restore.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                  0\n",
      "warm_up                                                     4.55749\n",
      "label_smoothing                                           0.0454467\n",
      "dropout                                                        0.15\n",
      "weight_decay                                            1.32922e-05\n",
      "final_div_factor                                            5756.37\n",
      "div_factor                                                  28.9679\n",
      "mom_max                                                    0.901919\n",
      "mom_min                                                    0.827043\n",
      "lr                                                       3.1398e-05\n",
      "epochs                                                           48\n",
      "batch_size                                                       24\n",
      "seed                                                            275\n",
      "max_len                                                         192\n",
      "dual                                                          False\n",
      "pooling                                                       first\n",
      "loss_fn                                                         bce\n",
      "optimizer                                                      LAMB\n",
      "gcs                                                        hm-eu-w4\n",
      "path                                 jigsaw/s07/20200515_055220_t2c\n",
      "tpu_id                                                          t2c\n",
      "model_id                                  jplu/tf-xlm-roberta-large\n",
      "dataset           gs://hm-eu-w4/jigsaw/translated-distilled-ds7-...\n",
      "Running on TPU  grpc://10.244.50.162:8470\n",
      "REPLICAS:  8\n",
      "gcs_path: gs://hm-eu-w4/jigsaw/s07/20200515_055220_t2c\n",
      "(8000, 192) (63812, 192) (8000,)\n",
      "batch_size: 192\n",
      "334/334 [==============================] - 21s 63ms/step\n",
      "2659/2659 [==============================] - 112s 42ms/step\n",
      "AUC: 0.961302614355538 toxic: 0.15375 pred: 0.18364666 ratio: 0.144875\n",
      "mean: 0.23866439 ratio: 0.2016862032219645\n",
      "                                                                  0\n",
      "warm_up                                                     4.55749\n",
      "label_smoothing                                           0.0454467\n",
      "dropout                                                        0.15\n",
      "weight_decay                                            1.32922e-05\n",
      "final_div_factor                                            5756.37\n",
      "div_factor                                                  28.9679\n",
      "mom_max                                                    0.901919\n",
      "mom_min                                                    0.827043\n",
      "lr                                                       3.1398e-05\n",
      "epochs                                                           48\n",
      "batch_size                                                       24\n",
      "seed                                                            275\n",
      "max_len                                                         192\n",
      "dual                                                          False\n",
      "pooling                                                       first\n",
      "loss_fn                                                         bce\n",
      "optimizer                                                      LAMB\n",
      "gcs                                                        hm-eu-w4\n",
      "path                                 jigsaw/s07/20200515_055220_t2c\n",
      "tpu_id                                                          t2c\n",
      "model_id                                  jplu/tf-xlm-roberta-large\n",
      "dataset           gs://hm-eu-w4/jigsaw/translated-distilled-ds7-...\n",
      "gcs_path               gs://hm-eu-w4/jigsaw/s07/20200515_055220_t2c\n",
      "auc                                                        0.961303\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.961302614355538"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from train import *\n",
    "\n",
    "def restore(\n",
    "    dataset='gs://hm-eu-w4/jigsaw/translated-distilled-ds7-subs/',\n",
    "    model_id='jplu/tf-xlm-roberta-large',\n",
    "tpu_id=                                                      't2c',\n",
    "path     =                      'jigsaw/s07/20200515_055220_t2c',\n",
    "gcs       =                                             'hm-eu-w4',\n",
    "optimizer  =                                                'LAMB',\n",
    "loss_fn        =                                           'bce',\n",
    "pooling        =                                           'first',\n",
    "dual            =                                           False,\n",
    "max_len        =                                             192,\n",
    "seed            =                                             275,\n",
    "batch_size       =                                            24,\n",
    "epochs                   =                                    48,\n",
    "lr               =                                        3.1398e-05,\n",
    "mom_min           =                                         0.827043,\n",
    "mom_max            =                                        0.901919,\n",
    "div_factor          =                                        28.9679,\n",
    "final_div_factor     =                                       5756.37,\n",
    "weight_decay          =                                  1.32922e-05,\n",
    "dropout                =                                        0.15,\n",
    "label_smoothing         =                                  0.0454467,\n",
    "warm_up                  =                                   4.55749):\n",
    "    \n",
    "    \n",
    "    \"\"\" build and train a TFAutoModel from npz or tfrec dataset \"\"\"\n",
    "    params = dict(locals())\n",
    "#     params.update(kwargs)\n",
    "    params = pd.DataFrame(params, index=[0])\n",
    "    kw_params = params.T[0].to_dict()\n",
    "    print(params.T)\n",
    "    gc.collect()\n",
    "\n",
    "    random.seed(seed)\n",
    "    np.random.seed(seed)\n",
