train_bgru.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/zake7749/.local/lib/python3.5/site-packages/h5py/__init__.py:34: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n",
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "\n",
    "from sotoxic.data_helper.data_transformer import DataTransformer\n",
    "from sotoxic.data_helper.data_loader import DataLoader\n",
    "from sotoxic.train.trainer import PyTorchModelTrainer\n",
    "from sotoxic.config import dataset_config "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "VOCAB_SIZE = 100000\n",
    "MAX_SEQUENCE_LENGTH = 300\n",
    "EMBEDDING_SIZE = 300"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prepare training and testing data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading the dataset\n",
      "Doing preprocessing...\n",
      "Preprocessed.\n"
     ]
    }
   ],
   "source": [
    "print(\"Loading the dataset\")\n",
    "data_transformer = DataTransformer(max_num_words=VOCAB_SIZE, max_sequence_length=MAX_SEQUENCE_LENGTH, char_level=False)\n",
    "data_loader = DataLoader()\n",
    "train_sequences, training_labels, test_sequences = data_transformer.prepare_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[686,\n",
       "  76,\n",
       "  1,\n",
       "  132,\n",
       "  124,\n",
       "  173,\n",
       "  28,\n",
       "  705,\n",
       "  4324,\n",
       "  9378,\n",
       "  1162,\n",
       "  81,\n",
       "  350,\n",
       "  44,\n",
       "  81,\n",
       "  12,\n",
       "  13594,\n",
       "  50,\n",
       "  6507,\n",
       "  15,\n",
       "  59,\n",
       "  2539,\n",
       "  143,\n",
       "  3,\n",
       "  2736,\n",
       "  31,\n",
       "  111,\n",
       "  1190,\n",
       "  16065,\n",
       "  2418,\n",
       "  6,\n",
       "  51,\n",
       "  20,\n",
       "  12,\n",
       "  239,\n",
       "  1,\n",
       "  349,\n",
       "  29,\n",
       "  1,\n",
       "  46,\n",
       "  136,\n",
       "  3,\n",
       "  38,\n",
       "  3789,\n",
       "  85],\n",
       " [337,\n",
       "  7768,\n",
       "  42,\n",
       "  2818,\n",
       "  14,\n",
       "  628,\n",
       "  3656,\n",
       "  3,\n",
       "  38,\n",
       "  4701,\n",
       "  2584,\n",
       "  22,\n",
       "  90,\n",
       "  46,\n",
       "  929,\n",
       "  3682,\n",
       "  987,\n",
       "  674,\n",
       "  10174],\n",
       " [429,\n",
       "  363,\n",
       "  3,\n",
       "  38,\n",
       "  118,\n",
       "  12,\n",
       "  250,\n",
       "  2,\n",
       "  77,\n",
       "  291,\n",
       "  10,\n",
       "  50,\n",
       "  9,\n",
       "  14,\n",
       "  551,\n",
       "  8,\n",
       "  2391,\n",
       "  496,\n",
       "  478,\n",
       "  99,\n",
       "  6,\n",
       "  568,\n",
       "  2,\n",
       "  39,\n",
       "  314,\n",
       "  132,\n",
       "  364,\n",
       "  4,\n",
       "  28,\n",
       "  46,\n",
       "  42,\n",
       "  183,\n",
       "  2,\n",
       "  440,\n",
       "  58,\n",
       "  34,\n",
       "  1,\n",
       "  2227,\n",
       "  88,\n",
       "  1,\n",
       "  696,\n",
       "  474],\n",
       " [58,\n",
       "  3,\n",
       "  134,\n",
       "  92,\n",
       "  56,\n",
       "  311,\n",
       "  1362,\n",
       "  15,\n",
       "  2063,\n",
       "  3,\n",
       "  5532,\n",
       "  23,\n",
       "  1,\n",
       "  109,\n",
       "  2330,\n",
       "  53,\n",
       "  18,\n",
       "  485,\n",
       "  15,\n",
       "  25,\n",
       "  5,\n",
       "  3714,\n",
       "  4,\n",
       "  1310,\n",
       "  4,\n",
       "  9973,\n",
       "  3,\n",
       "  64,\n",
       "  1,\n",
       "  273,\n",
       "  83,\n",
       "  115,\n",
       "  12893,\n",
       "  35,\n",
       "  9,\n",
       "  44,\n",
       "  16,\n",
       "  41,\n",
       "  11,\n",
       "  1,\n",
       "  1469,\n",
       "  131,\n",
       "  1245,\n",
       "  2128,\n",
       "  418,\n",
       "  1245,\n",
       "  304,\n",
       "  3,\n",
       "  40,\n",
       "  20,\n",
       "  9,\n",
       "  485,\n",
       "  15,\n",
       "  23,\n",
       "  48,\n",
       "  47,\n",
       "  290,\n",
       "  68,\n",
       "  108,\n",
       "  23,\n",
       "  7,\n",
       "  17,\n",
       "  56,\n",
       "  4271,\n",
       "  13,\n",
       "  2227,\n",
       "  305,\n",
       "  15,\n",
       "  273,\n",
       "  25,\n",
       "  101,\n",
       "  2,\n",
       "  20,\n",
       "  10,\n",
       "  198,\n",
       "  51,\n",
       "  194,\n",
       "  39,\n",
       "  72,\n",
       "  36,\n",
       "  525,\n",
       "  2,\n",
       "  18,\n",
       "  5,\n",
       "  6073,\n",
       "  15,\n",
       "  45,\n",
       "  13,\n",
       "  367,\n",
       "  35,\n",
       "  3,\n",
       "  604,\n",
       "  36,\n",
       "  83,\n",
       "  18,\n",
       "  5,\n",
       "  2118,\n",
       "  369,\n",
       "  5,\n",
       "  2841,\n",
       "  2937,\n",
       "  71,\n",
       "  10,\n",
       "  420,\n",
       "  11,\n",
       "  1,\n",
       "  478,\n",
       "  597,\n",
       "  1397,\n",
       "  95,\n",
       "  3085,\n",
       "  5259],\n",
       " [7, 1690, 16, 28, 2776, 56, 1052, 7, 564, 37, 9, 15],\n",
       " [2728, 29, 39, 19, 96, 79, 1, 2348, 96, 46],\n",
       " [2525, 142, 7, 2077, 303, 15, 28, 135],\n",
       " [21,\n",
       "  208,\n",
       "  2,\n",
       "  1,\n",
       "  3581,\n",
       "  43,\n",
       "  57,\n",
       "  350,\n",
       "  51,\n",
       "  20,\n",
       "  12,\n",
