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7983/10000 [08:26<02:05, 16.08it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7985/10000 [08:27<02:05, 16.02it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7987/10000 [08:27<02:07, 15.81it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7989/10000 [08:27<02:04, 16.21it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7991/10000 [08:27<02:01, 16.59it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7993/10000 [08:27<02:00, 16.62it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7995/10000 [08:27<01:58, 16.90it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7997/10000 [08:27<01:57, 17.09it/s]\n",
      "Loss: 0.579 :  80%|███████▉  | 7999/10000 [08:27<01:58, 16.87it/s]\n",
      "Loss: 0.668 :  80%|███████▉  | 7999/10000 [08:28<01:58, 16.87it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8001/10000 [08:28<02:13, 15.01it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8003/10000 [08:28<02:11, 15.21it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8005/10000 [08:28<02:07, 15.64it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8007/10000 [08:28<02:07, 15.69it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8009/10000 [08:28<02:05, 15.86it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8011/10000 [08:28<02:04, 15.95it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8013/10000 [08:28<02:00, 16.52it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8015/10000 [08:28<01:59, 16.67it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8017/10000 [08:29<01:59, 16.59it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8019/10000 [08:29<01:57, 16.86it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8021/10000 [08:29<01:59, 16.53it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8023/10000 [08:29<01:58, 16.65it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8025/10000 [08:29<01:57, 16.77it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8027/10000 [08:29<01:57, 16.84it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8029/10000 [08:29<01:58, 16.62it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8031/10000 [08:29<01:58, 16.59it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8033/10000 [08:30<01:57, 16.81it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8035/10000 [08:30<01:55, 17.03it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8037/10000 [08:30<01:57, 16.65it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8039/10000 [08:30<02:00, 16.32it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8041/10000 [08:30<01:55, 16.97it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8043/10000 [08:30<01:55, 17.00it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8045/10000 [08:30<01:55, 16.98it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8047/10000 [08:30<01:57, 16.58it/s]\n",
      "Loss: 0.668 :  80%|████████  | 8049/10000 [08:30<01:56, 16.72it/s]\n",
      "Loss: 0.665 :  80%|████████  | 8049/10000 [08:31<01:56, 16.72it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8051/10000 [08:31<02:10, 14.98it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8053/10000 [08:31<02:18, 14.05it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8055/10000 [08:31<02:20, 13.81it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8057/10000 [08:31<02:17, 14.10it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8059/10000 [08:31<02:13, 14.57it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8061/10000 [08:31<02:11, 14.74it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8063/10000 [08:31<02:10, 14.89it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8065/10000 [08:32<02:08, 15.03it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8067/10000 [08:32<02:05, 15.42it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8069/10000 [08:32<02:03, 15.57it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8071/10000 [08:32<02:02, 15.76it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8073/10000 [08:32<02:02, 15.72it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8075/10000 [08:32<01:59, 16.10it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8077/10000 [08:32<01:56, 16.49it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8079/10000 [08:32<01:53, 16.90it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8081/10000 [08:33<01:55, 16.68it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8083/10000 [08:33<01:51, 17.14it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8085/10000 [08:33<01:51, 17.18it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8087/10000 [08:33<01:50, 17.33it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8089/10000 [08:33<01:50, 17.36it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8091/10000 [08:33<01:50, 17.24it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8093/10000 [08:33<01:51, 17.12it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8095/10000 [08:33<01:49, 17.37it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8097/10000 [08:33<01:50, 17.26it/s]\n",
      "Loss: 0.665 :  81%|████████  | 8099/10000 [08:34<01:51, 17.01it/s]\n",
      "Loss: 0.527 :  81%|████████  | 8099/10000 [08:34<01:51, 17.01it/s]\n",
      "Loss: 0.527 :  81%|████████  | 8101/10000 [08:34<02:03, 15.35it/s]\n",
      "Loss: 0.527 :  81%|████████  | 8103/10000 [08:34<02:02, 15.54it/s]\n",
      "Loss: 0.527 :  81%|████████  | 8105/10000 [08:34<01:58, 15.99it/s]\n",
      "Loss: 0.527 :  81%|████████  | 8107/10000 [08:34<01:54, 16.53it/s]\n",
      "Loss: 0.527 :  81%|████████  | 8109/10000 [08:34<01:53, 16.70it/s]\n",
      "Loss: 0.527 :  81%|████████  | 8111/10000 [08:34<01:51, 17.00it/s]\n",
      "Loss: 0.527 :  81%|████████  | 8113/10000 [08:34<01:52, 16.85it/s]\n",
      "Loss: 0.527 :  81%|████████  | 8115/10000 [08:35<01:51, 16.87it/s]\n",
      "Loss: 0.527 :  81%|████████  | 8117/10000 [08:35<01:50, 16.98it/s]\n",
      "Loss: 0.527 :  81%|████████  | 8119/10000 [08:35<01:50, 17.06it/s]\n",
      "Loss: 0.527 :  81%|████████  | 8121/10000 [08:35<01:50, 17.01it/s]\n",
      "Loss: 0.527 :  81%|████████  | 8123/10000 [08:35<01:51, 16.89it/s]\n",
      "Loss: 0.527 :  81%|████████▏ | 8125/10000 [08:35<01:52, 16.69it/s]\n",
      "Loss: 0.527 :  81%|████████▏ | 8127/10000 [08:35<01:51, 16.79it/s]\n",
      "Loss: 0.527 :  81%|████████▏ | 8129/10000 [08:35<01:50, 16.87it/s]\n",
      "Loss: 0.527 :  81%|████████▏ | 8131/10000 [08:36<01:51, 16.80it/s]\n",
      "Loss: 0.527 :  81%|████████▏ | 8133/10000 [08:36<01:49, 16.99it/s]\n",
      "Loss: 0.527 :  81%|████████▏ | 8135/10000 [08:36<01:50, 16.82it/s]\n",
      "Loss: 0.527 :  81%|████████▏ | 8137/10000 [08:36<01:48, 17.15it/s]\n",
      "Loss: 0.527 :  81%|████████▏ | 8139/10000 [08:36<01:52, 16.51it/s]\n",
      "Loss: 0.527 :  81%|████████▏ | 8141/10000 [08:36<02:02, 15.19it/s]\n",
      "Loss: 0.527 :  81%|████████▏ | 8143/10000 [08:36<02:05, 14.75it/s]\n",
      "Loss: 0.527 :  81%|████████▏ | 8145/10000 [08:36<02:06, 14.70it/s]\n",
      "Loss: 0.527 :  81%|████████▏ | 8147/10000 [08:37<02:03, 14.95it/s]\n",
      "Loss: 0.527 :  81%|████████▏ | 8149/10000 [08:37<02:01, 15.20it/s]\n",
      "Loss: 0.616 :  81%|████████▏ | 8149/10000 [08:37<02:01, 15.20it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8151/10000 [08:37<02:09, 14.27it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8153/10000 [08:37<02:05, 14.67it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8155/10000 [08:37<02:01, 15.14it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8157/10000 [08:37<01:59, 15.41it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8159/10000 [08:37<01:55, 15.92it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8161/10000 [08:37<01:54, 16.01it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8163/10000 [08:38<01:51, 16.48it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8165/10000 [08:38<01:52, 16.32it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8167/10000 [08:38<01:53, 16.17it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8169/10000 [08:38<01:52, 16.29it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8171/10000 [08:38<01:50, 16.57it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8173/10000 [08:38<01:49, 16.65it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8175/10000 [08:38<01:50, 16.57it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8177/10000 [08:38<01:51, 16.41it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8179/10000 [08:39<01:49, 16.58it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8181/10000 [08:39<01:50, 16.41it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8183/10000 [08:39<01:50, 16.44it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8185/10000 [08:39<01:50, 16.38it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8187/10000 [08:39<01:51, 16.32it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8189/10000 [08:39<01:48, 16.65it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8191/10000 [08:39<01:46, 16.96it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8193/10000 [08:39<01:46, 16.91it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8195/10000 [08:40<01:45, 17.05it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8197/10000 [08:40<01:46, 16.86it/s]\n",
      "Loss: 0.616 :  82%|████████▏ | 8199/10000 [08:40<01:47, 16.79it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8199/10000 [08:40<01:47, 16.79it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8201/10000 [08:40<01:55, 15.51it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8203/10000 [08:40<01:53, 15.82it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8205/10000 [08:40<01:52, 15.97it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8207/10000 [08:40<01:50, 16.20it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8209/10000 [08:40<01:47, 16.63it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8211/10000 [08:41<01:47, 16.57it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8213/10000 [08:41<01:46, 16.84it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8215/10000 [08:41<01:43, 17.23it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8217/10000 [08:41<01:43, 17.26it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8219/10000 [08:41<01:45, 16.96it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8221/10000 [08:41<01:45, 16.90it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8223/10000 [08:41<01:44, 16.99it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8225/10000 [08:41<01:49, 16.28it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8227/10000 [08:41<01:58, 15.02it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8229/10000 [08:42<02:31, 11.72it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8231/10000 [08:42<02:26, 12.09it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8233/10000 [08:42<02:19, 12.67it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8235/10000 [08:42<02:14, 13.17it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8237/10000 [08:42<02:08, 13.68it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8239/10000 [08:42<02:07, 13.87it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8241/10000 [08:43<02:02, 14.39it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8243/10000 [08:43<01:59, 14.73it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8245/10000 [08:43<01:54, 15.31it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8247/10000 [08:43<01:52, 15.59it/s]\n",