    "    tf.random.set_seed(seed)\n",
    "    os.environ['PYTHONHASHSEED'] = str(seed)\n",
    "    os.environ['TF_DETERMINISTIC_OPS'] = '1'\n",
    "\n",
    "    strategy = setup_tpu(tpu_id)\n",
    "\n",
    "    ## Configuration\n",
    "#     path = f'{path}/{time.strftime(\"%Y%m%d_%H%M%S\")}_{tpu_id}'\n",
    "    gcs_path = f'gs://{gcs}/{path}'\n",
    "    checkpoint_path = f\"{gcs_path}/best_model.tf\"\n",
    "    print('gcs_path:', gcs_path)\n",
    "    params['gcs_path'] = gcs_path\n",
    "    \n",
    "    if dual: ## HACK: dynamic import :/\n",
    "        from dual_model import build_model, tf_dataset, np_dataset\n",
    "    else:\n",
    "        from single_model import build_model, tf_dataset, np_dataset\n",
    "\n",
    "    if dataset.startswith('gs://'):\n",
    "        datasets = tf_dataset(dataset, batch_size, max_len, seed)\n",
    "    else:\n",
    "        datasets = np_dataset(dataset, batch_size, seed)\n",
    "    train_dataset, valid_dataset, test_dataset = datasets\n",
    "\n",
    "    ## Load and Train\n",
    "    with strategy.scope():\n",
    "        model = build_model(**kw_params)\n",
    "#     model, preds, sub = train_model(model, strategy, checkpoint_path, **kw_params)\n",
    "\n",
    "#     ## Save results\n",
    "#     plot_history(model.history, path, gcs)\n",
    "#     history = pd.DataFrame(model.history.history)\n",
    "#     print(history)\n",
    "#     history.to_csv(f'{gcs_path}/history.csv', index=False)\n",
    "\n",
    "    # load best\n",
    "    # latest = tf.train.latest_checkpoint(checkpoint_dir)\n",
    "    checkpoint_path = f'{gcs_path}/best_model.tf'\n",
    "    with strategy.scope():\n",
    "        model.load_weights(checkpoint_path)\n",
    "\n",
    "    batch_size = batch_size * strategy.num_replicas_in_sync\n",
    "    print('batch_size:', batch_size)\n",
    "\n",
    "    preds = model.predict(valid_dataset, verbose=1)\n",
    "    sub_y = model.predict(test_dataset, verbose=1)\n",
    "\n",
    "    ## Load Dataset\n",
    "    comp_ds = '../input/jigsaw-multilingual-toxic-comment-classification'\n",
    "    valid = pd.read_csv(f'{comp_ds}/validation.csv')\n",
    "#     test = pd.read_csv(f'{comp_ds}/test.csv')\n",
    "    sub = pd.read_csv(f'{comp_ds}/sample_submission.csv')\n",
    "\n",
    "    valid['pred'] = preds\n",
    "    valid.to_csv(f'{gcs_path}/valid_oof.csv', index=False)\n",
    "\n",
    "    valid.groupby('toxic').pred.hist(bins=100, log=True, alpha=0.5)\n",
    "    plt.legend([0, 1])\n",
    "    save_fig('valid_hist.png', path, gcs)\n",
    "\n",
    "    valid[valid.toxic == 1].groupby('lang').pred.hist(bins=50, log=True, alpha=0.34)\n",
    "    plt.legend(valid.lang.unique())\n",
    "    save_fig('valid_toxic_hist.png', path, gcs)\n",
    "\n",
    "    valid_auc = roc_auc_score(valid.toxic, valid.pred)\n",
    "    print('AUC:', valid_auc,\n",
    "          'toxic:', valid.toxic.mean(),\n",
    "          'pred:', valid.pred.mean(),\n",
    "          'ratio:', (valid.pred > 0.5).mean())\n",
    "\n",
    "    ## Submission\n",
    "    sub['toxic'] = sub_y\n",
    "    sub.to_csv(f'{gcs_path}/submission.csv', index=False)\n",
    "\n",
    "    sub.toxic.hist(bins=100, log=True)\n",
    "    save_fig('sub_hist.png', path, gcs)\n",
    "    print('mean:', sub.toxic.mean(), 'ratio:', (sub.toxic > 0.5).mean())\n",
    "#     assert sub.toxic.mean() < 0.3\n",
    "\n",
    "    ## Save params\n",
    "    params['auc'] = valid_auc\n",
    "    params.to_csv(f'{gcs_path}/params{valid_auc:04f}.csv', index=False)\n",
    "    print(params.T)\n",
    "\n",
    "    return valid_auc\n",
    "\n",
    "restore()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAW4AAAD8CAYAAABXe05zAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nO3de5wcVZ3w/8/p29wnk8nMZCaZ3BPAcAmXgDGCP9BHEXx+oC5ikBXwgc2zKor+dlX2p+6zy6O76or+5LWiCz9YYVEgigIioFwCKBDCAEnIPZPJJJO5X/t+q6rz/FE9k5nJzPRcuqe7Jt/36zWv6a6qrvp2Vfe3T50655TSWiOEEMI5XLkOQAghxNRI4hZCCIeRxC2EEA4jiVsIIRxGErcQQjiMJG4hhHCYrCVupdRHlVIHlFKNSqnbs7UdIYQ41ahstONWSrmBg8CHgePAm8B1Wuu9Gd+YEEKcYrJV4r4IaNRaN2mtE8AjwNVZ2pYQQpxSPFla72KgZdjz48B7x1u4qqpKL1++PEuhCCGE8zQ3N9PT06PGmpetxJ2WUmozsBlg6dKlNDQ05CoUIYTIO+vXrx93XraqSlqBJcOe16emDdFa36O1Xq+1Xl9dXZ2lMIQQYu7JVuJ+E1ijlFqhlPIBm4Ans7QtIYQ4pWSlqkRrbSilbgX+CLiB+7XWe7KxLSGEONVkrY5ba/008HS21i+EEKcq6TkphBAOI4lbCCEcRhK3EEI4jCRuIYRwGEncQgjhMJK4hRDCYSRxCyGEw0jiFkIIh5HELYQQDiOJWwghHEYStxBCOIwkbiGEcBhJ3EII4TCSuIUQwmEkcQshhMNI4hZCCIeRxC2EEA4jiVsIIRxGErcQQjiMJG4hhHCYrN0seC4JBALce++9FBcXjznfNE1cLhdKqVmOLDMMw8DjceZHwcmxa62xLAu3253rUKbMsiwAXC5nlv0Mw8Dtdk/4nY1Go1x//fUsXLhwFiObHGd+4mdZOBzm0KFDfP7znx9zfktLC1VVVRQVFc1yZJmxf/9+zjjjjFyHMS1Ojj0UChEIBFi0aFGuQ5mygYEBEokENTU1uQ5lWo4cOUJ9fT1er3fcZX75y1/S398viduplFLU1NSwbt26MeeXlJRQW1tLaWnpLEeWGYZhcM455zjujEFrjWma4x6XfDcwMEB/fz8rVqzIdShT1t3dTTwep76+PtehTIvP52PlypUUFBSMu8wLL7yQt98JSdxptLe388orr5BMJnMdihBCAJK40yopKWHJkiXs2rUr16EIMWdprUkkEvT09GCaZta3ZxgGbW1tI64vlJaWUlFR4Yh6e0ncaZSXl7Ny5UpHXkASwkm6urqYP38+RUVFWa+iqKqqwufzDSVpy7Lo7e0lFApRXl6e1W1ngiRuIUReMAyDkpKSWalXVkrhcrmGErdSiqKiIgzDyPq2M0ESdxa8dKCL/nCC99SVc0Zd/v96C5EvBpO21joj6xktHo/zm9/8hhdeeIErrriCa665BqVU3l6EHI8k7jSi0Sjd3d1D7VYno20gSmcgTu08ZzYPFCLXHtp2lEBseqXfq9YtYknlyX0utNYcO3aMhoYGvvWtb/Gd73yHCy64gJUrV8403Fk3o8StlGoGgoAJGFrr9UqpSuBRYDnQDFyrte6fWZi5093dzfbt2x1zCiXEXJAwLBLG5AtLw1kTlNb379/P6tWrqaurY8OGDTQ0NDgycWfi8ullWutztdbrU89vB17QWq8BXkg9d6ylS5fysY99DJ/Pl+tQhBAzFIlEhnpAl5SUEIlEchzR9GSj3cvVwAOpxw8AH8/CNoQQYsqqq6uHqj7b2tryslfkZMw0cWvgT0qpt5RSm1PTFmqt21OPO4Ax94xSarNSqkEp1dDd3T3DMIQQIr21a9fS0dHBQw89RFNTExs2bMh1SNMy04uTF2utW5VSNcBzSqn9w2dqrbVSaswKJ631PcA9AOvXr5/ZJWQhxJxy6Rk1JM3p1XFXlY7djV0pRVVVFZ/73Oc4ePAgt956K/Pnz59JmDkzo8SttW5N/e9SSv0OuAjoVErVaa3blVJ1QFcG4hRCnEJWVWdn3B+Px8PZZ5/N6tWrJxynJN9Nu6pEKVWilCobfAx8BNgNPAncmFrsRuCJmQYphBDihJmUuBcCv0s1XPcAv9JaP6uUehPYopS6GTgKXDvzMHOnubmZZ555hkQiketQhBACmEHi1lo3ASeNp6m17gU+NJOg8kldXR2XXXYZDz30UK5DEUIIQG5dllZBQYFjRgwTQkwsHo/z8ssv88c//pF4PJ7rcKZNurwLIfJP08uQjE7vtfXroaRqzFnxeJwdO3bw7LPPsmHDBsfetUoStxAi/3TugZh/eq+tPn3cxF1aWsr111/PgQMHZhBc7sn5vxDilOFyuSgoKHDcaICjSeIWQgiHkcQthDhlWJZFNBolmUwSiUQcO+qn1HELIfJPxRKIV07vtd7xLzgGAgHuvPNOOjo6+Jd/+Re+/OUvc84550wzyNyRxC2EyD/v+b+zstqKigq+//3vE41GKSgocGwz3zmduC3LGurxOHhBwjRNkskkSim8Xi+7d+/mqaeeYuHChXz2s5+VcbeFEHlvTifuw4cP893vfpfDhw+zdetW3G43v/3tb3n88ccpKyvj1ltv5YwzzqCkpITf/e53mKaZ65CFOKVprXPS4mOm97icbc48T5ik+vp6vvnNb1JXVweA3+/n6aef5ic/+Qkf+chHePXVV0kkEhw/fjwrY5Ec74/w+uEe3j0+kPF1CzHXuFwuEolETpKo1ppEIuGYqpM5XeIuKipiwYIFQwejo6ODBQsWMG/ePFasWMGOHTtoaGhg586duN3uk37pjx07xgMPPMDhw4fRWrNjx44xtxMIBBgYGMDjsXenzx9iQdIk2gmHOqHY68bsyc4wlZnQ39/Pzp07cx3GtPT19Y17XPJdIpEgHo/j90+zo0kOxWIxTNOkp6cnY+scbOkx+D3KJtM0cbvdQ8+11kSjUQzDGJqezzd4mdOJe7Thp2FKKZRSvPe97+X9738/SqkRBxLsEvs3vvENOjo6uOeeezj77LPHXO/hw4epra2lpKQEgL1vttAbiA3N984r5Oyzl2TpXc3cjh07xn1v+c7JsQcCAfr7+1m2bFmuQ5my3t5e4vE4ixYtyuh6Z6uaZP/+/SxbtuykMbmHl/ZffPHFWYllOk6pxF1bW0tPTw+BQIAjR45QV1eH1+sd9xfe5XLR19fHtm3bTvqFHr2cy+U6MV8pUMNPuVzjvjZfuFwux/UmG/whzvd9O57BwoMT43dy7GDHP/i9dSJnRj1JTU1NfOtb32L79u187Wtfo7e3l4985CPcdtttPPvss2zcuDHtB6+4uJjFixc79gALIeaeOV3iXrZsGT/84Q/5wQ9+MDRGwYoVK7j66qtxuVz4fL60pUyPx0NxcbHjSqNCiLlrTidut9tNcXHxSdPHmjaevr4+du7c6diusUKIuUfO/9NYvHgxl19+uXTMEULkjTld4s6EgwcPsmXLFmKxWPqFhRBiFkiJO43TTz+dv/mbv6GwsDDXoQghBCCJOy2ndYUVQsx9UlWSxoEDB6SqRAiRV6TEncYZZ5zB5s2bpapECJE3JHELIYTDSFVJGvv375eqEiFEXpESdxqnn366VJUIIfKKJO40pKu7ECLfSOIWQgiHkcSdRn9/P3v27JHbmgkh8kbaxK2Uul8p1aWU2j1sWqVS6jml1KHU//mp6UopdZdSqlEptUspdX42g58NhmEQDoelI44QIm9MpsT9C+Cjo6bdDrygtV4DvJB6DnAFsCb1txn4WWbCzJ3q6mouuuiiWbmdkhBCTEbaxK21fgXoGzX5auCB1OMHgI8Pm/6gtm0DKpRSdZkKVgghxPTruBdqrdtTjzuAhanHi4GWYcsdT007iVJqs1KqQSnVkM835cyEQCzJywe6eOVgN6YlVS5CiJmZ8cVJbVf+Tjkbaa3v0Vqv11qvr66unmkYWWOaJolEYkZ13JGEydvHBnj7aL8kbiHEjE03cXcOVoGk/nelprcCw29nXp+a5lhNTU088sgjJBKJXIcihBDA9BP3k8CNqcc3Ak8Mm35DqnXJBsA/rErFkdasWcMNN9xAQUFBrkMRQghgEmOVKKUeBi4FqpRSx4H/BXwP2KKUuhk4ClybWvxp4EqgEYgAn8tCzEIIcUpLm7i11teNM+tDYyyrgS/ONCghhBDjk56TQgjhMJK4hRDCYSRxCyGEw0jiTqOjo4M///nPJJPJXIcihBCA3AEnreLiYurq6nC5Zv4bp4End7bhUvC+VQuom1c08wCFEKccKXGnUV5ezurVq3G73RlZX0tfhKO9EaIJGSZWCDE9kriFEMJhJHELIYTDSOIWQgiHkcQthBAOI4k7h7TWcks0IcSUSXPAHHl2dwdul2LtonIuWZO/45ELIfLPKZ24tdY0NTXR2tpKbW0tq1evzkh77cmIGxYAidR/IYSYrDldVdLS0sI///M/881vfhPTtNtNHzx4kNtvv5277rqLnp4elFJYlsWzzz4rN0sQQjjCnE7cxcXFbNy4kb1796K1JplMcuedd3LNNdfgcrl45ZVXWLJkCS6Xi5UrV+bkTu5JUxOKGUTixqxvW4hTQqQfXvt3eO2nkIzkOpqMmNNVJQsWLOCCCy4YunvNsWPHKCgo4JxzzsGyLP74xz9SUVHBiy++yFVXXYVSatZj3NceYF97gPIiLzdfvGLWty/EnKdNiAdTj+dGY4A5nbhHi0ajFBcXo5SiqKiIWCxGaWkpl1xyCYWFhSct397ezpNPPsmRI0fo7e3lwIEDY663t7eXaDSKz+cDwBsaoCI5tRJ0oXZz4EBuqmoCgQAHDx7MybZnQmtNIBAY97jku1gsRjQadWQVXTgcxjRNwuFwrkNJLxGBWJX9uLEJ3D56e3uxLGvCs+z+/v5ZCnDqTqnEPW/ePAKBAKZpMjAwQHl5Oeedd95Qwh1twYIFfOxjH6O5uZnf//73LF26dMzlLMuipqaG4uJiAN7udRM04lOKLQQ8f8yuh7/63EUU+04cGpdLUeDJXq2W3+8f973lM601wWDQkbGD/YPp9/tZsmRJ+oXzTF9fH/F4nLq6ulyHkl6kH7oG7Mf1i8BbTDweZ/HixRPeS7a8vHyWApy6OZ24/X4/r732Gu3t7fz5z3/m3HPPpaKigscee4zGxkY2btw44S+uz+cjHo/T0NAAQFHR2KP5eb1eCgoKhuZrtxfTNf3WIr/e0TXi+ZqaUv77ukXTXl86brebwsLCnFQVzYTWGrfbPe5xyXfxeByv1+vI+H0+H1prZ8RuhcGVOgMuLARf0dB+nyhxZ2pguWyY0xcnk8kk0WiU6667jr6+PgzD4NZbbwXgwgsv5L3vfW/a5n+rVq1i06ZNOb3L+9yolRNCZMqcLnFXVVXxqU996qTp119//aTXceDAAbZs2UIsFstkaFNyrDfCA6814/O42HThEseVjIUQmTWnS9yZcNppp3HLLbeMefFytiRMi75wgv6w8y5iCSEyb06XuDOhtbWV559/Pi+u/BuW5pWDPaDggqXzKS2UwyfEqUhK3GlUVlaybt26DHfOmV6ttWlp3j7Wz9tH+4gkjKFpScPCMKXrvBCnCimypZHJe06eMNM66hOv39bUy5vNfSyuKOJT653XrEwIMXWSuNNobW1l69atGbzLu2bmiRt2tAxQUuChpS+C1jAQSfJqYw8KeO/KBbhdcgFTiLlKEnca8+bN4/TTT2f//v0ZWmNmEuqetsCI56G4wfYjfbgUXLSiMmPbEULkH0ncaZSWlg4NRCWEcJB4EGJ++2+OSZuNlFL3K6W6lFK7h037J6VUq1JqR+rvymHz/kEp1aiUOqCUujxbgc+W9vZ2Xn755QxWleSXxq4QPaEEh7tDuQ5FiMzq2gdv/xfsfTLXkWTcZIqRvwA+Osb0H2utz039PQ2glFoLbALOTL3mbqVU/vYbnYTS0lKWLl3q6BL3oc4gv25o4dnd7SfNa+wK0h2Mc7BDErcQTpE2G2mtXwH6Jrm+q4FHtNZxrfURoBG4aAbx5VxZWRkrVqzI63ELxtMfTtAZiNHuj3G8P0qHP3e9P4XIKq0h2AGBdpjiAG9ONJM67luVUjcADcDfaa37gcXAtmHLHE9NEzmw9UAXR3vnxsDxQkxIW3a1iGXAuk1QObfHtp/u+f/PgFXAuUA7cOdUV6CU2qyUalBKNXR3d08zjOyTO7ELIfLNtErcWuvOwcdKqXuBp1JPW4HhvUDqU9PGWsc9wD0A69evz9vMePDgQX7961/ndJCp2RA3THpCI08xK4q9eBxcty/EXDWtxK2UqtNaD17p+gQw2OLkSeBXSqkfAYuANcD2GUeZQ6effjq33HILd999d65DyarmnjDNo6pVbnjfMhaU5m44WyHSigeha79dVaJPnWEf0iZupdTDwKVAlVLqOPC/gEuVUudidwNsBv4ngNZ6j1JqC7AXMIAvaq3N7IQuhDjlRfuh8flcRzHr0iZurfV1Y0y+b4Llvwt8dyZBiewwLT00NOy8Yi8uGddb5KNECIwEuH1QUJrraPKS9Jw8hQRiBr94rRmAzR9YSUlB5g5/d9BudljodXPawrKMrVecgg6/BB3vQvUZcNYn7JYiHe/a5/c1Z4DXAbdLyzJJ3Gk0NTXxhz/8gXjcOW1DNWSlJUxPKM4bTfZ4KJefWYtr2EBWzb0R/nKoB5/HxaFOuzPPh9cuxJfmJsfD45Q7+5zixvvMmkk48Kz9eN5iSdxI4k5ryZIlXHHFFTzwwAMzXFNmRgVMt15Lw/1/OQJALDn+5YVQ3LDH8jbHT/ChuIHX7aLA66LA4yaSMDnYGcStFB85c+zXJAyLg51BAD54Rk3a6B99swV/NMn7V1dx1uJ5aZcXc5TW0HA/JMInOtBYSXucEWOKLboSYft1yWjm48wTkrjT8Hq9lJaWZqA0mK3S5MnrjSTSXw/+1RvHhh5Xj7PMb9+2W3KevrCMZVXF9E3z1mnH+yP4o0kqinwsnj+ytBRNmkQSJkm5EYRIROykO6j3MLw+TmuuYAeEuiHSe/K8fb/PTnx5RBK3SOtAZ5ADqVI0gKU1f9rTgUsp1i+fn7bJ4M4WPwc7g6ytKz8pcY/HtDQv7OtEa7hoZSXzi30zeg9ijunaB8e2pV9ujpLELaZMA/s77ES+dlE5C8ZZLpIwMLXGsKZemra0Zl97AEvD2fXzmF88/XhFHov54e2H7MfDS9vj2fkIKPfUq0/mGEncImsefP1orkMQ+c6yIB5Iv9ygyST3U4AkbiFEdkT74ejroFyw5r+BS9JNpsieTMMwDCKRiAw0NY64YRFNmFm9y/zgNjxuhdftojcU58kdbQC8xycXNfNOy3Zofdtufx0PAgpWfzDXUc0pkrjTOHr0KE8++SSJxPRaVMx1v9/ZlvVtPP6O3brlohWVvH91FaalGYgmAY2Wa5b5Jxm1S9siayRxp7Fq1Squu+66OT/I1GzoCMTYur9rxLToqKaLvaE4u477sbTGmsJJzu5WP93BOHUVhZxRW56JcOem3ibobYTCcli6wW4/fXirXTpedC6Upm97nxEDx+zBoUzndGzLJ5K456xsdfiZvr5wIm1b8GDMYEfLwJTX3dQd5nB3CMMql8Q9kWA7tL4FZbV24kZD2ztgJmD+8tlL3KFuOw4xLTLY8pyVX0lbCJE5kriFEMJhpKpEzMBsVMfktsrH0ppfNxzHMC0uPb2axdITaGYOPQf+43b3djFtkrjFDGQuoQ5EkrT0RegOjb5YleMqn1AXse4mwhQTN8brIzpXTKPJq7ZgoMV+XFYLnjR3TIr22+OMiBmRxH0KWhA+TFGyj7Cviv7ikXfDrg28i8eKMVC0lFDBwkmuceal4h0tA6MuSk5indq+4PlaYw994fStEw53h+j0x5hf4uM9deVwvMHuiVe5EiqWjFg2GEvy7nE/VYef4PT2g7SVnQ28Z8rva1a0bLeb4C1YbQ97Om3TOIZmEnY8DGioWwe+ErvFyAgaml+1O+JE+mYQnxgkifsUVBk9woJIE90laxgoWsrwklZtaA+FRoCku3gKiTsbpeKT16lT456YwzpD9YUTHDzSN2wZMCwLhcKdGi/ctDQaTVN3mN2tflZUlbBmYSmu1ndQ4W6U2zeUuO1miJpgJM72pm7O6A6R94PNtr4F0QHwlc4wcY+iNYx358Gx7u/YvnP8dZ3CA0JlwymduLXWhEIhOjo6qKuro7T01LpNUlW4kQXhw/TOO4du36Jch5NWQ3M/bx8dQE9wSr+3LcD+9iClhR7+x8X22cR/vd5MIGpgpRL+kZ4wd794mHWdPXygbuTrtx/p482mbi44/iAXmgaKU7hnZjwI234+zsz5sPqKWQ1HnDCnW5V0dXXxn//5n9x7771YqRHqWlpa+PnPf85jjz2G3++nv7+fLVu2cODAgRxHO/sUGhcWyiHd+TVgpumYM7iMOWwh09KYemS6N7Ue84YrOvVabZr2vslU8I6UKnGP9ydyZk4nbtM08Xg8PP7441iWhWVZ3HXXXdTW1nLw4EG2bdtGfX09Z599dq5DPaWs7nmB93Q+RUV0ZqMHLvG/SUW0hdrg7gxFNrFXG3v4TUMLbx9N0507GYN3fmX/xfzT36DW8O5v7PX0j9pXycjQNoxokD1tfva8+SLxNx+y79k4HUdesdfZsn36MYtZMaerSmpra7nyyiv5wx/+ANil7UgkwhVXXEFNTQ0vv/wyGzdupKenh8LCQizLwuU68VsWi8Vob2+nra2NWCxGMBgcczvxeHzEQFQuI47Xyt+uvFp5MFxFw567KLDs5lmWy4fhKsIFWXsPxckBfGaEQjMyo20UGQFcLh+F2jppPR7LHDpebjOO10riSpUSLWV/7E3lI2QBMQNSyxrxKF4rgeEuGloeAOXCHwjiB+Z5TYKVw7462rTrfJXLHis6GYXednteIADJsctHkUiEeDw+xudKo03D/jz1tOMyE1A5AJ7KE4skI0PbMCwXfQkvJAIECjWFyjf0fk4SM8D02TEFg4C2n5tAwA/4wV0BFUG7yZ459mAwUTwkImGCAS9YvvHvF5lvQmHwWiQSCUKh0IRjECWTyVkMbGrmdOJWSo245VgwGKSsrAyXy0VZWRnhcJj+/n4syyIUCmGa5ojEHQgEePXVV2lubsbv99PV1TXWZgiFQiilKCiwm0K54gGKzPw96NpdQMI3f+i5W1uc3rvVfuLykfD5cCkoMkNZ2b7hLQd3AW70jLZhuYtAeUC5T1qPL+keOl7eZJCqaBfV4YNYysOxigvtODyldBoaFUiAx142EQxTZIVJeitG1G8r5RnahhE26eoaltQ7dttN3MpqofYsMBJgpK6X9PaDb+zkEI1GiUajJ3+u4iECB/9MXziBx62