       "  20,\n",
       "  10,\n",
       "  119,\n",
       "  25,\n",
       "  7,\n",
       "  33,\n",
       "  18,\n",
       "  727],\n",
       " [263,\n",
       "  23,\n",
       "  1,\n",
       "  252,\n",
       "  598,\n",
       "  24,\n",
       "  1406,\n",
       "  2,\n",
       "  7,\n",
       "  463,\n",
       "  3,\n",
       "  38,\n",
       "  12,\n",
       "  9317,\n",
       "  2,\n",
       "  332,\n",
       "  204,\n",
       "  11,\n",
       "  1,\n",
       "  1364,\n",
       "  44,\n",
       "  32,\n",
       "  2796,\n",
       "  15,\n",
       "  39,\n",
       "  13,\n",
       "  208,\n",
       "  3,\n",
       "  38,\n",
       "  1082,\n",
       "  1135,\n",
       "  9,\n",
       "  10,\n",
       "  18,\n",
       "  58,\n",
       "  1230,\n",
       "  35,\n",
       "  47,\n",
       "  40,\n",
       "  79,\n",
       "  10,\n",
       "  13,\n",
       "  450,\n",
       "  19,\n",
       "  5,\n",
       "  259,\n",
       "  3,\n",
       "  17,\n",
       "  57,\n",
       "  2,\n",
       "  1,\n",
       "  4881,\n",
       "  8619,\n",
       "  26,\n",
       "  10,\n",
       "  609,\n",
       "  5,\n",
       "  1275,\n",
       "  10,\n",
       "  652,\n",
       "  2,\n",
       "  2269,\n",
       "  8619,\n",
       "  55,\n",
       "  8,\n",
       "  5,\n",
       "  632,\n",
       "  7657,\n",
       "  9,\n",
       "  1003,\n",
       "  7,\n",
       "  48,\n",
       "  474,\n",
       "  36,\n",
       "  229,\n",
       "  18,\n",
       "  145,\n",
       "  303,\n",
       "  22,\n",
       "  3510,\n",
       "  11,\n",
       "  11416],\n",
       " [11625, 15, 14, 218, 6, 55, 16, 1893, 2, 133, 4, 78106],\n",
       " [281,\n",
       "  79,\n",
       "  906,\n",
       "  13,\n",
       "  65931,\n",
       "  90,\n",
       "  13,\n",
       "  1475,\n",
       "  65931,\n",
       "  3,\n",
       "  430,\n",
       "  1,\n",
       "  147,\n",
       "  2967,\n",
       "  9,\n",
       "  1,\n",
       "  147,\n",
       "  8,\n",
       "  86,\n",
       "  126,\n",
       "  173,\n",
       "  281,\n",
       "  79,\n",
       "  26,\n",
       "  36,\n",
       "  8,\n",
       "  48,\n",
       "  686,\n",
       "  25,\n",
       "  906,\n",
       "  19,\n",
       "  2,\n",
       "  76,\n",
       "  98,\n",
       "  79,\n",
       "  11,\n",
       "  45,\n",
       "  2640,\n",
       "  281,\n",
       "  79,\n",
       "  11,\n",
       "  764,\n",
       "  2,\n",
       "  1,\n",
       "  7834,\n",
       "  281,\n",
       "  79,\n",
       "  349,\n",
       "  7,\n",
       "  229,\n",
       "  65,\n",
       "  332,\n",
       "  74,\n",
       "  15,\n",
       "  1,\n",
       "  147,\n",
       "  684,\n",
       "  5,\n",
       "  464,\n",
       "  686,\n",
       "  25,\n",
       "  906,\n",
       "  13,\n",
       "  76,\n",
       "  197,\n",
       "  14,\n",
       "  147,\n",
       "  11,\n",
       "  409,\n",
       "  8,\n",
       "  1879,\n",
       "  22,\n",
       "  281,\n",
       "  79,\n",
       "  51,\n",
       "  107,\n",
       "  2,\n",
       "  1,\n",
       "  147,\n",
       "  684,\n",
       "  6,\n",
       "  77,\n",
       "  10,\n",
       "  2,\n",
       "  425,\n",
       "  5,\n",
       "  281,\n",
       "  79,\n",
       "  906,\n",
       "  23,\n",
       "  7,\n",
       "  17,\n",
       "  701,\n",
       "  61,\n",
       "  281,\n",
       "  79,\n",
       "  296,\n",
       "  394,\n",
       "  1409,\n",
       "  9,\n",
       "  7,\n",
       "  17,\n",
       "  2073,\n",
       "  1,\n",
       "  281,\n",
       "  79,\n",
       "  906,\n",
       "  15,\n",
       "  133,\n",
       "  45,\n",
       "  149,\n",
       "  7,\n",
       "  40,\n",
       "  129,\n",
       "  5,\n",
       "  153,\n",
       "  4,\n",
       "  147,\n",
       "  45,\n",
       "  7,\n",
       "  17,\n",
       "  592,\n",
       "  30,\n",
       "  1758,\n",
       "  15,\n",
       "  1,\n",
       "  28,\n",
       "  398,\n",
       "  158,\n",
       "  10,\n",
       "  8,\n",
       "  1772,\n",
       "  31,\n",
       "  1,\n",
       "  97,\n",
       "  308,\n",
       "  4,\n",
       "  56,\n",
       "  78,\n",
       "  7,\n",
       "  16,\n",
       "  1697,\n",
       "  11,\n",
       "  6,\n",
       "  80,\n",
       "  3163,\n",
       "  147,\n",
       "  29,\n",
       "  1,\n",
       "  3582,\n",
       "  872,\n",
       "  181,\n",
       "  9,\n",
       "  56,\n",
       "  281,\n",
       "  79,\n",
       "  344,\n",
       "  701,\n",
       "  143,\n",
       "  377,\n",
       "  83,\n",
       "  624,\n",
       "  6,\n",
       "  2534,\n",
       "  103,\n",
       "  27,\n",
       "  686,\n",
       "  33,\n",
       "  18,\n",
       "  140,\n",
       "  47,\n",
       "  702,\n",
       "  143,\n",
       "  44,\n",
       "  17,\n",
       "  57,\n",
       "  701,\n",
       "  19,\n",
       "  897,\n",
       "  15,\n",
       "  410,\n",
       "  13,\n",
       "  270,\n",
       "  113,\n",
       "  23,\n",
       "  7,\n",
       "  17,\n",
       "  56,\n",
       "  244,\n",
       "  51,\n",
       "  214,\n",
       "  84,\n",
       "  31,\n",
       "  1,\n",
       "  296,\n",
       "  249,\n",
       "  244,\n",
       "  128,\n",
       "  7,\n",
       "  46,\n",
       "  710,\n",
       "  6886,\n",
       "  112,\n",
       "  13,\n",
       "  65931,\n",
       "  90,\n",
       "  13,\n",
       "  1475,\n",
       "  65931,\n",
       "  3,\n",
       "  612,\n",
       "  9,\n",