      "Loss: 0.785 :  82%|████████▏ | 8249/10000 [08:43<01:51, 15.74it/s]\n",
      "Loss: 0.815 :  82%|████████▏ | 8249/10000 [08:43<01:51, 15.74it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8251/10000 [08:43<02:00, 14.49it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8253/10000 [08:43<01:56, 15.01it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8255/10000 [08:43<01:51, 15.63it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8257/10000 [08:44<01:49, 15.96it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8259/10000 [08:44<01:47, 16.17it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8261/10000 [08:44<01:45, 16.50it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8263/10000 [08:44<01:45, 16.51it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8265/10000 [08:44<01:43, 16.82it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8267/10000 [08:44<01:42, 16.86it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8269/10000 [08:44<01:43, 16.75it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8271/10000 [08:44<01:42, 16.82it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8273/10000 [08:45<01:43, 16.74it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8275/10000 [08:45<01:41, 16.95it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8277/10000 [08:45<01:38, 17.45it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8279/10000 [08:45<01:40, 17.18it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8281/10000 [08:45<01:39, 17.28it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8283/10000 [08:45<01:39, 17.25it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8285/10000 [08:45<01:39, 17.28it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8287/10000 [08:45<01:42, 16.77it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8289/10000 [08:45<01:44, 16.45it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8291/10000 [08:46<01:43, 16.44it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8293/10000 [08:46<01:43, 16.56it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8295/10000 [08:46<01:40, 17.00it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8297/10000 [08:46<01:40, 16.99it/s]\n",
      "Loss: 0.815 :  83%|████████▎ | 8299/10000 [08:46<01:40, 16.87it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8299/10000 [08:46<01:40, 16.87it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8301/10000 [08:46<01:49, 15.47it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8303/10000 [08:46<01:46, 15.88it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8305/10000 [08:46<01:45, 16.01it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8307/10000 [08:47<01:43, 16.36it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8309/10000 [08:47<01:46, 15.88it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8311/10000 [08:47<01:53, 14.92it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8313/10000 [08:47<01:55, 14.64it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8315/10000 [08:47<01:52, 14.97it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8317/10000 [08:47<01:55, 14.52it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8319/10000 [08:47<02:00, 13.98it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8321/10000 [08:48<02:00, 13.90it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8323/10000 [08:48<02:00, 13.86it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8325/10000 [08:48<01:59, 14.02it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8327/10000 [08:48<01:56, 14.33it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8329/10000 [08:48<01:52, 14.81it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8331/10000 [08:48<01:51, 15.02it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8333/10000 [08:48<01:51, 14.98it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8335/10000 [08:49<01:49, 15.19it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8337/10000 [08:49<01:48, 15.32it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8339/10000 [08:49<01:46, 15.58it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8341/10000 [08:49<01:42, 16.11it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8343/10000 [08:49<01:41, 16.32it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8345/10000 [08:49<01:38, 16.72it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8347/10000 [08:49<01:38, 16.82it/s]\n",
      "Loss: 0.515 :  83%|████████▎ | 8349/10000 [08:49<01:35, 17.29it/s]\n",
      "Loss: 0.622 :  83%|████████▎ | 8349/10000 [08:49<01:35, 17.29it/s]\n",
      "Loss: 0.622 :  84%|████████▎ | 8351/10000 [08:50<01:46, 15.45it/s]\n",
      "Loss: 0.622 :  84%|████████▎ | 8353/10000 [08:50<01:44, 15.81it/s]\n",
      "Loss: 0.622 :  84%|████████▎ | 8355/10000 [08:50<01:42, 16.08it/s]\n",
      "Loss: 0.622 :  84%|████████▎ | 8357/10000 [08:50<01:40, 16.32it/s]\n",
      "Loss: 0.622 :  84%|████████▎ | 8359/10000 [08:50<01:39, 16.42it/s]\n",
      "Loss: 0.622 :  84%|████████▎ | 8361/10000 [08:50<01:38, 16.69it/s]\n",
      "Loss: 0.622 :  84%|████████▎ | 8363/10000 [08:50<01:37, 16.84it/s]\n",
      "Loss: 0.622 :  84%|████████▎ | 8365/10000 [08:50<01:36, 16.98it/s]\n",
      "Loss: 0.622 :  84%|████████▎ | 8367/10000 [08:50<01:36, 16.85it/s]\n",
      "Loss: 0.622 :  84%|████████▎ | 8369/10000 [08:51<01:36, 16.85it/s]\n",
      "Loss: 0.622 :  84%|████████▎ | 8371/10000 [08:51<01:35, 17.14it/s]\n",
      "Loss: 0.622 :  84%|████████▎ | 8373/10000 [08:51<01:36, 16.80it/s]\n",
      "Loss: 0.622 :  84%|████████▍ | 8375/10000 [08:51<01:37, 16.67it/s]\n",
      "Loss: 0.622 :  84%|████████▍ | 8377/10000 [08:51<01:37, 16.71it/s]\n",
      "Loss: 0.622 :  84%|████████▍ | 8379/10000 [08:51<01:36, 16.74it/s]\n",
      "Loss: 0.622 :  84%|████████▍ | 8381/10000 [08:51<01:35, 17.01it/s]\n",
      "Loss: 0.622 :  84%|████████▍ | 8383/10000 [08:51<01:34, 17.08it/s]\n",
      "Loss: 0.622 :  84%|████████▍ | 8385/10000 [08:52<01:37, 16.58it/s]\n",
      "Loss: 0.622 :  84%|████████▍ | 8387/10000 [08:52<01:37, 16.59it/s]\n",
      "Loss: 0.622 :  84%|████████▍ | 8389/10000 [08:52<01:34, 16.98it/s]\n",
      "Loss: 0.622 :  84%|████████▍ | 8391/10000 [08:52<01:35, 16.81it/s]\n",
      "Loss: 0.622 :  84%|████████▍ | 8393/10000 [08:52<01:38, 16.34it/s]\n",
      "Loss: 0.622 :  84%|████████▍ | 8395/10000 [08:52<01:46, 15.04it/s]\n",
      "Loss: 0.622 :  84%|████████▍ | 8397/10000 [08:52<01:49, 14.66it/s]\n",
      "Loss: 0.622 :  84%|████████▍ | 8399/10000 [08:52<01:47, 14.93it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8399/10000 [08:53<01:47, 14.93it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8401/10000 [08:53<01:54, 13.92it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8403/10000 [08:53<01:51, 14.33it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8405/10000 [08:53<01:47, 14.81it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8407/10000 [08:53<01:45, 15.07it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8409/10000 [08:53<01:43, 15.39it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8411/10000 [08:53<01:40, 15.77it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8413/10000 [08:53<01:38, 16.06it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8415/10000 [08:53<01:38, 16.10it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8417/10000 [08:54<01:36, 16.38it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8419/10000 [08:54<01:35, 16.56it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8421/10000 [08:54<01:34, 16.67it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8423/10000 [08:54<01:35, 16.53it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8425/10000 [08:54<01:34, 16.66it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8427/10000 [08:54<01:34, 16.71it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8429/10000 [08:54<01:33, 16.80it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8431/10000 [08:54<01:32, 16.90it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8433/10000 [08:55<01:33, 16.75it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8435/10000 [08:55<01:33, 16.79it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8437/10000 [08:55<01:33, 16.78it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8439/10000 [08:55<01:32, 16.90it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8441/10000 [08:55<01:33, 16.75it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8443/10000 [08:55<01:31, 17.05it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8445/10000 [08:55<01:31, 16.95it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8447/10000 [08:55<01:31, 17.02it/s]\n",
      "Loss: 0.760 :  84%|████████▍ | 8449/10000 [08:55<01:32, 16.69it/s]\n",
      "Loss: 0.487 :  84%|████████▍ | 8449/10000 [08:56<01:32, 16.69it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8451/10000 [08:56<01:42, 15.07it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8453/10000 [08:56<01:40, 15.43it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8455/10000 [08:56<01:37, 15.84it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8457/10000 [08:56<01:35, 16.23it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8459/10000 [08:56<01:31, 16.81it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8461/10000 [08:56<01:44, 14.79it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8463/10000 [08:56<01:38, 15.61it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8465/10000 [08:57<01:37, 15.79it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8467/10000 [08:57<01:35, 16.12it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8469/10000 [08:57<01:33, 16.43it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8471/10000 [08:57<01:32, 16.58it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8473/10000 [08:57<01:32, 16.58it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8475/10000 [08:57<01:31, 16.71it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8477/10000 [08:57<01:31, 16.70it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8479/10000 [08:57<01:35, 15.99it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8481/10000 [08:58<01:40, 15.15it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8483/10000 [08:58<01:40, 15.12it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8485/10000 [08:58<01:40, 15.02it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8487/10000 [08:58<01:39, 15.14it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8489/10000 [08:58<01:38, 15.38it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8491/10000 [08:58<01:36, 15.58it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8493/10000 [08:58<01:35, 15.80it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8495/10000 [08:58<01:35, 15.82it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8497/10000 [08:59<01:33, 16.06it/s]\n",