oryiF/jAYw5Yz4kPbMC1Nwmd/bnusInxRF4zzWSWYii3mtZfRGpJFdvIdFLbsecnYifcxSshyYRzahmqKASVjbysfdfeCx0c4HKanpwePZ/wUGInkb1vzOZ24RysuLiYcDqO1JhKJUFRURF1dHbfccsuYy9fU1LBhwwY6OjqoqKhg1apVYy6ntaa2tnbo4uYbPUcJZGjwHLcVZ3XPiwAcqbyEhOfkL5LSJmt6nkdpi2MV7yXqq6Qy0kR1aGS9/aGqD2G5fCw0whTHTozql3SXUBRrG1Gf21m0goB34nbLC8KNVIUPEfPM42jlxkm/p4J4NwVmiI6iVQS8CyiNd7LY//aIZY7O30jMO0Z7Dm1xWs/zKG1SmugmWrSEhPKeFGvYpdjlt39I+1zzKaeX4lgbhstHwFMJSlGQ7GOVT6NqimGZfWy76CHg76Yo1o6LE8nZ711AwLsAnxFiqf9NVpXMgzOvAk8hxPdAog/m18GqVXYTw7ZUS5eli6Ck6kRgwQ5oehlcbgbqP0j/gJ8VK0Y2ySTURfvBLmKxCIVeF6sK5qMXzefZzmJiSZMLlleytMQc2kbStOiN2dU3yzzzKFYJCJn2+NdnfcI+E2h8AcI9YPZDQT+4wxBqsLfn6wGtOd4fIRgzqKisoW7VKruap33s5nvdRjFxy0O9bwo3QcgHK5aCr4REIsHy5cuHCltjmTcvf9sTzenEHYvFOHbsGMFgkObmZmpra7Esi23btvHWW29x1llnTfiLC7B8+XKuueYa7rvvvlmKeiSlLSpiLSjgqB6vFK+piLbgwqLNWgdAYTLA/FjLsCXsdWVSoWFvI+Sb2d20vWZkRKwAx63xT2HnxY7j1saE3UVMS9Pcm/kSk0snKQo3Q18pWNO4QJeMQl+TnVTrp/bSY30RIgmT02vLoWSCy6bJSGob3hPT/K0QaD3x3IzbywwTihsMRJP4JnGzaZFbczpx9/T08Nxzz3Haaafx5JNPcu211/LVr36Vhx9+mNraWi6++OIRVSNj6evrY+fOnZimMz7MldFmipP9lCY6s7eNSBNeM0ppPFUVYUZZGNwzYpmBoiXEPfk9Sp+lNR2BGKGuEGFll1g7/OnvZRhJmLT7Y/g7gpyx1K4m6ArGCLvDlIfiLJjMGOHaQnftI9AbZZ/ycfrSRbhcM2/D0hNK4HUnKS/0UlLkJRgzONwdZk3CmFGFRlcwhmlp5hV5KfadSBuW1nQF4mg0VaUFeN1zur1D3pjTibu+vp6vfe1rJ03/9re/Pel1+Hw+ysvLR9SV57O6WWhhURd4l7JhPwwFZpgV/a+OWOaA+8N5n7i1hubeCMeMftoGuoempzvSwZhBIGbQeKSPM5bWAtDSF6Ut7GfFqhgLqifxtdIW+ug2ero0x/q8rFlShysDjQ9bB+yzn2WVxZQUFdEfSbD1QDc17uSEhfR0WvqjJAyLlVUlJyXuI732fSDLCr2SuGeJ7OU0KioqeM973oPb7c51KEIIAczxEncm9PT00NDQgGEYs7K98lgbhYafmKecQGEGuy9PUVm8A61cxN2l+IuWpH9BPtGaBZFG3NogUFBHzFsx41UWJfuoCe3Da56oN68MHoS2xMS36ereD/5SKF+U/iYFwU77RgfxE035TEvTGYjh7wximJUTvGjtXiUAABn/SURBVHjUquIGnYEI4ea3qAkN4C2NTv7bHumBth12ffwogZh9nSWIl8LCkSvsCycIxw1KCjyUFkhqyaa0e1cptQR4EFiIfY3rHq31T5RSlcCjwHKgGbhWa92v7DqFnwBXAhHgJq3122Ot2wlcLteEV54zrTq0n+pIIz3Fq3OauKsih6mKHKa/cInzEjewbOANfGaEw5X/V0YSd0WslYrYiYt7CljU+xoc2DfxC4/82f6/+kPpE3dfEzS9NGJS0tQ09YQ5qHtJlEz+fQzdbajr96wECgvLweNN+zrAHu3vwNg/Rj2hBD2hBKbPy/LCohHzBqtplswvksSdZZOpKjGAv9NarwU2AF9USq0Fbgde0FqvAV5IPQe4AliT+tsM/CzjUc+i0tJSli5dmtGqEpeVwGeETpTetMZrhPEZIdx6/JK9z4ziM0K4JmhxkWkubeIzQviMUE47WahUHAVmiHGHH9WWHasZHorVY8XsfT3YQUeDz7TfT6Zb2bi0gTsRHFFiHothWsQNk6SZ+04rWmvihkncMKd0j0+RW2l/FrXW7UB76nFQKbUPWAxcDVyaWuwB4CXgG6npD2q7G+E2pVSFUqoutR7HaWtr4/nnn89oL6qa0H6WD2wj6qlg56JrATin4zG81sQtGtZ2PQVAc8UGOsrPyVg8E5kXb+P8tl9hKg9v1t9ErsbHLkn0cFbnExMuU2AEOa/90RHTlg28wbKBN4aee3SC89sezkqMldFmFuy6F8oLJ1yuzR+jdSBK0l0CnrKsxDIVO1v8I0ZcFPlvShcnlVLLgfOAN4CFw5JxB3ZVCthJfXij3OOpaY60fPly/uqv/gqfbzJtvIQQIvsmnbiVUqXAY8BXtNYjukulStdT+slWSm1WSjUopRq6u7vTv0DkiblSMpsr70OciiZ1BUEp5cVO2r/UWv82NblzsApEKVUHDA6O0AoMv5pVn5o2gtb6HuAegPXr1+ftt6i/v5/du3dnpVWJWyeoDB8G7DrcfKa0xYJIExqFJ02VDkBZvBOlLeKecsIF1bisJBVR+84orgnq8SeWyftPnlhPcbJv6DjYcyZXDrEsTU/qVmvWqKqGuGESjA17nx0tkCyAcBfRUT0Ty2Jt9DXvwtV7HFLrC8dHLZM4MfZIQb8ftGdoG2bayuncfL2SpoU/mkQBlSW+EX0hokmTcNzA7VLML5az2amaTKsSBdwH7NNa/2jYrCeBG4Hvpf4/MWz6rUqpR4D3An6n1m8DGIZBNDqzLt3j8ZkRTut9ISvrzjQXFmt6X5z08ouCuwDoLD2DIwXVeKxYBt5rdurXqyKNVEUap/y6pKU51DX2IFmhmDFyXtfrwOtjLrvY/xaHX3prwm3VBd9l8J4P5a5SKCsgOHob48rNdYlY0hyK76LllbiHhTEQSdDcG6HA42L+UkncUzWZqpL3A58FPqiU2pH6uxI7YX9YKXUI+G+p5wBPA01AI3Av8IXMhz17qqurufDCC9OOaSKEELNlMq1K/sL4P9kfGmN5DXxxhnEJIYQYhxQj02hpaeG5556bcMD1CWmLpQPbUVh0lq7NSGeQymgzBWaImGcenWVnznh9+aguuItkpIgBh3UA6gknCCdMYsn8vmYxkWDM4EhPeMQ0w5x8m/f+SJKkGSY5xms6AzEiCZNIYuR1Dq01Lf3REfX1JQVuasomblp5qpLEnUZVVRUXXHABTU1N6Rceg0JTG9yDC5P+wqUZSdzl8Q7K4x34CxfP2cRdFbEvFhquQkclbn80iT+avzfRmIxo0iQ6gx+eUNwgFB/7AnRfOMHAOPunKxgb0SlpQYlPEvc4ZJCpNIqKili4cGHa4V+FEGK2SDYSWZS3rTzFuCZzzOS45pokbjED6b7AzhjDXAw3mWOWbhlJ7NkmddxpDN72bLbvgOO1opTF2vBMMKCU24pTFmvDpa1Up5HZlv5L7jMjlMXa8Jont4UvNPyUxdooTk4wLOokDG7DZ+bvzV0zKZo08UeTJ3XkmUg4bg8iVeBxUeh1Y1p6qB5aT+Ozo1NxBMfsmJaZH+zBDjzDlRZ4cA+7U5BhWoQTJgooK/Q45oYnMyWJO41QKERzczOWldmR5NKZF2tlXuykDqcjlCZ6ODM18FR+0syPHmN+qsfkaIuCu1gU3MlMv+iV0WYqo80zWoeTtA5Eh4ZQnazBu9QsrihiaWUx0aTJ3vaxbvQ7md6pGsuyby7ck5h4JMSZCMSMk2I8Z/E8SoYNGRtJ2O9DYXfyOUXytlSVpFNXV8ell16K1zvJsYzFMJk47RazS46ZE0jiFkIIh5HELYQQDiOJWwghHEYStxBCOIwkbiGEcBhpDphGY2MjTzzxBPF4PNehzNia3hfRuChJ9OQ6lEmrCh+iNN6Fx3L+/s8HvaE44cRkbr6QWQc7g6AYdwyTqeoKxugNJzDy4IbLuSAl7jTmyj0nFVARO8782DF8lnM6qhQZfubHjlGW6Mx1KHNCzLAYiCRH3p1nFgxEkwxEkhgZ+sGIJkwGIsmM/RA4jSTuNDweD4WFhXnQI+vULFk4mxwzkR2SuB0j1z8ckB+JKB9imCw5ZrZ8iGFukcQtpiAfElE+xOAk+bC/8iGGuUUStxBCOIwkbjFFuTztlVPu6ZFjNtdI4hZTlMvTXjnlnh45ZnONJG4xRVJ6cx45ZnONJG4xRVJ6cx45ZnPNnO45aRgGgYA9EPv8+fNRSpFIJAgGg3g8HkpLS2lsbOThhx9m+fLlfOYzn3F8RxshxNw3pxN3c3MzP/zhD9m9ezcvvfQSbrebhx9+mFdffRWPx8PNN9/Mvn37+MxnPsPjjz9OMpmcQ4l7MncyEbZ82Vf5EofId3O6qqS+vp6vf/3rLF68GID+/n5efvllvv/97/PRj36Ubdu2kUwmcbnm4m6QBDB5+bKv8iUOke/mdIm7sLCQioqKoe7qnZ2dVFVVUVpaytKlS2loaODqq6/moYceoq6u7qTbkzU3N3PffffR2NiI1+vl7bffHnM7wWCQvr4+PB57d/r8YapSN3JVaIKlq1BaU51oZ0GyG48VJ1CyJovvfGqS3nKCeRTPZGns2PNpX06FpTxYLi+GpyzXoUyZ6fKhlYuEd/6M1qMUvNLvBiBatCDttcw3wy5cEUVSL8AoGbnwO7ECFPZNjE1L43W7KPCMKpTt2gMuD36/n1AohNvtHndbXV1d03lLs2JOJ+7RXC7X0E1/tda4XC7OOusszj///DGXX7ZsGXfccQcdHR3cfffdnHfeeWMu19jYSF1dHSUlJQDs336MHn8MAIXFiq5nUdokX7+egZI1lIUP5TqMaQk6OHbDXULSU0ZRvCPXoUxZ0jMPy+WlIIMjTU7l+zE6cbmAc6srcSnFPn+AgWiSunmFLCsrObGQAs75FHiL2bdvHytXrqSgoGDcbbz00ktTiGh2nVKJu7a2lp6eHgYGBjh8+DCLFy/G7XaPO4CUUorjx4/z/PPPk0wmJ1xu+P/UE/u/tv+ccBLshBiHGyxvOS3u0Zwcfz7FrjjxtRuMa8yvrFKoYX9OlLZyVym1RCm1VSm1Vym1Ryl1W2r6PymlWpVSO1J/Vw57zT8opRqVUgeUUpdn8w1MpLGxkb//+7+noaGBL33pS3R2dvLJT36SL3zhCzz33HNcfPHFE54qAVRWVnLuuecOVYMIIUSuTSYbGcDfaa3fVkqVAW8ppZ5Lzfux1vqHwxdWSq0FNgFnAouA55VSp2mtzUwGPhmrVq3i3nvvRWuNUgqXy8WaNWv42Mc+BthVJ+l+cQ3DIBwOo7V0JBBC5Ie0iVtr3Q60px4HlVL7gMUTvORq4BGtdRw4opRqBC4CXs9AvFOilBqzRJ2ulD1cJBKhtbV1qG5cCCFybUrt4JRSy4HzgDdSk25VSu1SSt2vlBq8vLwYaBn2suOMkeiVUpuVUg1KqYbu7u4pBz5bampq2Lhxo1SVCCHyxqSzkVKqFHgM+IrWOqCU+hnwv7GvEf1v4E7gf0x2fVrre4B7ANavX5+39RCNjY089thjtLS08M4774y5zLFjx+js7KS4uBiA1sZO+sIJwG5VcrhrADX7NUWTFvAdpzwxkOswpk5BwHOc8qQDYwcMV4ykK0yR4bz4k64EpvJQaAZyHQpgn117sXChaO4NE4obDAQK6PdHRy5Yugs8hRw+fJhAIDBhh7v29va8PdOeVOJWSnmxk/Yvtda/BdBadw6bfy/wVOppK7Bk2MvrU9Mc6bTTTuOLX/wi9913H2+++eaYy7z00kucddZZVFVVATAv9TcoVnNB9gOdgd88/DCbNm1y3BV2rTWPPfoomzZtynUo09La2sqxY8d43/vel+tQpqyxsZFgMMh5512U61CGvDVYNqqw/7UALaPLS++8C8AzzzzDxRdfTFnZ+I0Q6+vrqayszHicmZA2cSv723wfsE9r/aNh0+tS9d8AnwB2px4/CfxKKfUj7IuTa4DtGY16lpWXl/PVr3513PnBYJCrrrqKNWuc2RHktddeY/PmzY5L3JZl8cYbb7B58+ZchzIt77zzDtu3b3dk/Fu3bqWzs9OxP5otLS389V//NQsXLsx1KNMymRL3+4HPAu8qpXakpv2/wHVKqXOxq0qagf8JoLXeo5TaAuzFbpHyxVy0KJlNpaWljq4Dr6ioyHUI0+bk2L1e71D1mtMUFBRQVFSU6zCmrby8fEqNFPKNyodmbuvXr9cNDQ25DmPajh49SnV1tWO/hHv27GHt2rWOK3Frrdm7dy9nnnlmrkOZlmAwiN/vp76+PtehTFlfXx+JRILa2tpchzIthw8fZsmSJXk9qNz69etpaGgY80vp3GJijmith/4Ge14tW7ZsaNrgxYzJtBGfbaNjHxxca+3atUOx53OPsuHxD+7fwaQ9OB3Iy/jHir2srIyysjLH7PvhMQ7W/Q7/zOdj/GPtd7D7eAzGPnqeE8zFYfGyKhaL8b3vfY9Pf/rT3HHHHSSTScD+gPzlL3/hxhtv5LOf/Sy7d+9Os6bZpbUmEAhwxx13sGnTJv7t3/5tKPZIJMK//uu/cuONN7J582YOHcq/sT+01uzatYvPf/7z3HTTTTz99NOYpjli/r//+7/zuc99jv7+/hxGejKtNW+++SY333wzN998M1u3bh1KdoZh8Pjjj3PDDTfwpS99icbGxhxHezKtNb/85S/59Kc/zW233UZ7e/vQj+SOHTu46aabuOGGG7j77rvzsqPat7/9bS677DK2bt06YnpbWxt/+7d/y6ZNm/jd736Xl7GPRxL3FB09epTu7m4efPBBwuEwO3bY1f6JRII//elPfOUrX+H222/nP/7jP/Lug3Do0CEsy+LBBx+kvb2dPXv2AHZ95ec+9zkeeOABrrrqKrZs2ZLjSE+WTCZ58cUX+dSnPsU//uM/sn37dnp6TgxwtHPnTpqamqiurs67/R6Px3n22Wf5/Oc/z2233ca2bdvw+/2AfUxef/11fv7zn/PjH/+YFStW5Djakw0MDPD0009z//33c+aZZ/Lqq68OzXvqqaf45Cc/yQ9+8ANeeumlvGw+d/PNN/OJT3wCwzBGTH/66af5wAc+wM9+9jO2bNlCPB7PUYRTJ4l7ipqamlixYgXFxcWcf/75Q8nPNE3a29tZuXIla9asoaurK+8+CM3NzSxfvpzCwkLWrVvHgQMHAPB4PCxatAiXy0VbW1te1luGw2ESiQR1dXXMnz+f4uJi+vr60Frj9/t5+umnufbaa/NybPXBJF1TU0N1dTVg128D7Nu3D6UU3/nOd3jooYfycniFo0ePsnjxYsrKyjjzzDOHPjcAl112GQ8++CBf+cpXuOSSS/Ju/yulqK+vH7Mue+/evaxbt46qqiqKioryehjX0fJrLzuAaZpDV6PdbvfQ6fpgfdngB1cplXdfwOGxu1yuEVUNhmHw4osvsnv3bq677rpchTiuseqvB0t3L7zwAuXl5ViWhd/vJxAI5NW+H349ZDD2wfgSiQTt7e186Utfoquri9deey2vYoeJPzf79u3j+uuv54477uD1118nkUjkKswpsyxrzO+yE0jinqJly5Zx9OhR4vE4e/bsGWq77Xa7qampoaWlhePHj1NeXk5hYWGOox1pyZIltLS0kEgk2L9/PytXrgTsL+Ybb7zBH//4R77+9a/nZeuY4uJiPB4PPT09BAIBYrEYFRUVaK2prKykra2NLVu2sGvXLvbu3ZtXX8KysjIsy6Kvr2+o/n1w7Pb6+nrWrl1LdXU1VVVVhEKhvEvc9fX1tLa2Eo1GOXTo0NDnBuD1119n7dq1LFu2jHg87qjEvXr1avbt2zf0Y19TU5PrkCZNWpVM0cqVK3G73Xz5y1+mpKSERYsW8fzzz7NhwwYuueQSfvKTn2BZFjfddFPeXaU+7bTTeOKJJ7jtttuYP38+5eXlbN26lbPOOosf/ehHRKNR7rrrLjZu3MgnP/nJXIc7QkFBARdddBGPPvoopmlyySWXcODAAXp7e7n00ku59NJLaW1t5a677mLDhg151a6+uLiYjRs3Do1Uefnll/PWW2+xdOlS1q1bx2uvvcbXv/513G43t9xyS95VN9TU1HDBBRdw22234fF4uO222/jDH/7AhRdeyMc//nHuvPPOoarDwR+kfKG15qc//SmPPfYY8+bNwzAMPB4PZ599Npdffjk/+MEPeOaZZ/jwhz+clwWW8Ug77inSWhOJRIhGoxQWFlJYWEgymaSgoADTNIfqLufNm5d3Dfy11oTDYWKxGEVFRfh8PgzDwOfzMTAwMHRKX1hYmHdfQLCrc4LBIFprSkpKhqqmBu9iYpomkUiEkpKSvEt+w2MvLS0dqn7wer1Eo1Gi0Sher5eSkpK8+9yA3ZoqFArh9XopLS0lFotRWFg41FpJa01xcTGFhYV5VWAZjC8ej6OUGvpc+3w+XC4XgUAAwzAoKyvD5/PlVewTteOWxC2EEHloosSdX8USIYQQaUniFkIIh5HELYQQDiOJWwghHEYStxBCOIwkbiGEcBhJ3EII4TCSuIUQwmEkcQshhMNI4hZCCIeRxC2EEA4jiVsIIRxGErcQQjiMJG4hhHCYvBjWVSnVDYSBnnTLzrIq8i8mkLimKh/jyseYQOKaqmzGtUxrXT3WjLxI3ABKqQat9fpcxzFcPsYEEtdU5WNc+RgTSFxTlau4pKpECCEcRhK3EEI4TD4l7ntyHcAY8jEmkLimKh/jyseYQOKaqpzElTd13EIIISYnn0rcQgghJiHniVsp9VGl1AGlVKNS6vYcx9KslHpXKbVDKdWQmlaplHpOKXUo9X/+LMRxv1KqSym1e9i0MeNQtrtS+2+XUur8WYzpn5RSran9tUMpdeWwef+QiumAUurybMSU2s4SpdRWpdRepdQepdRtqem53l/jxZWzfaaUKlRKbVdK7UzF9M+p6SuUUm+ktv2oUsqXml6Qet6Ymr880zGliesXSqkjw/bVuanps3IMh8XnVkq9o5R6KvU8p/sLAK11zv4AN3AYWAn4gJ3A2hzG0wxUjZr2A+D21OPbge/PQhwfAM4HdqeLA7gSeAZQwAbgjVmM6Z+Avx9j2bWpY1kArEgdY3eW4qoDzk89LgMOpraf6/01Xlw522ep91yaeuwF3kjtgy3AptT0nwOfTz3+AvDz1ONNwKNZ2lfjxfUL4Joxlp+VYzhse/8P8CvgqdTznO4vrXXOS9wXAY1a6yatdQJ4BLg6xzGNdjXwQOrxA8DHs71BrfUrQN8k47gaeFDbtgEVSqm6WYppPFcDj2it41rrI0Aj9rHOOK11u9b67dTjILAPWEzu99d4cY0n6/ss9Z5Dqafe1J8GPgj8JjV99L4a3Ie/AT6klFKZjClNXOOZlWMIoJSqBz4G/P+p54oc7y/IfVXJYqBl2PPjTPzhzjYN/Ekp9ZZSanNq2kKtdXvqcQewMDehjRtHrvfhranT1fuHVSPlJKbUqel52CW2vNlfo+KCHO6z1Gn/DqALeA67ZD+gtTbG2O5QTKn5fmBBpmMaKy6t9eC++m5qX/1YKVUwOq4xYs60/w/4OmClni8gD/ZXrhN3vrlYa30+cAXwRaXUB4bP1PY5UM6b4eRLHMDPgFXAuUA7cGeuAlFKlQKPAV/RWgeGz8vl/hojrpzuM621qbU+F6jHLtGfMZvbH8/ouJRSZwH/gB3fhUAl8I3ZjEkp9d+BLq31W7O53cnIdeJuBZYMe16fmpYTWuvW1P8u4HfYH+zOwdOw1P+uHIU3Xhw524da687UF84C7uXEqf2sxqSU8mInx19qrX+bmpzz/TVWXPmyz7TWA8BW4H3YVQ2eMbY7FFNq/jygN1sxjYrro6nqJq21jgP/yezvq/cDVymlmrGrcT8I/IQ82F+5TtxvAmtSV2l92BX6T+YiEKVUiVKqbPAx8BFgdyqeG1OL3Qg8kYv4JojjSeCG1JX2DYB/WBVBVo2qV/wE9v4ajGlT6ir7CmANsD1LMSjgPmCf1vpHw2bldH+NF1cu95lSqlopVZF6XAR8GLvufStwTWqx0ftqcB9eA7yYOnvJqHHi2j/sh1dh1yMP31dZP4Za63/QWtdrrZdj56YXtdbXk+P9NRhcTv+wrxAfxK5r+2YO41iJfVV/J7BnMBbsOqoXgEPA80DlLMTyMPZpdBK7Du3m8eLAvrL+09T+exdYP4sx/Vdqm7uwP7R1w5b/ZiqmA8AVWdxXF2NXg+wCdqT+rsyD/TVeXDnbZ8A5wDupbe8G/nHYZ3879gXRXwMFqemFqeeNqfkrs7SvxovrxdS+2g08xImWJ7NyDEfFeCknWpXkdH9praXnpBBCOE2uq0qEEEJMkSRuIYRwGEncQgjhMJK4hRDCYSRxCyGEw0jiFkIIh5HELYQQDiOJWwghHOb/ALA35r0ou+JKAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.image as mpimg\n",
    "img = mpimg.imread('sub_hist.png')\n",
    "_ = plt.imshow(img)\n",
    "plt.show()\n",
    "img = mpimg.imread('valid_hist.png')\n",
    "_ = plt.imshow(img)\n",
    "plt.show()\n",
    "img = mpimg.imread('valid_toxic_hist.png')\n",
    "_ = plt.imshow(img)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import pdb; pdb.pm()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.6.9"
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "state": {
     "03df225fb40c4a3cbddfdb6beddb0e4b": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "DescriptionStyleModel",
      "state": {
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "DescriptionStyleModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/base",
       "_view_module_version": "1.2.0",
       "_view_name": "StyleView",
       "description_width": ""
      }
     },
     "15ada43973ce43deabc26c111bc346e6": {
      "model_module": "@jupyter-widgets/base",
      "model_module_version": "1.2.0",
      "model_name": "LayoutModel",
      "state": {
       "_model_module": "@jupyter-widgets/base",
       "_model_module_version": "1.2.0",
       "_model_name": "LayoutModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/base",
       "_view_module_version": "1.2.0",
       "_view_name": "LayoutView",
       "align_content": null,
       "align_items": null,
       "align_self": null,
       "border": null,
       "bottom": null,
       "display": null,
       "flex": null,
       "flex_flow": null,
       "grid_area": null,
       "grid_auto_columns": null,
       "grid_auto_flow": null,
       "grid_auto_rows": null,
       "grid_column": null,
       "grid_gap": null,
       "grid_row": null,
       "grid_template_areas": null,
       "grid_template_columns": null,
       "grid_template_rows": null,
       "height": null,
       "justify_content": null,
       "justify_items": null,
       "left": null,
       "margin": null,
       "max_height": null,
       "max_width": null,
       "min_height": null,
       "min_width": null,
       "object_fit": null,
       "object_position": null,
       "order": null,
       "overflow": null,
       "overflow_x": null,
       "overflow_y": null,
       "padding": null,
       "right": null,
       "top": null,
       "visibility": null,
       "width": null
      }
     },
     "1ec7dd5ce7e04e4184f5c16e9d410e3b": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "ProgressStyleModel",
      "state": {