       "  1,\n",
       "  736,\n",
       "  684,\n",
       "  507,\n",
       "  68,\n",
       "  12,\n",
       "  2129,\n",
       "  60,\n",
       "  322,\n",
       "  1,\n",
       "  166,\n",
       "  35,\n",
       "  1,\n",
       "  249,\n",
       "  663,\n",
       "  8,\n",
       "  2145,\n",
       "  23,\n",
       "  7,\n",
       "  67,\n",
       "  12,\n",
       "  447,\n",
       "  14,\n",
       "  736,\n",
       "  198,\n",
       "  80,\n",
       "  7,\n",
       "  33,\n",
       "  115,\n",
       "  2,\n",
       "  2129,\n",
       "  1,\n",
       "  2305,\n",
       "  4,\n",
       "  1,\n",
       "  249,\n",
       "  23,\n",
       "  7,\n",
       "  3671,\n",
       "  10,\n",
       "  29,\n",
       "  5,\n",
       "  387,\n",
       "  80,\n",
       "  5,\n",
       "  158,\n",
       "  2,\n",
       "  1,\n",
       "  387,\n",
       "  29,\n",
       "  55,\n",
       "  10,\n",
       "  24,\n",
       "  519,\n",
       "  785,\n",
       "  22,\n",
       "  5,\n",
       "  7553,\n",
       "  4,\n",
       "  9,\n",
       "  387,\n",
       "  625,\n",
       "  4,\n",
       "  79,\n",
       "  4,\n",
       "  98,\n",
       "  166,\n",
       "  8,\n",
       "  772,\n",
       "  1752,\n",
       "  99,\n",
       "  184,\n",
       "  23,\n",
       "  1,\n",
       "  249,\n",
       "  2178,\n",
       "  8,\n",
       "  254,\n",
       "  29,\n",
       "  1,\n",
       "  387,\n",
       "  2604,\n",
       "  80,\n",
       "  87,\n",
       "  249,\n",
       "  53,\n",
       "  65,\n",
       "  18,\n",
       "  3415,\n",
       "  19,\n",
       "  96,\n",
       "  19,\n",
       "  295,\n",
       "  1,\n",
       "  112,\n",
       "  51,\n",
       "  144,\n",
       "  5,\n",
       "  765,\n",
       "  249,\n",
       "  3672,\n",
       "  258,\n",
       "  23,\n",
       "  1,\n",
       "  736,\n",
       "  68,\n",
       "  12,\n",
       "  17,\n",
       "  47,\n",
       "  223,\n",
       "  23,\n",
       "  7,\n",
       "  322,\n",
       "  605,\n",
       "  1,\n",
       "  458,\n",
       "  2968,\n",
       "  25,\n",
       "  782,\n",
       "  80,\n",
       "  1,\n",
       "  258,\n",
       "  40,\n",
       "  18,\n",
       "  126,\n",
       "  2,\n",
       "  916,\n",
       "  10,\n",
       "  173,\n",
       "  1,\n",
       "  176,\n",
       "  1390,\n",
       "  794,\n",
       "  23,\n",
       "  7,\n",
       "  207,\n",
       "  1,\n",
       "  296,\n",
       "  1307,\n",
       "  1,\n",
       "  410,\n",
       "  31,\n",
       "  281,\n",
       "  79,\n",
       "  79,\n",
       "  5,\n",
       "  258,\n",
       "  103,\n",
       "  19,\n",
       "  25,\n",
       "  47,\n",
       "  4,\n",
       "  1,\n",
       "  61,\n",
       "  725,\n",
       "  420,\n",
       "  31,\n",
       "  147,\n",
       "  249,\n",
       "  725,\n",
       "  281,\n",
       "  79,\n",
       "  63,\n",
       "  147,\n",
       "  249,\n",
       "  725,\n",
       "  13,\n",
       "  1,\n",
       "  481,\n",
       "  153,\n",
       "  4,\n",
       "  249,\n",
       "  725,\n",
       "  9,\n",
       "  7,\n",
       "  40,\n",
       "  79,\n",
       "  23,\n",
       "  7,\n",
       "  17,\n",
       "  701,\n",
       "  61,\n",
       "  1464,\n",
       "  394,\n",
       "  1409,\n",
       "  9,\n",
       "  7,\n",
       "  17,\n",
       "  2073,\n",
       "  87,\n",
       "  112,\n",
       "  6,\n",
       "  857,\n",
       "  84,\n",
       "  149,\n",
       "  7,\n",
       "  40,\n",
       "  129,\n",
       "  5,\n",
       "  153,\n",
       "  4,\n",
       "  1464,\n",
       "  7,\n",
       "  17,\n",
       "  701,\n",
       "  30,\n",
       "  427,\n",
       "  14,\n",
       "  158,\n",
       "  1208,\n",
       "  6,\n",
       "  4702,\n",
       "  344,\n",
       "  83,\n",
       "  18,\n",
       "  140,\n",
       "  47,\n",
       "  702,\n",
       "  143,\n",
       "  44,\n",
       "  17,\n",
       "  57,\n",
       "  857,\n",
       "  19,\n",
       "  897,\n",
       "  15,\n",
       "  410,\n",
       "  13,\n",
       "  270,\n",
       "  113,\n",
       "  23,\n",
       "  1,\n",
       "  147,\n",
       "  8,\n",
       "  1444,\n",
       "  173,\n",
       "  5,\n",
       "  262,\n",
       "  176,\n",
       "  794,\n",
       "  466,\n",
       "  281,\n",
       "  79,\n",
       "  80,\n",
       "  1,\n",
       "  147,\n",
       "  33,\n",
       "  18,\n",
       "  140,\n",
       "  2324,\n",
       "  808,\n",
       "  143,\n",
       "  23,\n",
       "  7,\n",
       "  17,\n",
       "  56,\n",
       "  244,\n",
       "  51,\n",
       "  214,\n",
       "  84,\n",
       "  31,\n",
       "  1,\n",
       "  296,\n",
       "  249,\n",
       "  244,\n",
       "  128,\n",
       "  7,\n",
       "  46,\n",
       "  710],\n",
       " [22499, 18, 5, 363, 6, 2141, 530, 10, 335, 139, 1, 2682],\n",
       " [429,\n",
       "  37,\n",
       "  8,\n",
       "  10,\n",
       "  46,\n",
       "  37,\n",
       "  8,\n",
       "  10,\n",
       "  27,\n",
       "  4228,\n",
       "  412,\n",
       "  4,\n",
       "  59,\n",
       "  683,\n",
       "  47220,\n",
       "  60,\n",
       "  16,\n",
       "  95,\n",
       "  31,\n",
       "  5519,\n",
       "  535,\n",
       "  4558,\n",
       "  47221,\n",
       "  60,\n",
       "  3495,\n",
       "  71,\n",
       "  56,\n",
       "  47,\n",
       "  60,\n",
       "  5172,\n",
       "  84,\n",
       "  244,\n",
       "  20460,\n",
       "  87,\n",
       "  528,\n",
       "  1154,\n",
       "  6,\n",
       "  5789,\n",
       "  262,\n",
       "  1308,\n",
       "  31,\n",
       "  683,\n",
       "  214,\n",
       "  78107,\n",
       "  2,\n",
       "  1479,\n",
       "  71,\n",
       "  66,\n",
       "  954,\n",
       "  88,\n",
       "  251,\n",
       "  39,\n",
       "  6246,\n",
       "  1354],\n",
       " [142,\n",
       "  7,\n",
       "  404,\n",
       "  3938,\n",
       "  1418,\n",
       "  6,\n",
       "  1354,\n",
       "  31,\n",
       "  39,\n",
       "  2141,\n",
       "  367,\n",
       "  1,\n",
       "  77,\n",
       "  508,\n",
       "  271,\n",
       "  1535,\n",
       "  4825,\n",
       "  582,\n",
       "  8,\n",
       "  12,\n",
       "  155,\n",
       "  2,\n",
       "  13154,\n",
       "  21,\n",
       "  581,\n",
       "  10,\n",
       "  33,\n",
       "  1082,\n",
       "  92,\n",
       "  10,\n",
       "  138,\n",
       "  49,\n",
       "  7,\n",
       "  16,\n",
       "  2969,\n",
       "  21,\n",
       "  616,\n",
       "  19,\n",
       "  27,\n",
       "  395,\n",
       "  85,\n",
       "  1,\n",
       "  77,\n",
       "  508,\n",
       "  8,\n",
       "  478,\n",
       "  14,\n",
       "  8,\n",
       "  266,\n",
       "  1,\n",
       "  494,\n",
       "  125,\n",
       "  3294,\n",
       "  34,\n",
       "  1095,\n",
       "  8566,\n",
       "  42,\n",
       "  514,\n",
       "  19,\n",
       "  4,\n",
       "  1293,\n",
       "  66,\n",
       "  3648,\n",
       "  8,\n",
       "  320,\n",
       "  136,\n",
       "  42,\n",
       "  8,\n",
       "  1,\n",
       "  75,\n",
       "  791,\n",
       "  2949,\n",
       "  1195,\n",
       "  60,\n",
       "  67,\n",
       "  12,\n",
       "  5160,\n",
       "  9,\n",
       "  517,\n",
       "  58,\n",
       "  320,\n",
       "  88,\n",
       "  66,\n",
       "  32566,\n",
       "  27,\n",
       "  7009,\n",
       "  3,\n",
       "  2641,\n",
       "  2,\n",
       "  384,\n",
       "  14,\n",
       "  77,\n",
       "  11,\n",
       "  5891,\n",
       "  4,\n",
       "  15855,\n",
       "  1,\n",
       "  753,\n",
       "  4,\n",
       "  27,\n",
       "  395,\n",
       "  9,\n",
       "  8,\n",
       "  1263,\n",
       "  2,\n",
       "  138,\n",
       "  31,\n",
       "  1,\n",
       "  251,\n",
       "  508,\n",
       "  6,\n",
       "  12,\n",
       "  2470,\n",
       "  1418,\n",
       "  303,\n",
       "  353,\n",
       "  35,\n",
       "  15856,\n",
       "  354,\n",
       "  23,\n",
       "  7,\n",
       "  3290,\n",
       "  5,\n",
       "  781,\n",
       "  4,\n",
       "  2020,\n",
       "  117,\n",
       "  7,\n",
       "  40,\n",
       "  20,\n",
       "  14,\n",
       "  52,\n",
       "  40,\n",
       "  17,\n",
       "  5,\n",
       "  3883,\n",
       "  130,\n",
       "  15,\n",
       "  1,\n",
       "  424,\n",
       "  6,\n",
       "  2310,\n",
       "  1,\n",
       "  338,\n",
       "  16066],\n",
       " [396,\n",
       "  6,\n",
       "  1,\n",
       "  1539,\n",
       "  238,\n",
       "  492,\n",
       "  164,\n",
       "  965,\n",
       "  22,\n",
       "  39,\n",
       "  163,\n",
       "  2584,\n",
       "  164,\n",
       "  3945,\n",
       "  117,\n",
       "  10,\n",
       "  68,\n",
       "  12,\n",
       "  1265,\n",
       "  3,\n",
       "  207,\n",
       "  1,\n",
       "  581,\n",
       "  24,\n",
       "  313,\n",
       "  39,\n",
       "  6,\n",
       "  26,\n",
       "  49,\n",
       "  3,\n",
       "  157,\n",
       "  1,\n",
       "  883,\n",
       "  24,\n",
       "  3487,\n",
       "  6,\n",
       "  3,\n",
       "  6948,\n",
       "  90],\n",
       " [34158,\n",
       "  1051,\n",
       "  11,\n",
       "  2401,\n",
       "  34158,\n",
       "  19627,\n",
       "  24,\n",
       "  871,\n",
       "  255,\n",
       "  324,\n",
       "  80,\n",
       "  550,\n",
       "  1225,\n",
       "  5533,\n",
       "  55,\n",
       "  378,\n",
       "  34158,\n",
       "  1324,\n",
       "  921,\n",
       "  271,\n",
       "  1742,\n",
       "  22,\n",
       "  1,\n",
       "  23127,\n",
       "  1,\n",
       "  735,\n",
       "  43605,\n",
       "  2,\n",
       "  18,\n",
       "  2100,\n",
       "  2,\n",
       "  10403,\n",
       "  1856,\n",
       "  173,\n",
       "  6765,\n",
       "  5,\n",
       "  17194,\n",
       "  11,\n",
       "  1865,\n",
       "  42,\n",
       "  24,\n",
       "  596,\n",
       "  255,\n",
       "  324,\n",
       "  938,\n",
       "  74,\n",
       "  22,\n",
       "  66,\n",
       "  159,\n",
       "  3571,\n",
       "  19627,\n",
       "  1536,\n",
       "  6,\n",
       "  302,\n",
       "  35,\n",
       "  275,\n",
       "  42,\n",
       "  8,\n",
       "  851,\n",
       "  11,\n",
       "  6151,\n",
       "  42,\n",
       "  118,\n",
       "  8,\n",
       "  73,\n",
       "  93,\n",
       "  42,\n",
       "  18,\n",
       "  2155,\n",
       "  80,\n",
       "  9036,\n",
       "  6247,\n",
       "  6,\n",
       "  73,\n",
       "  93,\n",
       "  42,\n",
       "  18,\n",
       "  898,\n",
       "  78,\n",
       "  66,\n",
       "  3904,\n",
       "  2089,\n",
       "  1,\n",
       "  15596,\n",
       "  43605,\n",
       "  8,\n",
       "  967,\n",
       "  255,\n",
       "  324,\n",
       "  6151,\n",
       "  624,\n",
       "  34158,\n",
       "  649,\n",
       "  462,\n",
       "  9379,\n",
       "  23,\n",