      "Loss: 0.487 :  85%|████████▍ | 8499/10000 [08:59<01:31, 16.48it/s]\n",
      "Loss: 0.559 :  85%|████████▍ | 8499/10000 [08:59<01:31, 16.48it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8501/10000 [08:59<01:40, 14.97it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8503/10000 [08:59<01:37, 15.36it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8505/10000 [08:59<01:37, 15.35it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8507/10000 [08:59<01:34, 15.74it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8509/10000 [08:59<01:33, 15.98it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8511/10000 [08:59<01:32, 16.06it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8513/10000 [09:00<01:29, 16.61it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8515/10000 [09:00<01:29, 16.53it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8517/10000 [09:00<01:28, 16.71it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8519/10000 [09:00<01:28, 16.73it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8521/10000 [09:00<01:27, 16.82it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8523/10000 [09:00<01:27, 16.91it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8525/10000 [09:00<01:26, 17.02it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8527/10000 [09:00<01:26, 17.02it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8529/10000 [09:00<01:25, 17.13it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8531/10000 [09:01<01:26, 16.98it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8533/10000 [09:01<01:25, 17.17it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8535/10000 [09:01<01:25, 17.17it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8537/10000 [09:01<01:25, 17.06it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8539/10000 [09:01<01:24, 17.24it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8541/10000 [09:01<01:25, 17.09it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8543/10000 [09:01<01:24, 17.19it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8545/10000 [09:01<01:24, 17.14it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8547/10000 [09:02<01:23, 17.30it/s]\n",
      "Loss: 0.559 :  85%|████████▌ | 8549/10000 [09:02<01:24, 17.08it/s]\n",
      "Loss: 0.644 :  85%|████████▌ | 8549/10000 [09:02<01:24, 17.08it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8551/10000 [09:02<01:33, 15.46it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8553/10000 [09:02<01:31, 15.78it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8555/10000 [09:02<01:28, 16.35it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8557/10000 [09:02<01:27, 16.48it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8559/10000 [09:02<01:27, 16.55it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8561/10000 [09:02<01:26, 16.70it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8563/10000 [09:03<01:25, 16.90it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8565/10000 [09:03<01:27, 16.39it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8567/10000 [09:03<01:37, 14.67it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8569/10000 [09:03<01:45, 13.60it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8571/10000 [09:03<01:46, 13.43it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8573/10000 [09:03<01:51, 12.80it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8575/10000 [09:03<01:56, 12.28it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8577/10000 [09:04<02:00, 11.85it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8579/10000 [09:04<01:57, 12.07it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8581/10000 [09:04<01:54, 12.39it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8583/10000 [09:04<01:49, 12.92it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8585/10000 [09:04<01:45, 13.45it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8587/10000 [09:04<01:44, 13.48it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8589/10000 [09:05<01:43, 13.67it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8591/10000 [09:05<01:41, 13.87it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8593/10000 [09:05<01:38, 14.22it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8595/10000 [09:05<01:37, 14.41it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8597/10000 [09:05<01:35, 14.66it/s]\n",
      "Loss: 0.644 :  86%|████████▌ | 8599/10000 [09:05<01:35, 14.67it/s]\n",
      "Loss: 0.662 :  86%|████████▌ | 8599/10000 [09:05<01:35, 14.67it/s]\n",
      "Loss: 0.662 :  86%|████████▌ | 8601/10000 [09:05<01:40, 13.85it/s]\n",
      "Loss: 0.662 :  86%|████████▌ | 8603/10000 [09:06<01:36, 14.43it/s]\n",
      "Loss: 0.662 :  86%|████████▌ | 8605/10000 [09:06<01:33, 15.00it/s]\n",
      "Loss: 0.662 :  86%|████████▌ | 8607/10000 [09:06<01:32, 15.13it/s]\n",
      "Loss: 0.662 :  86%|████████▌ | 8609/10000 [09:06<01:27, 15.85it/s]\n",
      "Loss: 0.662 :  86%|████████▌ | 8611/10000 [09:06<01:34, 14.73it/s]\n",
      "Loss: 0.662 :  86%|████████▌ | 8613/10000 [09:06<01:36, 14.31it/s]\n",
      "Loss: 0.662 :  86%|████████▌ | 8615/10000 [09:06<01:35, 14.53it/s]\n",
      "Loss: 0.662 :  86%|████████▌ | 8617/10000 [09:06<01:35, 14.44it/s]\n",
      "Loss: 0.662 :  86%|████████▌ | 8619/10000 [09:07<01:35, 14.48it/s]\n",
      "Loss: 0.662 :  86%|████████▌ | 8621/10000 [09:07<01:36, 14.33it/s]\n",
      "Loss: 0.662 :  86%|████████▌ | 8623/10000 [09:07<01:34, 14.57it/s]\n",
      "Loss: 0.662 :  86%|████████▋ | 8625/10000 [09:07<01:31, 15.01it/s]\n",
      "Loss: 0.662 :  86%|████████▋ | 8627/10000 [09:07<01:29, 15.37it/s]\n",
      "Loss: 0.662 :  86%|████████▋ | 8629/10000 [09:07<01:28, 15.50it/s]\n",
      "Loss: 0.662 :  86%|████████▋ | 8631/10000 [09:07<01:27, 15.69it/s]\n",
      "Loss: 0.662 :  86%|████████▋ | 8633/10000 [09:07<01:26, 15.73it/s]\n",
      "Loss: 0.662 :  86%|████████▋ | 8635/10000 [09:08<01:25, 16.01it/s]\n",
      "Loss: 0.662 :  86%|████████▋ | 8637/10000 [09:08<01:24, 16.18it/s]\n",
      "Loss: 0.662 :  86%|████████▋ | 8639/10000 [09:08<01:23, 16.23it/s]\n",
      "Loss: 0.662 :  86%|████████▋ | 8641/10000 [09:08<01:27, 15.54it/s]\n",
      "Loss: 0.662 :  86%|████████▋ | 8643/10000 [09:08<01:31, 14.76it/s]\n",
      "Loss: 0.662 :  86%|████████▋ | 8645/10000 [09:08<01:33, 14.45it/s]\n",
      "Loss: 0.662 :  86%|████████▋ | 8647/10000 [09:08<01:31, 14.73it/s]\n",
      "Loss: 0.662 :  86%|████████▋ | 8649/10000 [09:09<01:28, 15.26it/s]\n",
      "Loss: 0.627 :  86%|████████▋ | 8649/10000 [09:09<01:28, 15.26it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8651/10000 [09:09<01:38, 13.64it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8653/10000 [09:09<01:36, 13.95it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8655/10000 [09:09<01:32, 14.47it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8657/10000 [09:09<01:29, 14.96it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8659/10000 [09:09<01:27, 15.26it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8661/10000 [09:09<01:26, 15.48it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8663/10000 [09:09<01:24, 15.81it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8665/10000 [09:10<01:24, 15.84it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8667/10000 [09:10<01:26, 15.42it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8669/10000 [09:10<01:26, 15.30it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8671/10000 [09:10<01:29, 14.89it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8673/10000 [09:10<01:30, 14.74it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8675/10000 [09:10<01:31, 14.49it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8677/10000 [09:10<01:31, 14.39it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8679/10000 [09:11<01:30, 14.59it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8681/10000 [09:11<01:30, 14.63it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8683/10000 [09:11<01:28, 14.84it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8685/10000 [09:11<01:26, 15.12it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8687/10000 [09:11<01:25, 15.42it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8689/10000 [09:11<01:23, 15.69it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8691/10000 [09:11<01:22, 15.81it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8693/10000 [09:11<01:21, 16.06it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8695/10000 [09:12<01:22, 15.91it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8697/10000 [09:12<01:22, 15.74it/s]\n",
      "Loss: 0.627 :  87%|████████▋ | 8699/10000 [09:12<01:21, 15.99it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8699/10000 [09:12<01:21, 15.99it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8701/10000 [09:12<01:28, 14.62it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8703/10000 [09:12<01:26, 14.93it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8705/10000 [09:12<01:23, 15.48it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8707/10000 [09:12<01:21, 15.81it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8709/10000 [09:12<01:19, 16.22it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8711/10000 [09:13<01:18, 16.46it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8713/10000 [09:13<01:15, 16.97it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8715/10000 [09:13<01:17, 16.69it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8717/10000 [09:13<01:16, 16.70it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8719/10000 [09:13<01:16, 16.68it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8721/10000 [09:13<01:16, 16.76it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8723/10000 [09:13<01:18, 16.18it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8725/10000 [09:13<01:24, 15.10it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8727/10000 [09:14<01:26, 14.72it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8729/10000 [09:14<01:25, 14.87it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8731/10000 [09:14<01:23, 15.23it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8733/10000 [09:14<01:22, 15.37it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8735/10000 [09:14<01:20, 15.79it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8737/10000 [09:14<01:19, 15.90it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8739/10000 [09:14<01:17, 16.17it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8741/10000 [09:14<01:17, 16.33it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8743/10000 [09:15<01:16, 16.34it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8745/10000 [09:15<01:16, 16.35it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8747/10000 [09:15<01:16, 16.46it/s]\n",