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "ProgressStyleModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/base",
       "_view_module_version": "1.2.0",
       "_view_name": "StyleView",
       "bar_color": null,
       "description_width": "initial"
      }
     },
     "23976f077e8c4855af1a1cfe193b60ef": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "HTMLModel",
      "state": {
       "_dom_classes": [],
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "HTMLModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/controls",
       "_view_module_version": "1.5.0",
       "_view_name": "HTMLView",
       "description": "",
       "description_tooltip": null,
       "layout": "IPY_MODEL_3fbc3a0e6fb14d8085d98270ec4b33b2",
       "placeholder": "​",
       "style": "IPY_MODEL_2a65d9d5af90464aa718587f9a121a1c",
       "value": " 5.07M/5.07M [00:01&lt;00:00, 2.89MB/s]"
      }
     },
     "2a65d9d5af90464aa718587f9a121a1c": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "DescriptionStyleModel",
      "state": {
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "DescriptionStyleModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/base",
       "_view_module_version": "1.2.0",
       "_view_name": "StyleView",
       "description_width": ""
      }
     },
     "3e88f3383ac8472faef99555f2eb3cec": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "ProgressStyleModel",
      "state": {
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "ProgressStyleModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/base",
       "_view_module_version": "1.2.0",
       "_view_name": "StyleView",
       "bar_color": null,
       "description_width": "initial"
      }
     },
     "3fbc3a0e6fb14d8085d98270ec4b33b2": {
      "model_module": "@jupyter-widgets/base",
      "model_module_version": "1.2.0",
      "model_name": "LayoutModel",
      "state": {
       "_model_module": "@jupyter-widgets/base",
       "_model_module_version": "1.2.0",
       "_model_name": "LayoutModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/base",
       "_view_module_version": "1.2.0",
       "_view_name": "LayoutView",
       "align_content": null,
       "align_items": null,
       "align_self": null,
       "border": null,
       "bottom": null,
       "display": null,
       "flex": null,
       "flex_flow": null,
       "grid_area": null,
       "grid_auto_columns": null,
       "grid_auto_flow": null,
       "grid_auto_rows": null,
       "grid_column": null,
       "grid_gap": null,
       "grid_row": null,
       "grid_template_areas": null,
       "grid_template_columns": null,
       "grid_template_rows": null,
       "height": null,
       "justify_content": null,
       "justify_items": null,
       "left": null,
       "margin": null,
       "max_height": null,
       "max_width": null,
       "min_height": null,
       "min_width": null,
       "object_fit": null,
       "object_position": null,
       "order": null,
       "overflow": null,
       "overflow_x": null,
       "overflow_y": null,
       "padding": null,
       "right": null,
       "top": null,
       "visibility": null,
       "width": null
      }
     },
     "4a4e5352a8714816a1eb15d297ef2077": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "FloatProgressModel",
      "state": {
       "_dom_classes": [],
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "FloatProgressModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/controls",
       "_view_module_version": "1.5.0",
       "_view_name": "ProgressView",
       "bar_style": "success",
       "description": "Downloading: 100%",
       "description_tooltip": null,
       "layout": "IPY_MODEL_6481a898fe2e4e52bafb7bda11c7a9c6",
       "max": 5069051,
       "min": 0,
       "orientation": "horizontal",
       "style": "IPY_MODEL_3e88f3383ac8472faef99555f2eb3cec",
       "value": 5069051
      }
     },
     "5bbd64c445e149a1894b2f195c93ed6c": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "FloatProgressModel",
      "state": {
       "_dom_classes": [],
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "FloatProgressModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/controls",
       "_view_module_version": "1.5.0",
       "_view_name": "ProgressView",
       "bar_style": "success",
       "description": "Downloading: 100%",
       "description_tooltip": null,
       "layout": "IPY_MODEL_d73ea71b0f6f4970a15b4f60b4ce8c56",
       "max": 738,
       "min": 0,
       "orientation": "horizontal",
       "style": "IPY_MODEL_1ec7dd5ce7e04e4184f5c16e9d410e3b",
       "value": 738
      }
     },
     "5eb82dd559dd49a2bdd9078f0289f13a": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "FloatProgressModel",
      "state": {
       "_dom_classes": [],
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "FloatProgressModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/controls",
       "_view_module_version": "1.5.0",
       "_view_name": "ProgressView",
       "bar_style": "success",
       "description": "Downloading: 100%",
       "description_tooltip": null,
       "layout": "IPY_MODEL_e39f3a70f25841049ddde333e84f9b2f",
       "max": 3271420488,
       "min": 0,
       "orientation": "horizontal",
       "style": "IPY_MODEL_a2cbb4062d314b7faff832f5eab5cce6",
       "value": 3271420488
      }
     },
     "6481a898fe2e4e52bafb7bda11c7a9c6": {
      "model_module": "@jupyter-widgets/base",
      "model_module_version": "1.2.0",
      "model_name": "LayoutModel",
      "state": {
       "_model_module": "@jupyter-widgets/base",
       "_model_module_version": "1.2.0",
       "_model_name": "LayoutModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/base",
       "_view_module_version": "1.2.0",
       "_view_name": "LayoutView",
       "align_content": null,
       "align_items": null,
       "align_self": null,
       "border": null,
       "bottom": null,
       "display": null,
       "flex": null,
       "flex_flow": null,
       "grid_area": null,
       "grid_auto_columns": null,
       "grid_auto_flow": null,
       "grid_auto_rows": null,
       "grid_column": null,
       "grid_gap": null,
       "grid_row": null,
       "grid_template_areas": null,
       "grid_template_columns": null,
       "grid_template_rows": null,
       "height": null,
       "justify_content": null,
       "justify_items": null,
       "left": null,
       "margin": null,
       "max_height": null,
       "max_width": null,
       "min_height": null,
       "min_width": null,