       "  21,\n",
       "  891,\n",
       "  34,\n",
       "  9,\n",
       "  7515,\n",
       "  967,\n",
       "  107,\n",
       "  2,\n",
       "  21,\n",
       "  6,\n",
       "  161,\n",
       "  888,\n",
       "  66,\n",
       "  388,\n",
       "  4,\n",
       "  1622,\n",
       "  28,\n",
       "  462],\n",
       " [2500, 20, 12, 138, 253, 25, 64, 4, 17821, 151, 15180],\n",
       " [415, 46, 1728, 51733, 98417],\n",
       " [1, 114, 124, 48, 442, 76, 12, 1297, 2, 425, 4617, 15, 43606, 2, 425, 58, 99],\n",
       " [20,\n",
       "  12,\n",
       "  257,\n",
       "  2,\n",
       "  1573,\n",
       "  7,\n",
       "  3,\n",
       "  63,\n",
       "  9,\n",
       "  7,\n",
       "  16,\n",
       "  521,\n",
       "  154,\n",
       "  431,\n",
       "  496,\n",
       "  204,\n",
       "  629,\n",
       "  62,\n",
       "  6,\n",
       "  23,\n",
       "  7,\n",
       "  20,\n",
       "  396,\n",
       "  96,\n",
       "  26,\n",
       "  23,\n",
       "  12,\n",
       "  6,\n",
       "  7,\n",
       "  40,\n",
       "  40669,\n",
       "  530,\n",
       "  14,\n",
       "  22,\n",
       "  39,\n",
       "  80,\n",
       "  91,\n",
       "  188,\n",
       "  3,\n",
       "  32,\n",
       "  49,\n",
       "  2,\n",
       "  214,\n",
       "  7,\n",
       "  2,\n",
       "  152,\n",
       "  5,\n",
       "  2771,\n",
       "  138,\n",
       "  31,\n",
       "  1,\n",
       "  7769,\n",
       "  3683,\n",
       "  3851,\n",
       "  25952,\n",
       "  6,\n",
       "  1,\n",
       "  901,\n",
       "  420,\n",
       "  11,\n",
       "  10,\n",
       "  1588,\n",
       "  105,\n",
       "  901,\n",
       "  1265,\n",
       "  785,\n",
       "  11,\n",
       "  59,\n",
       "  25952,\n",
       "  8,\n",
       "  36,\n",
       "  204,\n",
       "  9,\n",
       "  7,\n",
       "  64,\n",
       "  52,\n",
       "  40,\n",
       "  20,\n",
       "  22,\n",
       "  1,\n",
       "  25952,\n",
       "  1495,\n",
       "  98418,\n",
       "  10]]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_sequences[:20]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading the pre-trained word embedding.\n",
      "Total 2000000 word vectors.\n",
      "Null word embeddings: 22573\n",
      "Loaded\n"
     ]
    }
   ],
   "source": [
    "print(\"Loading the pre-trained word embedding.\")\n",
    "embeddings_index = data_loader.load_embedding('features/crawl-300d-2M.vec')\n",
    "embedding_matrix = data_transformer.build_embedding_matrix(embeddings_index)\n",
    "print(\"Loaded\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Build model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<module 'sotoxic.train.trainer' from '/home/zake7749/PycharmProjects/CCP/sotoxic/train/trainer.py'>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import importlib\n",
    "import sotoxic.models.pytorch.bgru as bgru\n",
    "import sotoxic.train.trainer as trn\n",
    "importlib.reload(bgru)\n",
    "importlib.reload(trn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def get_bgru_network():\n",
    "    embedding = nn.Embedding(VOCAB_SIZE, EMBEDDING_SIZE)\n",
    "    embedding.weight.data.copy_(torch.from_numpy(embedding_matrix))\n",
    "    embedding.weight.requires_grad=False\n",
    "    return bgru.BayesianGRUClassifier(input_size=EMBEDDING_SIZE, hidden_size=60, embedding=embedding)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "trainer = trn.PyTorchModelTrainer(model_stamp=\"FASTTXT_BGRU_64_64\", epoch_num=300, learning_rate=1e-3,\n",
    "                                  verbose_round=40, shuffle_inputs=False, early_stopping_round=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Choose the torch base model.\n",
      "Training on fold 0\n",
      "Epoch:1 Batch:0 Log-loss0.7120054364204407\n",
      "Epoch:1 Batch:40 Log-loss0.1323106735944748\n",
      "Epoch:1 Batch:80 Log-loss0.07861244678497314\n",
      "Epoch:1 Batch:120 Log-loss0.050647079944610596\n",
      "Epoch:1 Batch:160 Log-loss0.04874030873179436\n",
      "Epoch:1 Batch:200 Log-loss0.05985138937830925\n",
      "Epoch:1 Batch:240 Log-loss0.037754934281110764\n",
      "Epoch:1 Batch:280 Log-loss0.050901126116514206\n",
      "Epoch:1 Batch:320 Log-loss0.06970404833555222\n",
      "Epoch:1 Batch:360 Log-loss0.06141173839569092\n",
      "Epoch:1 Batch:400 Log-loss0.04935069754719734\n",
      "Epoch:1 Batch:440 Log-loss0.0667911097407341\n",
      "Epoch:1 Batch:480 Log-loss0.04504765197634697\n",
      "Epoch:1 Batch:520 Log-loss0.055000707507133484\n",
      "Epoch:1 Batch:560 Log-loss0.036638546735048294\n",
      "Epoch average log loss:0.0793381747856204\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch1, val_loss:0.04555952425826937, best_val_loss:0.04555952425826937, best_auc:0.976878537666061\n",
      "Epoch:2 Batch:0 Log-loss0.06867814809083939\n",
      "Epoch:2 Batch:40 Log-loss0.03992528095841408\n",
      "Epoch:2 Batch:80 Log-loss0.046969298273324966\n",
      "Epoch:2 Batch:120 Log-loss0.042035061866045\n",
      "Epoch:2 Batch:160 Log-loss0.041205327957868576\n",
      "Epoch:2 Batch:200 Log-loss0.049056291580200195\n",
      "Epoch:2 Batch:240 Log-loss0.07705309242010117\n",
      "Epoch:2 Batch:280 Log-loss0.04704364761710167\n",
      "Epoch:2 Batch:320 Log-loss0.0487460158765316\n",
      "Epoch:2 Batch:360 Log-loss0.05659325793385506\n",
      "Epoch:2 Batch:400 Log-loss0.057180922478437424\n",
      "Epoch:2 Batch:440 Log-loss0.056150536984205246\n",
      "Epoch:2 Batch:480 Log-loss0.03524068742990494\n",