      "Loss: 0.522 :  87%|████████▋ | 8749/10000 [09:15<01:14, 16.79it/s]\n",
      "Loss: 0.729 :  87%|████████▋ | 8749/10000 [09:15<01:14, 16.79it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8751/10000 [09:15<01:21, 15.29it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8753/10000 [09:15<01:19, 15.74it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8755/10000 [09:15<01:19, 15.57it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8757/10000 [09:15<01:17, 15.99it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8759/10000 [09:16<01:17, 16.06it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8761/10000 [09:16<01:16, 16.26it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8763/10000 [09:16<01:16, 16.16it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8765/10000 [09:16<01:15, 16.46it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8767/10000 [09:16<01:13, 16.69it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8769/10000 [09:16<01:14, 16.59it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8771/10000 [09:16<01:12, 16.88it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8773/10000 [09:16<01:12, 16.94it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8775/10000 [09:17<01:12, 16.93it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8777/10000 [09:17<01:11, 17.20it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8779/10000 [09:17<01:12, 16.88it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8781/10000 [09:17<01:11, 17.14it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8783/10000 [09:17<01:10, 17.22it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8785/10000 [09:17<01:11, 17.05it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8787/10000 [09:17<01:10, 17.15it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8789/10000 [09:17<01:11, 16.98it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8791/10000 [09:17<01:09, 17.28it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8793/10000 [09:18<01:09, 17.41it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8795/10000 [09:18<01:10, 17.16it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8797/10000 [09:18<01:10, 17.15it/s]\n",
      "Loss: 0.729 :  88%|████████▊ | 8799/10000 [09:18<01:10, 16.99it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8799/10000 [09:18<01:10, 16.99it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8801/10000 [09:18<01:17, 15.44it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8803/10000 [09:18<01:14, 15.98it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8805/10000 [09:18<01:13, 16.20it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8807/10000 [09:18<01:12, 16.39it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8809/10000 [09:19<01:12, 16.40it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8811/10000 [09:19<01:18, 15.22it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8813/10000 [09:19<01:20, 14.71it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8815/10000 [09:19<01:19, 14.93it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8817/10000 [09:19<01:19, 14.95it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8819/10000 [09:19<01:17, 15.19it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8821/10000 [09:19<01:16, 15.43it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8823/10000 [09:20<01:14, 15.79it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8825/10000 [09:20<01:13, 15.97it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8827/10000 [09:20<01:14, 15.66it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8829/10000 [09:20<01:13, 16.04it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8831/10000 [09:20<01:12, 16.24it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8833/10000 [09:20<01:12, 16.15it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8835/10000 [09:20<01:09, 16.70it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8837/10000 [09:20<01:08, 16.92it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8839/10000 [09:20<01:09, 16.70it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8841/10000 [09:21<01:08, 16.86it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8843/10000 [09:21<01:08, 16.78it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8845/10000 [09:21<01:07, 17.01it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8847/10000 [09:21<01:09, 16.64it/s]\n",
      "Loss: 0.663 :  88%|████████▊ | 8849/10000 [09:21<01:08, 16.82it/s]\n",
      "Loss: 0.701 :  88%|████████▊ | 8849/10000 [09:21<01:08, 16.82it/s]\n",
      "Loss: 0.701 :  89%|████████▊ | 8851/10000 [09:21<01:14, 15.40it/s]\n",
      "Loss: 0.701 :  89%|████████▊ | 8853/10000 [09:21<01:13, 15.57it/s]\n",
      "Loss: 0.701 :  89%|████████▊ | 8855/10000 [09:21<01:11, 15.94it/s]\n",
      "Loss: 0.701 :  89%|████████▊ | 8857/10000 [09:22<01:11, 16.07it/s]\n",
      "Loss: 0.701 :  89%|████████▊ | 8859/10000 [09:22<01:08, 16.56it/s]\n",
      "Loss: 0.701 :  89%|████████▊ | 8861/10000 [09:22<01:08, 16.69it/s]\n",
      "Loss: 0.701 :  89%|████████▊ | 8863/10000 [09:22<01:08, 16.59it/s]\n",
      "Loss: 0.701 :  89%|████████▊ | 8865/10000 [09:22<01:07, 16.77it/s]\n",
      "Loss: 0.701 :  89%|████████▊ | 8867/10000 [09:22<01:07, 16.78it/s]\n",
      "Loss: 0.701 :  89%|████████▊ | 8869/10000 [09:22<01:07, 16.88it/s]\n",
      "Loss: 0.701 :  89%|████████▊ | 8871/10000 [09:22<01:06, 16.88it/s]\n",
      "Loss: 0.701 :  89%|████████▊ | 8873/10000 [09:23<01:06, 17.05it/s]\n",
      "Loss: 0.701 :  89%|████████▉ | 8875/10000 [09:23<01:05, 17.24it/s]\n",
      "Loss: 0.701 :  89%|████████▉ | 8877/10000 [09:23<01:05, 17.11it/s]\n",
      "Loss: 0.701 :  89%|████████▉ | 8879/10000 [09:23<01:05, 17.00it/s]\n",
      "Loss: 0.701 :  89%|████████▉ | 8881/10000 [09:23<01:05, 17.14it/s]\n",
      "Loss: 0.701 :  89%|████████▉ | 8883/10000 [09:23<01:04, 17.26it/s]\n",
      "Loss: 0.701 :  89%|████████▉ | 8885/10000 [09:23<01:04, 17.38it/s]\n",
      "Loss: 0.701 :  89%|████████▉ | 8887/10000 [09:23<01:04, 17.29it/s]\n",
      "Loss: 0.701 :  89%|████████▉ | 8889/10000 [09:23<01:04, 17.23it/s]\n",
      "Loss: 0.701 :  89%|████████▉ | 8891/10000 [09:24<01:04, 17.23it/s]\n",
      "Loss: 0.701 :  89%|████████▉ | 8893/10000 [09:24<01:03, 17.34it/s]\n",
      "Loss: 0.701 :  89%|████████▉ | 8895/10000 [09:24<01:04, 17.17it/s]\n",
      "Loss: 0.701 :  89%|████████▉ | 8897/10000 [09:24<01:07, 16.34it/s]\n",
      "Loss: 0.701 :  89%|████████▉ | 8899/10000 [09:24<01:12, 15.18it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8899/10000 [09:24<01:12, 15.18it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8901/10000 [09:24<01:21, 13.53it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8903/10000 [09:24<01:17, 14.20it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8905/10000 [09:25<01:16, 14.37it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8907/10000 [09:25<01:14, 14.60it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8909/10000 [09:25<01:11, 15.21it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8911/10000 [09:25<01:10, 15.42it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8913/10000 [09:25<01:09, 15.73it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8915/10000 [09:25<01:07, 16.01it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8917/10000 [09:25<01:07, 16.14it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8919/10000 [09:25<01:07, 16.12it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8921/10000 [09:26<01:07, 15.99it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8923/10000 [09:26<01:07, 16.04it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8925/10000 [09:26<01:05, 16.37it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8927/10000 [09:26<01:04, 16.55it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8929/10000 [09:26<01:03, 16.88it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8931/10000 [09:26<01:04, 16.57it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8933/10000 [09:26<01:03, 16.72it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8935/10000 [09:26<01:03, 16.84it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8937/10000 [09:26<01:03, 16.70it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8939/10000 [09:27<01:03, 16.84it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8941/10000 [09:27<01:02, 16.81it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8943/10000 [09:27<01:03, 16.65it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8945/10000 [09:27<01:03, 16.61it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8947/10000 [09:27<01:03, 16.48it/s]\n",
      "Loss: 0.504 :  89%|████████▉ | 8949/10000 [09:27<01:03, 16.52it/s]\n",
      "Loss: 0.637 :  89%|████████▉ | 8949/10000 [09:27<01:03, 16.52it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8951/10000 [09:27<01:10, 14.99it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8953/10000 [09:27<01:07, 15.60it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8955/10000 [09:28<01:04, 16.08it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8957/10000 [09:28<01:03, 16.36it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8959/10000 [09:28<01:03, 16.48it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8961/10000 [09:28<01:04, 16.20it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8963/10000 [09:28<01:03, 16.33it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8965/10000 [09:28<01:03, 16.41it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8967/10000 [09:28<01:02, 16.42it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8969/10000 [09:28<01:02, 16.52it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8971/10000 [09:29<01:01, 16.62it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8973/10000 [09:29<01:00, 16.84it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8975/10000 [09:29<01:00, 16.91it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8977/10000 [09:29<00:59, 17.13it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8979/10000 [09:29<00:58, 17.32it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8981/10000 [09:29<00:58, 17.44it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8983/10000 [09:29<01:01, 16.63it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8985/10000 [09:29<01:05, 15.44it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8987/10000 [09:30<01:07, 14.95it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8989/10000 [09:30<01:07, 15.04it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8991/10000 [09:30<01:06, 15.26it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8993/10000 [09:30<01:04, 15.67it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8995/10000 [09:30<01:03, 15.82it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8997/10000 [09:30<01:02, 16.08it/s]\n",