       "object_fit": null,
       "object_position": null,
       "order": null,
       "overflow": null,
       "overflow_x": null,
       "overflow_y": null,
       "padding": null,
       "right": null,
       "top": null,
       "visibility": null,
       "width": null
      }
     },
     "6b45213384a34a549bd24c55d6e4b583": {
      "model_module": "@jupyter-widgets/base",
      "model_module_version": "1.2.0",
      "model_name": "LayoutModel",
      "state": {
       "_model_module": "@jupyter-widgets/base",
       "_model_module_version": "1.2.0",
       "_model_name": "LayoutModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/base",
       "_view_module_version": "1.2.0",
       "_view_name": "LayoutView",
       "align_content": null,
       "align_items": null,
       "align_self": null,
       "border": null,
       "bottom": null,
       "display": null,
       "flex": null,
       "flex_flow": null,
       "grid_area": null,
       "grid_auto_columns": null,
       "grid_auto_flow": null,
       "grid_auto_rows": null,
       "grid_column": null,
       "grid_gap": null,
       "grid_row": null,
       "grid_template_areas": null,
       "grid_template_columns": null,
       "grid_template_rows": null,
       "height": null,
       "justify_content": null,
       "justify_items": null,
       "left": null,
       "margin": null,
       "max_height": null,
       "max_width": null,
       "min_height": null,
       "min_width": null,
       "object_fit": null,
       "object_position": null,
       "order": null,
       "overflow": null,
       "overflow_x": null,
       "overflow_y": null,
       "padding": null,
       "right": null,
       "top": null,
       "visibility": null,
       "width": null
      }
     },
     "76625d00261a4c16b05a29bd9885e7ac": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "HTMLModel",
      "state": {
       "_dom_classes": [],
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "HTMLModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/controls",
       "_view_module_version": "1.5.0",
       "_view_name": "HTMLView",
       "description": "",
       "description_tooltip": null,
       "layout": "IPY_MODEL_b4adcdd0fb934b4e8b0f7efbef525f64",
       "placeholder": "​",
       "style": "IPY_MODEL_03df225fb40c4a3cbddfdb6beddb0e4b",
       "value": " 3.27G/3.27G [01:26&lt;00:00, 37.7MB/s]"
      }
     },
     "7d865efc6e0f4be886075b93553f8a0d": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "DescriptionStyleModel",
      "state": {
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "DescriptionStyleModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/base",
       "_view_module_version": "1.2.0",
       "_view_name": "StyleView",
       "description_width": ""
      }
     },
     "9b6585770d8e4a83a6be952a064cc4dd": {
      "model_module": "@jupyter-widgets/base",
      "model_module_version": "1.2.0",
      "model_name": "LayoutModel",
      "state": {
       "_model_module": "@jupyter-widgets/base",
       "_model_module_version": "1.2.0",
       "_model_name": "LayoutModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/base",
       "_view_module_version": "1.2.0",
       "_view_name": "LayoutView",
       "align_content": null,
       "align_items": null,
       "align_self": null,
       "border": null,
       "bottom": null,
       "display": null,
       "flex": null,
       "flex_flow": null,
       "grid_area": null,
       "grid_auto_columns": null,
       "grid_auto_flow": null,
       "grid_auto_rows": null,
       "grid_column": null,
       "grid_gap": null,
       "grid_row": null,
       "grid_template_areas": null,
       "grid_template_columns": null,
       "grid_template_rows": null,
       "height": null,
       "justify_content": null,
       "justify_items": null,
       "left": null,
       "margin": null,
       "max_height": null,
       "max_width": null,
       "min_height": null,
       "min_width": null,
       "object_fit": null,
       "object_position": null,
       "order": null,
       "overflow": null,
       "overflow_x": null,
       "overflow_y": null,
       "padding": null,
       "right": null,
       "top": null,
       "visibility": null,
       "width": null
      }
     },
     "a2cbb4062d314b7faff832f5eab5cce6": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "ProgressStyleModel",
      "state": {
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "ProgressStyleModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/base",
       "_view_module_version": "1.2.0",
       "_view_name": "StyleView",
       "bar_color": null,
       "description_width": "initial"
      }
     },
     "b4adcdd0fb934b4e8b0f7efbef525f64": {
      "model_module": "@jupyter-widgets/base",
      "model_module_version": "1.2.0",
      "model_name": "LayoutModel",
      "state": {
       "_model_module": "@jupyter-widgets/base",
       "_model_module_version": "1.2.0",
       "_model_name": "LayoutModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/base",
       "_view_module_version": "1.2.0",
       "_view_name": "LayoutView",
       "align_content": null,
       "align_items": null,
       "align_self": null,
       "border": null,
       "bottom": null,
       "display": null,
       "flex": null,
       "flex_flow": null,
       "grid_area": null,
       "grid_auto_columns": null,
       "grid_auto_flow": null,
       "grid_auto_rows": null,
       "grid_column": null,
       "grid_gap": null,
       "grid_row": null,
       "grid_template_areas": null,
       "grid_template_columns": null,
       "grid_template_rows": null,
       "height": null,
       "justify_content": null,
       "justify_items": null,
       "left": null,
       "margin": null,
       "max_height": null,
       "max_width": null,
       "min_height": null,
       "min_width": null,
       "object_fit": null,
       "object_position": null,
       "order": null,
       "overflow": null,
       "overflow_x": null,
       "overflow_y": null,
       "padding": null,
       "right": null,
       "top": null,
       "visibility": null,
       "width": null
      }
     },
     "b56982b3e6e945b98bb6ceff75718713": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "HBoxModel",
      "state": {
       "_dom_classes": [],
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "HBoxModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/controls",
       "_view_module_version": "1.5.0",
       "_view_name": "HBoxView",
       "box_style": "",
       "children": [