      "Epoch:2 Batch:520 Log-loss0.04137372598052025\n",
      "Epoch:2 Batch:560 Log-loss0.05214753374457359\n",
      "Epoch average log loss:0.05001075967281524\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch2, val_loss:0.04434937140227999, best_val_loss:0.04434937140227999, best_auc:0.9804736220542257\n",
      "Epoch:3 Batch:0 Log-loss0.04830276966094971\n",
      "Epoch:3 Batch:40 Log-loss0.04096377268433571\n",
      "Epoch:3 Batch:80 Log-loss0.050506722182035446\n",
      "Epoch:3 Batch:120 Log-loss0.04582715407013893\n",
      "Epoch:3 Batch:160 Log-loss0.03807763382792473\n",
      "Epoch:3 Batch:200 Log-loss0.05335954949259758\n",
      "Epoch:3 Batch:240 Log-loss0.04983651638031006\n",
      "Epoch:3 Batch:280 Log-loss0.04252280667424202\n",
      "Epoch:3 Batch:320 Log-loss0.06060698255896568\n",
      "Epoch:3 Batch:360 Log-loss0.04445343092083931\n",
      "Epoch:3 Batch:400 Log-loss0.0501587875187397\n",
      "Epoch:3 Batch:440 Log-loss0.06167474016547203\n",
      "Epoch:3 Batch:480 Log-loss0.0546649731695652\n",
      "Epoch:3 Batch:520 Log-loss0.03097853623330593\n",
      "Epoch:3 Batch:560 Log-loss0.05456440895795822\n",
      "Epoch average log loss:0.04741293806250074\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch3, val_loss:0.04215973133101202, best_val_loss:0.04215973133101202, best_auc:0.9831964330014121\n",
      "Epoch:4 Batch:0 Log-loss0.044340312480926514\n",
      "Epoch:4 Batch:40 Log-loss0.029539242386817932\n",
      "Epoch:4 Batch:80 Log-loss0.04329937323927879\n",
      "Epoch:4 Batch:120 Log-loss0.052374426275491714\n",
      "Epoch:4 Batch:160 Log-loss0.04566134512424469\n",
      "Epoch:4 Batch:200 Log-loss0.06970560550689697\n",
      "Epoch:4 Batch:240 Log-loss0.057286981493234634\n",
      "Epoch:4 Batch:280 Log-loss0.06771456450223923\n",
      "Epoch:4 Batch:320 Log-loss0.041504886001348495\n",
      "Epoch:4 Batch:360 Log-loss0.05768328905105591\n",
      "Epoch:4 Batch:400 Log-loss0.0453309565782547\n",
      "Epoch:4 Batch:440 Log-loss0.03599371388554573\n",
      "Epoch:4 Batch:480 Log-loss0.048371702432632446\n",
      "Epoch:4 Batch:520 Log-loss0.06484667211771011\n",
      "Epoch:4 Batch:560 Log-loss0.03898661956191063\n",
      "Epoch average log loss:0.04605169754275786\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch4, val_loss:0.04104461211755689, best_val_loss:0.04104461211755689, best_auc:0.9848249535806568\n",
      "Epoch:5 Batch:0 Log-loss0.037953298538923264\n",
      "Epoch:5 Batch:40 Log-loss0.05954471230506897\n",
      "Epoch:5 Batch:80 Log-loss0.03514011576771736\n",
      "Epoch:5 Batch:120 Log-loss0.07466626912355423\n",
      "Epoch:5 Batch:160 Log-loss0.0445808507502079\n",
      "Epoch:5 Batch:200 Log-loss0.05395655706524849\n",
      "Epoch:5 Batch:240 Log-loss0.05939869210124016\n",
      "Epoch:5 Batch:280 Log-loss0.0463140495121479\n",
      "Epoch:5 Batch:320 Log-loss0.039417777210474014\n",
      "Epoch:5 Batch:360 Log-loss0.03911064937710762\n",
      "Epoch:5 Batch:400 Log-loss0.04967992380261421\n",
      "Epoch:5 Batch:440 Log-loss0.04370224103331566\n",
      "Epoch:5 Batch:480 Log-loss0.044944316148757935\n",
      "Epoch:5 Batch:520 Log-loss0.03316333889961243\n",
      "Epoch:5 Batch:560 Log-loss0.03341749310493469\n",
      "Epoch average log loss:0.04478505265433341\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch5, val_loss:0.04076002593755801, best_val_loss:0.04076002593755801, best_auc:0.9853210580933728\n",
      "Epoch:6 Batch:0 Log-loss0.035939380526542664\n",
      "Epoch:6 Batch:40 Log-loss0.055933430790901184\n",
      "Epoch:6 Batch:80 Log-loss0.04986649751663208\n",
      "Epoch:6 Batch:120 Log-loss0.05274675413966179\n",
      "Epoch:6 Batch:160 Log-loss0.07919551432132721\n",
      "Epoch:6 Batch:200 Log-loss0.06013499200344086\n",
      "Epoch:6 Batch:240 Log-loss0.03523259237408638\n",
      "Epoch:6 Batch:280 Log-loss0.03885224089026451\n",
      "Epoch:6 Batch:320 Log-loss0.047836143523454666\n",
      "Epoch:6 Batch:360 Log-loss0.03278129920363426\n",
      "Epoch:6 Batch:400 Log-loss0.04611663892865181\n",
      "Epoch:6 Batch:440 Log-loss0.04569028690457344\n",
      "Epoch:6 Batch:480 Log-loss0.04447239264845848\n",
      "Epoch:6 Batch:520 Log-loss0.04127364605665207\n",
      "Epoch:6 Batch:560 Log-loss0.029941245913505554\n",
      "Epoch average log loss:0.043567589760225796\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch6, val_loss:0.040026577256412024, best_val_loss:0.040026577256412024, best_auc:0.9863640882951649\n",
      "Epoch:7 Batch:0 Log-loss0.03992745652794838\n",
      "Epoch:7 Batch:40 Log-loss0.034982237964868546\n",
      "Epoch:7 Batch:80 Log-loss0.0431690514087677\n",
      "Epoch:7 Batch:120 Log-loss0.030462026596069336\n",
      "Epoch:7 Batch:160 Log-loss0.04396167770028114\n",
      "Epoch:7 Batch:200 Log-loss0.04974162206053734\n",
      "Epoch:7 Batch:240 Log-loss0.05741676315665245\n",
      "Epoch:7 Batch:280 Log-loss0.06804589182138443\n",
      "Epoch:7 Batch:320 Log-loss0.04352782294154167\n",
      "Epoch:7 Batch:360 Log-loss0.047447506338357925\n",
      "Epoch:7 Batch:400 Log-loss0.04978137090802193\n",
      "Epoch:7 Batch:440 Log-loss0.05918342247605324\n",
      "Epoch:7 Batch:480 Log-loss0.03971332311630249\n",
      "Epoch:7 Batch:520 Log-loss0.03241755813360214\n",
      "Epoch:7 Batch:560 Log-loss0.02896280400454998\n",
      "Epoch average log loss:0.04303903625612812\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch7, val_loss:0.039330826899868575, best_val_loss:0.039330826899868575, best_auc:0.9868008643454426\n",