      "Loss: 0.637 :  90%|████████▉ | 8999/10000 [09:30<01:01, 16.18it/s]\n",
      "Loss: 0.686 :  90%|████████▉ | 8999/10000 [09:30<01:01, 16.18it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9001/10000 [09:30<01:07, 14.72it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9003/10000 [09:31<01:05, 15.32it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9005/10000 [09:31<01:02, 15.83it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9007/10000 [09:31<01:02, 15.86it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9009/10000 [09:31<01:00, 16.28it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9011/10000 [09:31<01:01, 16.13it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9013/10000 [09:31<01:01, 16.09it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9015/10000 [09:31<01:01, 16.13it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9017/10000 [09:31<01:00, 16.13it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9019/10000 [09:32<01:01, 16.07it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9021/10000 [09:32<01:00, 16.18it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9023/10000 [09:32<01:00, 16.24it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9025/10000 [09:32<00:59, 16.42it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9027/10000 [09:32<00:58, 16.69it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9029/10000 [09:32<00:57, 16.95it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9031/10000 [09:32<00:57, 16.81it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9033/10000 [09:32<00:57, 16.73it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9035/10000 [09:33<00:57, 16.70it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9037/10000 [09:33<00:57, 16.87it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9039/10000 [09:33<00:56, 16.96it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9041/10000 [09:33<00:55, 17.32it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9043/10000 [09:33<00:55, 17.16it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9045/10000 [09:33<00:56, 16.96it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9047/10000 [09:33<00:56, 16.75it/s]\n",
      "Loss: 0.686 :  90%|█████████ | 9049/10000 [09:33<00:57, 16.60it/s]\n",
      "Loss: 0.490 :  90%|█████████ | 9049/10000 [09:33<00:57, 16.60it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9051/10000 [09:34<01:02, 15.14it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9053/10000 [09:34<00:59, 15.95it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9055/10000 [09:34<00:57, 16.34it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9057/10000 [09:34<00:56, 16.72it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9059/10000 [09:34<00:56, 16.57it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9061/10000 [09:34<00:56, 16.70it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9063/10000 [09:34<00:55, 16.84it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9065/10000 [09:34<00:54, 17.19it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9067/10000 [09:34<00:54, 17.16it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9069/10000 [09:35<00:55, 16.70it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9071/10000 [09:35<01:01, 15.14it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9073/10000 [09:35<01:03, 14.71it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9075/10000 [09:35<01:02, 14.91it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9077/10000 [09:35<01:01, 15.11it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9079/10000 [09:35<01:00, 15.22it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9081/10000 [09:35<01:00, 15.26it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9083/10000 [09:36<00:59, 15.38it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9085/10000 [09:36<00:59, 15.46it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9087/10000 [09:36<00:58, 15.51it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9089/10000 [09:36<00:57, 15.80it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9091/10000 [09:36<00:57, 15.90it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9093/10000 [09:36<00:55, 16.23it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9095/10000 [09:36<00:54, 16.52it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9097/10000 [09:36<00:56, 16.11it/s]\n",
      "Loss: 0.490 :  91%|█████████ | 9099/10000 [09:36<00:54, 16.57it/s]\n",
      "Loss: 0.644 :  91%|█████████ | 9099/10000 [09:37<00:54, 16.57it/s]\n",
      "Loss: 0.644 :  91%|█████████ | 9101/10000 [09:37<01:08, 13.20it/s]\n",
      "Loss: 0.644 :  91%|█████████ | 9103/10000 [09:37<01:03, 14.09it/s]\n",
      "Loss: 0.644 :  91%|█████████ | 9105/10000 [09:37<01:01, 14.64it/s]\n",
      "Loss: 0.644 :  91%|█████████ | 9107/10000 [09:37<00:58, 15.14it/s]\n",
      "Loss: 0.644 :  91%|█████████ | 9109/10000 [09:37<00:56, 15.66it/s]\n",
      "Loss: 0.644 :  91%|█████████ | 9111/10000 [09:37<00:55, 16.08it/s]\n",
      "Loss: 0.644 :  91%|█████████ | 9113/10000 [09:37<00:53, 16.48it/s]\n",
      "Loss: 0.644 :  91%|█████████ | 9115/10000 [09:38<00:53, 16.62it/s]\n",
      "Loss: 0.644 :  91%|█████████ | 9117/10000 [09:38<00:52, 16.95it/s]\n",
      "Loss: 0.644 :  91%|█████████ | 9119/10000 [09:38<00:51, 17.08it/s]\n",
      "Loss: 0.644 :  91%|█████████ | 9121/10000 [09:38<00:52, 16.82it/s]\n",
      "Loss: 0.644 :  91%|█████████ | 9123/10000 [09:38<00:52, 16.74it/s]\n",
      "Loss: 0.644 :  91%|█████████▏| 9125/10000 [09:38<00:52, 16.67it/s]\n",
      "Loss: 0.644 :  91%|█████████▏| 9127/10000 [09:38<00:51, 16.79it/s]\n",
      "Loss: 0.644 :  91%|█████████▏| 9129/10000 [09:38<00:51, 16.81it/s]\n",
      "Loss: 0.644 :  91%|█████████▏| 9131/10000 [09:38<00:51, 17.03it/s]\n",
      "Loss: 0.644 :  91%|█████████▏| 9133/10000 [09:39<00:51, 16.87it/s]\n",
      "Loss: 0.644 :  91%|█████████▏| 9135/10000 [09:39<00:50, 17.06it/s]\n",
      "Loss: 0.644 :  91%|█████████▏| 9137/10000 [09:39<00:51, 16.85it/s]\n",
      "Loss: 0.644 :  91%|█████████▏| 9139/10000 [09:39<00:50, 16.96it/s]\n",
      "Loss: 0.644 :  91%|█████████▏| 9141/10000 [09:39<00:50, 16.96it/s]\n",
      "Loss: 0.644 :  91%|█████████▏| 9143/10000 [09:39<00:50, 16.94it/s]\n",
      "Loss: 0.644 :  91%|█████████▏| 9145/10000 [09:39<00:50, 16.95it/s]\n",
      "Loss: 0.644 :  91%|█████████▏| 9147/10000 [09:39<00:50, 17.01it/s]\n",
      "Loss: 0.644 :  91%|█████████▏| 9149/10000 [09:40<00:49, 17.15it/s]\n",
      "Loss: 0.717 :  91%|█████████▏| 9149/10000 [09:40<00:49, 17.15it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9151/10000 [09:40<00:55, 15.43it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9153/10000 [09:40<00:53, 15.75it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9155/10000 [09:40<00:57, 14.58it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9157/10000 [09:40<00:58, 14.36it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9159/10000 [09:40<00:57, 14.66it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9161/10000 [09:40<00:55, 15.17it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9163/10000 [09:41<00:54, 15.48it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9165/10000 [09:41<00:54, 15.31it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9167/10000 [09:41<00:53, 15.52it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9169/10000 [09:41<00:52, 15.70it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9171/10000 [09:41<00:51, 16.04it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9173/10000 [09:41<00:51, 16.02it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9175/10000 [09:41<00:50, 16.30it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9177/10000 [09:41<00:50, 16.38it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9179/10000 [09:42<00:50, 16.29it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9181/10000 [09:42<00:50, 16.18it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9183/10000 [09:42<00:50, 16.22it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9185/10000 [09:42<00:49, 16.52it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9187/10000 [09:42<00:49, 16.51it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9189/10000 [09:42<00:48, 16.77it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9191/10000 [09:42<00:47, 17.00it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9193/10000 [09:42<00:48, 16.76it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9195/10000 [09:42<00:47, 16.88it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9197/10000 [09:43<00:46, 17.10it/s]\n",
      "Loss: 0.717 :  92%|█████████▏| 9199/10000 [09:43<00:46, 17.13it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9199/10000 [09:43<00:46, 17.13it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9201/10000 [09:43<00:52, 15.19it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9203/10000 [09:43<00:50, 15.72it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9205/10000 [09:43<00:49, 16.03it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9207/10000 [09:43<00:48, 16.34it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9209/10000 [09:43<00:47, 16.67it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9211/10000 [09:43<00:46, 17.13it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9213/10000 [09:44<00:45, 17.29it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9215/10000 [09:44<00:45, 17.28it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9217/10000 [09:44<00:45, 17.12it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9219/10000 [09:44<00:45, 17.33it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9221/10000 [09:44<00:45, 17.24it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9223/10000 [09:44<00:45, 17.04it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9225/10000 [09:44<00:45, 17.12it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9227/10000 [09:44<00:44, 17.21it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9229/10000 [09:44<00:43, 17.62it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9231/10000 [09:45<00:43, 17.68it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9233/10000 [09:45<00:43, 17.51it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9235/10000 [09:45<00:43, 17.44it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9237/10000 [09:45<00:44, 17.04it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9239/10000 [09:45<00:44, 17.02it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9241/10000 [09:45<00:44, 16.91it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9243/10000 [09:45<00:49, 15.37it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9245/10000 [09:45<00:51, 14.69it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9247/10000 [09:46<00:50, 14.95it/s]\n",
      "Loss: 0.670 :  92%|█████████▏| 9249/10000 [09:46<00:49, 15.13it/s]\n",