        "IPY_MODEL_4a4e5352a8714816a1eb15d297ef2077",
        "IPY_MODEL_23976f077e8c4855af1a1cfe193b60ef"
       ],
       "layout": "IPY_MODEL_9b6585770d8e4a83a6be952a064cc4dd"
      }
     },
     "cd0d1db27650445485ab49e819c306b9": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "HTMLModel",
      "state": {
       "_dom_classes": [],
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "HTMLModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/controls",
       "_view_module_version": "1.5.0",
       "_view_name": "HTMLView",
       "description": "",
       "description_tooltip": null,
       "layout": "IPY_MODEL_f4961b04478b45c48a12189728bf82a7",
       "placeholder": "​",
       "style": "IPY_MODEL_7d865efc6e0f4be886075b93553f8a0d",
       "value": " 738/738 [00:22&lt;00:00, 32.8B/s]"
      }
     },
     "ceb3bf5654dc4ab7aaab33ce7fcfa136": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "HBoxModel",
      "state": {
       "_dom_classes": [],
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "HBoxModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/controls",
       "_view_module_version": "1.5.0",
       "_view_name": "HBoxView",
       "box_style": "",
       "children": [
        "IPY_MODEL_5bbd64c445e149a1894b2f195c93ed6c",
        "IPY_MODEL_cd0d1db27650445485ab49e819c306b9"
       ],
       "layout": "IPY_MODEL_6b45213384a34a549bd24c55d6e4b583"
      }
     },
     "d73ea71b0f6f4970a15b4f60b4ce8c56": {
      "model_module": "@jupyter-widgets/base",
      "model_module_version": "1.2.0",
      "model_name": "LayoutModel",
      "state": {
       "_model_module": "@jupyter-widgets/base",
       "_model_module_version": "1.2.0",
       "_model_name": "LayoutModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/base",
       "_view_module_version": "1.2.0",
       "_view_name": "LayoutView",
       "align_content": null,
       "align_items": null,
       "align_self": null,
       "border": null,
       "bottom": null,
       "display": null,
       "flex": null,
       "flex_flow": null,
       "grid_area": null,
       "grid_auto_columns": null,
       "grid_auto_flow": null,
       "grid_auto_rows": null,
       "grid_column": null,
       "grid_gap": null,
       "grid_row": null,
       "grid_template_areas": null,
       "grid_template_columns": null,
       "grid_template_rows": null,
       "height": null,
       "justify_content": null,
       "justify_items": null,
       "left": null,
       "margin": null,
       "max_height": null,
       "max_width": null,
       "min_height": null,
       "min_width": null,
       "object_fit": null,
       "object_position": null,
       "order": null,
       "overflow": null,
       "overflow_x": null,
       "overflow_y": null,
       "padding": null,
       "right": null,
       "top": null,
       "visibility": null,
       "width": null
      }
     },
     "deb54208870b488999cc305ae36d2560": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "HBoxModel",
      "state": {
       "_dom_classes": [],
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "HBoxModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/controls",
       "_view_module_version": "1.5.0",
       "_view_name": "HBoxView",
       "box_style": "",
       "children": [
        "IPY_MODEL_5eb82dd559dd49a2bdd9078f0289f13a",
        "IPY_MODEL_76625d00261a4c16b05a29bd9885e7ac"
       ],
       "layout": "IPY_MODEL_15ada43973ce43deabc26c111bc346e6"
      }
     },
     "e39f3a70f25841049ddde333e84f9b2f": {
      "model_module": "@jupyter-widgets/base",
      "model_module_version": "1.2.0",
      "model_name": "LayoutModel",
      "state": {
       "_model_module": "@jupyter-widgets/base",
       "_model_module_version": "1.2.0",
       "_model_name": "LayoutModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/base",
       "_view_module_version": "1.2.0",
       "_view_name": "LayoutView",
       "align_content": null,
       "align_items": null,
       "align_self": null,
       "border": null,
       "bottom": null,
       "display": null,
       "flex": null,
       "flex_flow": null,
       "grid_area": null,
       "grid_auto_columns": null,
       "grid_auto_flow": null,
       "grid_auto_rows": null,
       "grid_column": null,
       "grid_gap": null,
       "grid_row": null,
       "grid_template_areas": null,
       "grid_template_columns": null,
       "grid_template_rows": null,
       "height": null,
       "justify_content": null,
       "justify_items": null,
       "left": null,
       "margin": null,
       "max_height": null,
       "max_width": null,
       "min_height": null,
       "min_width": null,
       "object_fit": null,
       "object_position": null,
       "order": null,
       "overflow": null,
       "overflow_x": null,
       "overflow_y": null,
       "padding": null,
       "right": null,
       "top": null,
       "visibility": null,
       "width": null
      }
     },
     "f4961b04478b45c48a12189728bf82a7": {
      "model_module": "@jupyter-widgets/base",
      "model_module_version": "1.2.0",
      "model_name": "LayoutModel",
      "state": {
       "_model_module": "@jupyter-widgets/base",
       "_model_module_version": "1.2.0",
       "_model_name": "LayoutModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/base",
       "_view_module_version": "1.2.0",
       "_view_name": "LayoutView",
       "align_content": null,
       "align_items": null,
       "align_self": null,
       "border": null,
       "bottom": null,
       "display": null,
       "flex": null,
       "flex_flow": null,
       "grid_area": null,
       "grid_auto_columns": null,
       "grid_auto_flow": null,
       "grid_auto_rows": null,
       "grid_column": null,
       "grid_gap": null,
       "grid_row": null,
       "grid_template_areas": null,
       "grid_template_columns": null,
       "grid_template_rows": null,
       "height": null,
       "justify_content": null,
       "justify_items": null,
       "left": null,
       "margin": null,
       "max_height": null,
       "max_width": null,
       "min_height": null,
       "min_width": null,
       "object_fit": null,
       "object_position": null,
       "order": null,
       "overflow": null,
       "overflow_x": null,
       "overflow_y": null,
       "padding": null,
       "right": null,
       "top": null,
       "visibility": null,
       "width": null
      }
     }
    },
    "version_major": 2,
    "version_minor": 0
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}