      "Epoch:8 Batch:0 Log-loss0.05936514213681221\n",
      "Epoch:8 Batch:40 Log-loss0.0302696842700243\n",
      "Epoch:8 Batch:80 Log-loss0.04756798967719078\n",
      "Epoch:8 Batch:120 Log-loss0.04152555391192436\n",
      "Epoch:8 Batch:160 Log-loss0.06145668029785156\n",
      "Epoch:8 Batch:200 Log-loss0.051588546484708786\n",
      "Epoch:8 Batch:240 Log-loss0.03861844167113304\n",
      "Epoch:8 Batch:280 Log-loss0.05779844522476196\n",
      "Epoch:8 Batch:320 Log-loss0.032732538878917694\n",
      "Epoch:8 Batch:360 Log-loss0.03579225391149521\n",
      "Epoch:8 Batch:400 Log-loss0.041509777307510376\n",
      "Epoch:8 Batch:440 Log-loss0.04704805091023445\n",
      "Epoch:8 Batch:480 Log-loss0.038060083985328674\n",
      "Epoch:8 Batch:520 Log-loss0.04304433986544609\n",
      "Epoch:8 Batch:560 Log-loss0.05566395819187164\n",
      "Epoch average log loss:0.04216378396709582\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch8, val_loss:0.03889493152577631, best_val_loss:0.03889493152577631, best_auc:0.9878324618173928\n",
      "Epoch:9 Batch:0 Log-loss0.0571688674390316\n",
      "Epoch:9 Batch:40 Log-loss0.04262637719511986\n",
      "Epoch:9 Batch:80 Log-loss0.02655140310525894\n",
      "Epoch:9 Batch:120 Log-loss0.043560054153203964\n",
      "Epoch:9 Batch:160 Log-loss0.05560952052474022\n",
      "Epoch:9 Batch:200 Log-loss0.04212874546647072\n",
      "Epoch:9 Batch:240 Log-loss0.02979004569351673\n",
      "Epoch:9 Batch:280 Log-loss0.03245263919234276\n",
      "Epoch:9 Batch:320 Log-loss0.026346318423748016\n",
      "Epoch:9 Batch:360 Log-loss0.040751054883003235\n",
      "Epoch:9 Batch:400 Log-loss0.02844347059726715\n",
      "Epoch:9 Batch:440 Log-loss0.037340354174375534\n",
      "Epoch:9 Batch:480 Log-loss0.03828814998269081\n",
      "Epoch:9 Batch:520 Log-loss0.05120089650154114\n",
      "Epoch:9 Batch:560 Log-loss0.03552517294883728\n",
      "Epoch average log loss:0.041780799171621245\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch9, val_loss:0.038531173758827544, best_val_loss:0.038531173758827544, best_auc:0.9882740412030578\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:10 Batch:0 Log-loss0.03695434331893921\n",
      "Epoch:10 Batch:40 Log-loss0.04090544953942299\n",
      "Epoch:10 Batch:80 Log-loss0.05570065602660179\n",
      "Epoch:10 Batch:120 Log-loss0.03997639939188957\n",
      "Epoch:10 Batch:160 Log-loss0.04213264212012291\n",
      "Epoch:10 Batch:200 Log-loss0.03962920978665352\n",
      "Epoch:10 Batch:240 Log-loss0.040476951748132706\n",
      "Epoch:10 Batch:280 Log-loss0.04190034046769142\n",
      "Epoch:10 Batch:320 Log-loss0.044796139001846313\n",
      "Epoch:10 Batch:360 Log-loss0.03308139368891716\n",
      "Epoch:10 Batch:400 Log-loss0.051052410155534744\n",
      "Epoch:10 Batch:440 Log-loss0.0357748419046402\n",
      "Epoch:10 Batch:480 Log-loss0.032477814704179764\n",
      "Epoch:10 Batch:520 Log-loss0.05190897360444069\n",
      "Epoch:10 Batch:560 Log-loss0.03744189813733101\n",
      "Epoch average log loss:0.04090871061531028\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch10, val_loss:0.03803986401146478, best_val_loss:0.03803986401146478, best_auc:0.9886834264795996\n",
      "Epoch:11 Batch:0 Log-loss0.03530322387814522\n",
      "Epoch:11 Batch:40 Log-loss0.03929544612765312\n",
      "Epoch:11 Batch:80 Log-loss0.04489068314433098\n",
      "Epoch:11 Batch:120 Log-loss0.034296635538339615\n",
      "Epoch:11 Batch:160 Log-loss0.035722535103559494\n",
      "Epoch:11 Batch:200 Log-loss0.04076075181365013\n",
      "Epoch:11 Batch:240 Log-loss0.027380459010601044\n",
      "Epoch:11 Batch:280 Log-loss0.05594487488269806\n",
      "Epoch:11 Batch:320 Log-loss0.04280964657664299\n",
      "Epoch:11 Batch:360 Log-loss0.040003590285778046\n",
      "Epoch:11 Batch:400 Log-loss0.048115745186805725\n",
      "Epoch:11 Batch:440 Log-loss0.028943320736289024\n",
      "Epoch:11 Batch:480 Log-loss0.0420382060110569\n",
      "Epoch:11 Batch:520 Log-loss0.0317797027528286\n",
      "Epoch:11 Batch:560 Log-loss0.052402906119823456\n",
      "Epoch average log loss:0.04069471282418817\n",
      "In Epoch11, val_loss:0.03819792237609317, best_val_loss:0.03803986401146478, best_auc:0.9886834264795996\n",
      "Epoch:12 Batch:0 Log-loss0.0224236398935318\n",
      "Epoch:12 Batch:40 Log-loss0.04290938749909401\n",
      "Epoch:12 Batch:80 Log-loss0.03902919963002205\n",
      "Epoch:12 Batch:120 Log-loss0.04078105464577675\n",
      "Epoch:12 Batch:160 Log-loss0.03167366608977318\n",
      "Epoch:12 Batch:200 Log-loss0.04649972915649414\n",
      "Epoch:12 Batch:240 Log-loss0.049783289432525635\n",
      "Epoch:12 Batch:280 Log-loss0.0434969998896122\n",
      "Epoch:12 Batch:320 Log-loss0.02270817570388317\n",
      "Epoch:12 Batch:360 Log-loss0.041368432343006134\n",
      "Epoch:12 Batch:400 Log-loss0.034349750727415085\n",
      "Epoch:12 Batch:440 Log-loss0.03214189037680626\n",
      "Epoch:12 Batch:480 Log-loss0.026861103251576424\n",
      "Epoch:12 Batch:520 Log-loss0.0471443235874176\n",
      "Epoch:12 Batch:560 Log-loss0.03870300576090813\n",
      "Epoch average log loss:0.040244564058126084\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch12, val_loss:0.03772386062401634, best_val_loss:0.03772386062401634, best_auc:0.9891847812434168\n",
      "Epoch:13 Batch:0 Log-loss0.060291845351457596\n",
      "Epoch:13 Batch:40 Log-loss0.04330018162727356\n",
      "Epoch:13 Batch:80 Log-loss0.044248636811971664\n",
      "Epoch:13 Batch:120 Log-loss0.032636284828186035\n",
      "Epoch:13 Batch:160 Log-loss0.045553743839263916\n",