      "Loss: 0.573 :  92%|█████████▏| 9249/10000 [09:46<00:49, 15.13it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9251/10000 [09:46<00:55, 13.52it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9253/10000 [09:46<00:52, 14.19it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9255/10000 [09:46<00:50, 14.77it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9257/10000 [09:46<00:49, 15.08it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9259/10000 [09:46<00:49, 15.11it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9261/10000 [09:47<00:47, 15.49it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9263/10000 [09:47<00:46, 15.77it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9265/10000 [09:47<00:45, 16.25it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9267/10000 [09:47<00:44, 16.45it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9269/10000 [09:47<00:44, 16.55it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9271/10000 [09:47<00:43, 16.57it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9273/10000 [09:47<00:43, 16.76it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9275/10000 [09:47<00:42, 16.86it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9277/10000 [09:47<00:42, 16.87it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9279/10000 [09:48<00:43, 16.75it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9281/10000 [09:48<00:42, 16.76it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9283/10000 [09:48<00:43, 16.60it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9285/10000 [09:48<00:42, 16.64it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9287/10000 [09:48<00:43, 16.48it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9289/10000 [09:48<00:42, 16.70it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9291/10000 [09:48<00:42, 16.64it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9293/10000 [09:48<00:42, 16.65it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9295/10000 [09:49<00:42, 16.71it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9297/10000 [09:49<00:41, 16.96it/s]\n",
      "Loss: 0.573 :  93%|█████████▎| 9299/10000 [09:49<00:41, 16.89it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9299/10000 [09:49<00:41, 16.89it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9301/10000 [09:49<00:46, 15.11it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9303/10000 [09:49<00:44, 15.70it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9305/10000 [09:49<00:43, 15.97it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9307/10000 [09:49<00:42, 16.46it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9309/10000 [09:49<00:40, 16.86it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9311/10000 [09:50<00:40, 16.91it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9313/10000 [09:50<00:41, 16.60it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9315/10000 [09:50<00:41, 16.52it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9317/10000 [09:50<00:41, 16.63it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9319/10000 [09:50<00:39, 17.03it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9321/10000 [09:50<00:39, 17.05it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9323/10000 [09:50<00:39, 17.13it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9325/10000 [09:50<00:40, 16.83it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9327/10000 [09:51<00:40, 16.68it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9329/10000 [09:51<00:43, 15.30it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9331/10000 [09:51<00:45, 14.79it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9333/10000 [09:51<00:45, 14.76it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9335/10000 [09:51<00:44, 14.94it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9337/10000 [09:51<00:43, 15.14it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9339/10000 [09:51<00:43, 15.29it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9341/10000 [09:51<00:42, 15.41it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9343/10000 [09:52<00:42, 15.46it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9345/10000 [09:52<00:41, 15.94it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9347/10000 [09:52<00:40, 15.96it/s]\n",
      "Loss: 0.691 :  93%|█████████▎| 9349/10000 [09:52<00:41, 15.84it/s]\n",
      "Loss: 0.553 :  93%|█████████▎| 9349/10000 [09:52<00:41, 15.84it/s]\n",
      "Loss: 0.553 :  94%|█████████▎| 9351/10000 [09:52<00:44, 14.70it/s]\n",
      "Loss: 0.553 :  94%|█████████▎| 9353/10000 [09:52<00:42, 15.27it/s]\n",
      "Loss: 0.553 :  94%|█████████▎| 9355/10000 [09:52<00:40, 15.86it/s]\n",
      "Loss: 0.553 :  94%|█████████▎| 9357/10000 [09:52<00:39, 16.27it/s]\n",
      "Loss: 0.553 :  94%|█████████▎| 9359/10000 [09:53<00:39, 16.20it/s]\n",
      "Loss: 0.553 :  94%|█████████▎| 9361/10000 [09:53<00:38, 16.42it/s]\n",
      "Loss: 0.553 :  94%|█████████▎| 9363/10000 [09:53<00:38, 16.48it/s]\n",
      "Loss: 0.553 :  94%|█████████▎| 9365/10000 [09:53<00:37, 16.75it/s]\n",
      "Loss: 0.553 :  94%|█████████▎| 9367/10000 [09:53<00:38, 16.48it/s]\n",
      "Loss: 0.553 :  94%|█████████▎| 9369/10000 [09:53<00:37, 16.67it/s]\n",
      "Loss: 0.553 :  94%|█████████▎| 9371/10000 [09:53<00:37, 16.75it/s]\n",
      "Loss: 0.553 :  94%|█████████▎| 9373/10000 [09:53<00:37, 16.94it/s]\n",
      "Loss: 0.553 :  94%|█████████▍| 9375/10000 [09:54<00:37, 16.81it/s]\n",
      "Loss: 0.553 :  94%|█████████▍| 9377/10000 [09:54<00:36, 17.09it/s]\n",
      "Loss: 0.553 :  94%|█████████▍| 9379/10000 [09:54<00:36, 17.05it/s]\n",
      "Loss: 0.553 :  94%|█████████▍| 9381/10000 [09:54<00:37, 16.68it/s]\n",
      "Loss: 0.553 :  94%|█████████▍| 9383/10000 [09:54<00:36, 16.70it/s]\n",
      "Loss: 0.553 :  94%|█████████▍| 9385/10000 [09:54<00:37, 16.51it/s]\n",
      "Loss: 0.553 :  94%|█████████▍| 9387/10000 [09:54<00:36, 16.86it/s]\n",
      "Loss: 0.553 :  94%|█████████▍| 9389/10000 [09:54<00:35, 17.15it/s]\n",
      "Loss: 0.553 :  94%|█████████▍| 9391/10000 [09:54<00:35, 17.07it/s]\n",
      "Loss: 0.553 :  94%|█████████▍| 9393/10000 [09:55<00:35, 17.20it/s]\n",
      "Loss: 0.553 :  94%|█████████▍| 9395/10000 [09:55<00:35, 16.98it/s]\n",
      "Loss: 0.553 :  94%|█████████▍| 9397/10000 [09:55<00:35, 16.96it/s]\n",
      "Loss: 0.553 :  94%|█████████▍| 9399/10000 [09:55<00:35, 16.96it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9399/10000 [09:55<00:35, 16.96it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9401/10000 [09:55<00:39, 15.26it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9403/10000 [09:55<00:37, 15.83it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9405/10000 [09:55<00:37, 16.00it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9407/10000 [09:55<00:36, 16.24it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9409/10000 [09:56<00:35, 16.42it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9411/10000 [09:56<00:36, 16.29it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9413/10000 [09:56<00:36, 16.19it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9415/10000 [09:56<00:38, 15.12it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9417/10000 [09:56<00:39, 14.67it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9419/10000 [09:56<00:39, 14.81it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9421/10000 [09:56<00:38, 15.22it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9423/10000 [09:57<00:37, 15.26it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9425/10000 [09:57<00:36, 15.60it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9427/10000 [09:57<00:35, 16.06it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9429/10000 [09:57<00:34, 16.33it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9431/10000 [09:57<00:34, 16.34it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9433/10000 [09:57<00:34, 16.54it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9435/10000 [09:57<00:33, 16.78it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9437/10000 [09:57<00:33, 16.64it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9439/10000 [09:57<00:33, 16.52it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9441/10000 [09:58<00:33, 16.70it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9443/10000 [09:58<00:33, 16.76it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9445/10000 [09:58<00:32, 16.88it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9447/10000 [09:58<00:32, 17.02it/s]\n",
      "Loss: 0.505 :  94%|█████████▍| 9449/10000 [09:58<00:32, 16.99it/s]\n",
      "Loss: 0.683 :  94%|█████████▍| 9449/10000 [09:58<00:32, 16.99it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9451/10000 [09:58<00:36, 15.17it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9453/10000 [09:58<00:34, 15.77it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9455/10000 [09:58<00:34, 16.02it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9457/10000 [09:59<00:33, 16.43it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9459/10000 [09:59<00:32, 16.42it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9461/10000 [09:59<00:32, 16.55it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9463/10000 [09:59<00:32, 16.41it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9465/10000 [09:59<00:31, 16.74it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9467/10000 [09:59<00:31, 16.91it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9469/10000 [09:59<00:31, 16.72it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9471/10000 [09:59<00:31, 16.78it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9473/10000 [10:00<00:31, 16.81it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9475/10000 [10:00<00:31, 16.76it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9477/10000 [10:00<00:31, 16.64it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9479/10000 [10:00<00:31, 16.78it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9481/10000 [10:00<00:30, 16.85it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9483/10000 [10:00<00:36, 14.36it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9485/10000 [10:00<00:39, 13.12it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9487/10000 [10:01<00:42, 12.20it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9489/10000 [10:01<00:44, 11.53it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9491/10000 [10:01<00:44, 11.51it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9493/10000 [10:01<00:42, 11.84it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9495/10000 [10:01<00:43, 11.49it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9497/10000 [10:02<00:46, 10.78it/s]\n",
      "Loss: 0.683 :  95%|█████████▍| 9499/10000 [10:02<00:45, 10.91it/s]\n",
      "Loss: 0.564 :  95%|█████████▍| 9499/10000 [10:02<00:45, 10.91it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9501/10000 [10:02<00:49, 10.02it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9503/10000 [10:02<00:48, 10.29it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9505/10000 [10:02<00:45, 10.98it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9507/10000 [10:02<00:43, 11.31it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9509/10000 [10:03<00:41, 11.72it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9511/10000 [10:03<00:39, 12.29it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9513/10000 [10:03<00:38, 12.62it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9515/10000 [10:03<00:38, 12.72it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9517/10000 [10:03<00:37, 12.74it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9519/10000 [10:03<00:36, 13.33it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9521/10000 [10:03<00:35, 13.55it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9523/10000 [10:04<00:34, 13.96it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9525/10000 [10:04<00:33, 13.99it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9527/10000 [10:04<00:33, 13.95it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9529/10000 [10:04<00:34, 13.71it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9531/10000 [10:04<00:34, 13.55it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9533/10000 [10:04<00:34, 13.69it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9535/10000 [10:04<00:33, 13.80it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9537/10000 [10:05<00:33, 13.99it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9539/10000 [10:05<00:31, 14.41it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9541/10000 [10:05<00:30, 14.89it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9543/10000 [10:05<00:30, 15.22it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9545/10000 [10:05<00:28, 15.70it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9547/10000 [10:05<00:28, 16.00it/s]\n",
      "Loss: 0.564 :  95%|█████████▌| 9549/10000 [10:05<00:28, 15.97it/s]\n",
      "Loss: 0.635 :  95%|█████████▌| 9549/10000 [10:05<00:28, 15.97it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9551/10000 [10:05<00:30, 14.75it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9553/10000 [10:06<00:28, 15.59it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9555/10000 [10:06<00:27, 15.98it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9557/10000 [10:06<00:27, 16.05it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9559/10000 [10:06<00:26, 16.43it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9561/10000 [10:06<00:26, 16.66it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9563/10000 [10:06<00:26, 16.58it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9565/10000 [10:06<00:26, 16.62it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9567/10000 [10:06<00:25, 16.94it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9569/10000 [10:07<00:25, 16.87it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9571/10000 [10:07<00:28, 15.28it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9573/10000 [10:07<00:29, 14.71it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9575/10000 [10:07<00:28, 14.81it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9577/10000 [10:07<00:28, 14.98it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9579/10000 [10:07<00:28, 14.86it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9581/10000 [10:07<00:27, 15.22it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9583/10000 [10:08<00:26, 15.52it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9585/10000 [10:08<00:26, 15.91it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9587/10000 [10:08<00:25, 15.93it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9589/10000 [10:08<00:26, 15.78it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9591/10000 [10:08<00:25, 15.77it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9593/10000 [10:08<00:25, 16.19it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9595/10000 [10:08<00:24, 16.53it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9597/10000 [10:08<00:24, 16.40it/s]\n",
      "Loss: 0.635 :  96%|█████████▌| 9599/10000 [10:08<00:24, 16.34it/s]\n",
      "Loss: 0.569 :  96%|█████████▌| 9599/10000 [10:09<00:24, 16.34it/s]\n",
      "Loss: 0.569 :  96%|█████████▌| 9601/10000 [10:09<00:26, 15.16it/s]\n",
      "Loss: 0.569 :  96%|█████████▌| 9603/10000 [10:09<00:25, 15.71it/s]\n",
      "Loss: 0.569 :  96%|█████████▌| 9605/10000 [10:09<00:24, 16.16it/s]\n",
      "Loss: 0.569 :  96%|█████████▌| 9607/10000 [10:09<00:23, 16.52it/s]\n",
      "Loss: 0.569 :  96%|█████████▌| 9609/10000 [10:09<00:23, 16.59it/s]\n",
      "Loss: 0.569 :  96%|█████████▌| 9611/10000 [10:09<00:23, 16.83it/s]\n",
      "Loss: 0.569 :  96%|█████████▌| 9613/10000 [10:09<00:23, 16.82it/s]\n",
      "Loss: 0.569 :  96%|█████████▌| 9615/10000 [10:09<00:22, 16.89it/s]\n",
      "Loss: 0.569 :  96%|█████████▌| 9617/10000 [10:10<00:22, 16.77it/s]\n",
      "Loss: 0.569 :  96%|█████████▌| 9619/10000 [10:10<00:22, 16.75it/s]\n",
      "Loss: 0.569 :  96%|█████████▌| 9621/10000 [10:10<00:22, 16.98it/s]\n",
      "Loss: 0.569 :  96%|█████████▌| 9623/10000 [10:10<00:22, 17.08it/s]\n",
      "Loss: 0.569 :  96%|█████████▋| 9625/10000 [10:10<00:21, 17.16it/s]\n",
      "Loss: 0.569 :  96%|█████████▋| 9627/10000 [10:10<00:21, 17.07it/s]\n",
      "Loss: 0.569 :  96%|█████████▋| 9629/10000 [10:10<00:22, 16.65it/s]\n",
      "Loss: 0.569 :  96%|█████████▋| 9631/10000 [10:10<00:21, 16.84it/s]\n",
      "Loss: 0.569 :  96%|█████████▋| 9633/10000 [10:11<00:21, 16.69it/s]\n",
      "Loss: 0.569 :  96%|█████████▋| 9635/10000 [10:11<00:21, 16.99it/s]\n",
      "Loss: 0.569 :  96%|█████████▋| 9637/10000 [10:11<00:21, 17.01it/s]\n",
      "Loss: 0.569 :  96%|█████████▋| 9639/10000 [10:11<00:21, 16.89it/s]\n",
      "Loss: 0.569 :  96%|█████████▋| 9641/10000 [10:11<00:21, 16.99it/s]\n",
      "Loss: 0.569 :  96%|█████████▋| 9643/10000 [10:11<00:21, 16.94it/s]\n",
      "Loss: 0.569 :  96%|█████████▋| 9645/10000 [10:11<00:20, 17.38it/s]\n",
      "Loss: 0.569 :  96%|█████████▋| 9647/10000 [10:11<00:20, 17.22it/s]\n",
      "Loss: 0.569 :  96%|█████████▋| 9649/10000 [10:11<00:20, 17.05it/s]\n",
      "Loss: 0.584 :  96%|█████████▋| 9649/10000 [10:12<00:20, 17.05it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9651/10000 [10:12<00:22, 15.41it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9653/10000 [10:12<00:21, 15.85it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9655/10000 [10:12<00:21, 15.92it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9657/10000 [10:12<00:23, 14.71it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9659/10000 [10:12<00:23, 14.56it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9661/10000 [10:12<00:23, 14.62it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9663/10000 [10:12<00:22, 15.00it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9665/10000 [10:13<00:21, 15.30it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9667/10000 [10:13<00:21, 15.41it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9669/10000 [10:13<00:21, 15.48it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9671/10000 [10:13<00:20, 15.95it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9673/10000 [10:13<00:20, 15.70it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9675/10000 [10:13<00:20, 16.20it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9677/10000 [10:13<00:19, 16.64it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9679/10000 [10:13<00:19, 16.62it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9681/10000 [10:14<00:19, 16.78it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9683/10000 [10:14<00:18, 16.78it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9685/10000 [10:14<00:18, 16.65it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9687/10000 [10:14<00:18, 16.53it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9689/10000 [10:14<00:18, 16.93it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9691/10000 [10:14<00:18, 16.99it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9693/10000 [10:14<00:17, 17.11it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9695/10000 [10:14<00:17, 16.95it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9697/10000 [10:14<00:18, 16.81it/s]\n",
      "Loss: 0.584 :  97%|█████████▋| 9699/10000 [10:15<00:17, 16.96it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9699/10000 [10:15<00:17, 16.96it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9701/10000 [10:15<00:19, 15.60it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9703/10000 [10:15<00:18, 15.96it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9705/10000 [10:15<00:18, 16.11it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9707/10000 [10:15<00:17, 16.30it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9709/10000 [10:15<00:17, 16.54it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9711/10000 [10:15<00:17, 16.87it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9713/10000 [10:15<00:17, 16.82it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9715/10000 [10:16<00:17, 16.74it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9717/10000 [10:16<00:17, 16.51it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9719/10000 [10:16<00:17, 16.41it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9721/10000 [10:16<00:16, 16.73it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9723/10000 [10:16<00:16, 17.06it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9725/10000 [10:16<00:16, 17.07it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9727/10000 [10:16<00:15, 17.22it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9729/10000 [10:16<00:15, 17.42it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9731/10000 [10:17<00:18, 14.89it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9733/10000 [10:17<00:17, 15.52it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9735/10000 [10:17<00:16, 15.95it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9737/10000 [10:17<00:16, 16.35it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9739/10000 [10:17<00:15, 16.59it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9741/10000 [10:17<00:15, 16.56it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9743/10000 [10:17<00:16, 15.55it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9745/10000 [10:17<00:17, 14.92it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9747/10000 [10:18<00:17, 14.75it/s]\n",
      "Loss: 0.747 :  97%|█████████▋| 9749/10000 [10:18<00:16, 14.98it/s]\n",
      "Loss: 0.494 :  97%|█████████▋| 9749/10000 [10:18<00:16, 14.98it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9751/10000 [10:18<00:17, 13.87it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9753/10000 [10:18<00:17, 14.47it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9755/10000 [10:18<00:16, 15.02it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9757/10000 [10:18<00:15, 15.49it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9759/10000 [10:18<00:15, 15.99it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9761/10000 [10:18<00:14, 16.02it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9763/10000 [10:19<00:14, 16.17it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9765/10000 [10:19<00:14, 16.51it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9767/10000 [10:19<00:13, 16.75it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9769/10000 [10:19<00:13, 17.00it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9771/10000 [10:19<00:13, 17.08it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9773/10000 [10:19<00:13, 17.26it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9775/10000 [10:19<00:13, 16.89it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9777/10000 [10:19<00:13, 16.99it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9779/10000 [10:20<00:13, 16.92it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9781/10000 [10:20<00:13, 16.76it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9783/10000 [10:20<00:12, 16.76it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9785/10000 [10:20<00:12, 16.99it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9787/10000 [10:20<00:12, 17.08it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9789/10000 [10:20<00:12, 17.12it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9791/10000 [10:20<00:12, 17.29it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9793/10000 [10:20<00:11, 17.46it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9795/10000 [10:20<00:11, 17.51it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9797/10000 [10:21<00:11, 17.48it/s]\n",
      "Loss: 0.494 :  98%|█████████▊| 9799/10000 [10:21<00:11, 17.50it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9799/10000 [10:21<00:11, 17.50it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9801/10000 [10:21<00:12, 15.66it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9803/10000 [10:21<00:12, 16.11it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9805/10000 [10:21<00:12, 16.16it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9807/10000 [10:21<00:11, 16.50it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9809/10000 [10:21<00:11, 16.73it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9811/10000 [10:21<00:11, 17.02it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9813/10000 [10:22<00:10, 17.27it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9815/10000 [10:22<00:11, 16.80it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9817/10000 [10:22<00:11, 16.64it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9819/10000 [10:22<00:10, 16.77it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9821/10000 [10:22<00:10, 17.10it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9823/10000 [10:22<00:10, 17.45it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9825/10000 [10:22<00:10, 17.31it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9827/10000 [10:22<00:09, 17.31it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9829/10000 [10:22<00:10, 16.85it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9831/10000 [10:23<00:11, 15.13it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9833/10000 [10:23<00:11, 14.73it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9835/10000 [10:23<00:11, 14.70it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9837/10000 [10:23<00:10, 14.92it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9839/10000 [10:23<00:10, 14.95it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9841/10000 [10:23<00:10, 15.08it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9843/10000 [10:23<00:10, 15.63it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9845/10000 [10:24<00:09, 15.79it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9847/10000 [10:24<00:09, 16.08it/s]\n",
      "Loss: 0.647 :  98%|█████████▊| 9849/10000 [10:24<00:09, 16.54it/s]\n",
      "Loss: 0.636 :  98%|█████████▊| 9849/10000 [10:24<00:09, 16.54it/s]\n",
      "Loss: 0.636 :  99%|█████████▊| 9851/10000 [10:24<00:10, 14.77it/s]\n",
      "Loss: 0.636 :  99%|█████████▊| 9853/10000 [10:24<00:09, 15.26it/s]\n",
      "Loss: 0.636 :  99%|█████████▊| 9855/10000 [10:24<00:09, 15.62it/s]\n",
      "Loss: 0.636 :  99%|█████████▊| 9857/10000 [10:24<00:11, 12.21it/s]\n",
      "Loss: 0.636 :  99%|█████████▊| 9859/10000 [10:25<00:10, 13.17it/s]\n",
      "Loss: 0.636 :  99%|█████████▊| 9861/10000 [10:25<00:10, 13.86it/s]\n",
      "Loss: 0.636 :  99%|█████████▊| 9863/10000 [10:25<00:09, 14.27it/s]\n",
      "Loss: 0.636 :  99%|█████████▊| 9865/10000 [10:25<00:09, 14.83it/s]\n",
      "Loss: 0.636 :  99%|█████████▊| 9867/10000 [10:25<00:08, 15.32it/s]\n",
      "Loss: 0.636 :  99%|█████████▊| 9869/10000 [10:25<00:08, 15.64it/s]\n",
      "Loss: 0.636 :  99%|█████████▊| 9871/10000 [10:25<00:08, 15.90it/s]\n",
      "Loss: 0.636 :  99%|█████████▊| 9873/10000 [10:25<00:07, 16.17it/s]\n",
      "Loss: 0.636 :  99%|█████████▉| 9875/10000 [10:26<00:07, 16.08it/s]\n",
      "Loss: 0.636 :  99%|█████████▉| 9877/10000 [10:26<00:07, 16.08it/s]\n",
      "Loss: 0.636 :  99%|█████████▉| 9879/10000 [10:26<00:07, 16.35it/s]\n",
      "Loss: 0.636 :  99%|█████████▉| 9881/10000 [10:26<00:07, 16.56it/s]\n",
      "Loss: 0.636 :  99%|█████████▉| 9883/10000 [10:26<00:06, 16.89it/s]\n",
      "Loss: 0.636 :  99%|█████████▉| 9885/10000 [10:26<00:06, 17.03it/s]\n",
      "Loss: 0.636 :  99%|█████████▉| 9887/10000 [10:26<00:06, 17.17it/s]\n",
      "Loss: 0.636 :  99%|█████████▉| 9889/10000 [10:26<00:06, 17.08it/s]\n",
      "Loss: 0.636 :  99%|█████████▉| 9891/10000 [10:27<00:06, 16.93it/s]\n",
      "Loss: 0.636 :  99%|█████████▉| 9893/10000 [10:27<00:06, 17.12it/s]\n",
      "Loss: 0.636 :  99%|█████████▉| 9895/10000 [10:27<00:06, 17.35it/s]\n",
      "Loss: 0.636 :  99%|█████████▉| 9897/10000 [10:27<00:05, 17.34it/s]\n",
      "Loss: 0.636 :  99%|█████████▉| 9899/10000 [10:27<00:05, 17.21it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9899/10000 [10:27<00:05, 17.21it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9901/10000 [10:27<00:06, 15.58it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9903/10000 [10:27<00:06, 16.16it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9905/10000 [10:27<00:05, 16.73it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9907/10000 [10:27<00:05, 16.92it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9909/10000 [10:28<00:05, 16.69it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9911/10000 [10:28<00:05, 16.46it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9913/10000 [10:28<00:05, 15.84it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9915/10000 [10:28<00:05, 14.18it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9917/10000 [10:28<00:06, 13.40it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9919/10000 [10:28<00:05, 13.50it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9921/10000 [10:28<00:05, 13.54it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9923/10000 [10:29<00:05, 13.81it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9925/10000 [10:29<00:05, 14.06it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9927/10000 [10:29<00:05, 14.29it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9929/10000 [10:29<00:04, 14.37it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9931/10000 [10:29<00:04, 14.47it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9933/10000 [10:29<00:04, 14.53it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9935/10000 [10:29<00:04, 14.48it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9937/10000 [10:30<00:04, 14.69it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9939/10000 [10:30<00:04, 14.72it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9941/10000 [10:30<00:04, 14.58it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9943/10000 [10:30<00:03, 14.51it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9945/10000 [10:30<00:03, 14.52it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9947/10000 [10:30<00:03, 14.64it/s]\n",
      "Loss: 0.590 :  99%|█████████▉| 9949/10000 [10:30<00:03, 14.54it/s]\n",
      "Loss: 0.716 :  99%|█████████▉| 9949/10000 [10:31<00:03, 14.54it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9951/10000 [10:31<00:03, 13.06it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9953/10000 [10:31<00:03, 13.48it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9955/10000 [10:31<00:03, 13.79it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9957/10000 [10:31<00:03, 14.11it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9959/10000 [10:31<00:02, 14.27it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9961/10000 [10:31<00:02, 14.49it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9963/10000 [10:31<00:02, 14.66it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9965/10000 [10:32<00:02, 14.69it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9967/10000 [10:32<00:02, 14.54it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9969/10000 [10:32<00:02, 14.72it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9971/10000 [10:32<00:01, 14.95it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9973/10000 [10:32<00:01, 15.30it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9975/10000 [10:32<00:01, 15.53it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9977/10000 [10:32<00:01, 15.97it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9979/10000 [10:32<00:01, 16.12it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9981/10000 [10:33<00:01, 16.43it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9983/10000 [10:33<00:00, 17.04it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9985/10000 [10:33<00:00, 16.77it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9987/10000 [10:33<00:00, 16.79it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9989/10000 [10:33<00:00, 16.95it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9991/10000 [10:33<00:00, 16.93it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9993/10000 [10:33<00:00, 16.13it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9995/10000 [10:33<00:00, 15.09it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9997/10000 [10:34<00:00, 14.65it/s]\n",
      "Loss: 0.716 : 100%|█████████▉| 9999/10000 [10:34<00:00, 14.84it/s]\n",
      "Loss: 0.716 : 100%|██████████| 10000/10000 [10:34<00:00, 15.77it/s]\n"
     ]
    }
   ],
   "source": [
    "!python location_finding.py --num-steps 10000 --num-inner-samples 64 --num-outer-samples 128 --design-network-type static --num-experiments 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "dad_vtorch",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.11"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}