      "Epoch:13 Batch:200 Log-loss0.02872871421277523\n",
      "Epoch:13 Batch:240 Log-loss0.04538087174296379\n",
      "Epoch:13 Batch:280 Log-loss0.03926897794008255\n",
      "Epoch:13 Batch:320 Log-loss0.045715559273958206\n",
      "Epoch:13 Batch:360 Log-loss0.04150130972266197\n",
      "Epoch:13 Batch:400 Log-loss0.03738865256309509\n",
      "Epoch:13 Batch:440 Log-loss0.03178480267524719\n",
      "Epoch:13 Batch:480 Log-loss0.04769943282008171\n",
      "Epoch:13 Batch:520 Log-loss0.02698323130607605\n",
      "Epoch:13 Batch:560 Log-loss0.040955834090709686\n",
      "Epoch average log loss:0.039668001653626564\n",
      "In Epoch13, val_loss:0.0377531890928375, best_val_loss:0.03772386062401634, best_auc:0.9891847812434168\n",
      "Epoch:14 Batch:0 Log-loss0.03509269654750824\n",
      "Epoch:14 Batch:40 Log-loss0.030732452869415283\n",
      "Epoch:14 Batch:80 Log-loss0.050405409187078476\n",
      "Epoch:14 Batch:120 Log-loss0.045347388833761215\n",
      "Epoch:14 Batch:160 Log-loss0.040238894522190094\n",
      "Epoch:14 Batch:200 Log-loss0.04657888785004616\n",
      "Epoch:14 Batch:240 Log-loss0.04376591369509697\n",
      "Epoch:14 Batch:280 Log-loss0.03076501004397869\n",
      "Epoch:14 Batch:320 Log-loss0.05770501121878624\n",
      "Epoch:14 Batch:360 Log-loss0.0357428602874279\n",
      "Epoch:14 Batch:400 Log-loss0.054260771721601486\n",
      "Epoch:14 Batch:440 Log-loss0.03408332169055939\n",
      "Epoch:14 Batch:480 Log-loss0.04128219559788704\n",
      "Epoch:14 Batch:520 Log-loss0.028443820774555206\n",
      "Epoch:14 Batch:560 Log-loss0.03651941940188408\n",
      "Epoch average log loss:0.039487611149836864\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch14, val_loss:0.03770914891211926, best_val_loss:0.03770914891211926, best_auc:0.9891352863867117\n",
      "Epoch:15 Batch:0 Log-loss0.04915967956185341\n",
      "Epoch:15 Batch:40 Log-loss0.03083048202097416\n",
      "Epoch:15 Batch:80 Log-loss0.037839557975530624\n",
      "Epoch:15 Batch:120 Log-loss0.04802577197551727\n",
      "Epoch:15 Batch:160 Log-loss0.03994175046682358\n",
      "Epoch:15 Batch:200 Log-loss0.0371636338531971\n",
      "Epoch:15 Batch:240 Log-loss0.03508050739765167\n",
      "Epoch:15 Batch:280 Log-loss0.026659322902560234\n",
      "Epoch:15 Batch:320 Log-loss0.03200378641486168\n",
      "Epoch:15 Batch:360 Log-loss0.04112040624022484\n",
      "Epoch:15 Batch:400 Log-loss0.036452144384384155\n",
      "Epoch:15 Batch:440 Log-loss0.04047171398997307\n",
      "Epoch:15 Batch:480 Log-loss0.03129281848669052\n",
      "Epoch:15 Batch:520 Log-loss0.05017142370343208\n",
      "Epoch:15 Batch:560 Log-loss0.036765579134225845\n",
      "Epoch average log loss:0.039163759788165664\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch15, val_loss:0.03769336894943021, best_val_loss:0.03769336894943021, best_auc:0.9892264731964048\n",
      "Epoch:16 Batch:0 Log-loss0.033707521855831146\n",
      "Epoch:16 Batch:40 Log-loss0.04967847466468811\n",
      "Epoch:16 Batch:80 Log-loss0.061272840946912766\n",
      "Epoch:16 Batch:120 Log-loss0.0377248115837574\n",
      "Epoch:16 Batch:160 Log-loss0.033359382301568985\n",
      "Epoch:16 Batch:200 Log-loss0.05108053609728813\n",
      "Epoch:16 Batch:240 Log-loss0.02579716593027115\n",
      "Epoch:16 Batch:280 Log-loss0.034421276301145554\n",
      "Epoch:16 Batch:320 Log-loss0.04456759989261627\n",
      "Epoch:16 Batch:360 Log-loss0.04775548353791237\n",
      "Epoch:16 Batch:400 Log-loss0.04677410423755646\n",
      "Epoch:16 Batch:440 Log-loss0.04755590856075287\n",
      "Epoch:16 Batch:480 Log-loss0.03218359872698784\n",
      "Epoch:16 Batch:520 Log-loss0.04928584396839142\n",
      "Epoch:16 Batch:560 Log-loss0.03954445943236351\n",
      "Epoch average log loss:0.03911433703864792\n",
      "In Epoch16, val_loss:0.03793768018918047, best_val_loss:0.03769336894943021, best_auc:0.9892264731964048\n",
      "Epoch:17 Batch:0 Log-loss0.04665709659457207\n",
      "Epoch:17 Batch:40 Log-loss0.030883660539984703\n",
      "Epoch:17 Batch:80 Log-loss0.029673494398593903\n",
      "Epoch:17 Batch:120 Log-loss0.02570403181016445\n",
      "Epoch:17 Batch:160 Log-loss0.03809676691889763\n",
      "Epoch:17 Batch:200 Log-loss0.03503228351473808\n",
      "Epoch:17 Batch:240 Log-loss0.04171283543109894\n",
      "Epoch:17 Batch:280 Log-loss0.03663165494799614\n",
      "Epoch:17 Batch:320 Log-loss0.05492253601551056\n",
      "Epoch:17 Batch:360 Log-loss0.03774653747677803\n",
      "Epoch:17 Batch:400 Log-loss0.03281967341899872\n",
      "Epoch:17 Batch:440 Log-loss0.04743754491209984\n",
      "Epoch:17 Batch:480 Log-loss0.042677734047174454\n",
      "Epoch:17 Batch:520 Log-loss0.03689684718847275\n",
      "Epoch:17 Batch:560 Log-loss0.03080877847969532\n",
      "Epoch average log loss:0.039086649447147334\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch17, val_loss:0.03733626075097543, best_val_loss:0.03733626075097543, best_auc:0.9890794771232595\n",
      "Epoch:18 Batch:0 Log-loss0.031091362237930298\n",
      "Epoch:18 Batch:40 Log-loss0.04164903983473778\n",
      "Epoch:18 Batch:80 Log-loss0.05218102037906647\n",
      "Epoch:18 Batch:120 Log-loss0.04123418405652046\n",
      "Epoch:18 Batch:160 Log-loss0.028827277943491936\n",
      "Epoch:18 Batch:200 Log-loss0.0318727046251297\n",
      "Epoch:18 Batch:240 Log-loss0.032309211790561676\n",
      "Epoch:18 Batch:280 Log-loss0.045698728412389755\n",
      "Epoch:18 Batch:320 Log-loss0.06347060948610306\n",
      "Epoch:18 Batch:360 Log-loss0.040234923362731934\n",
      "Epoch:18 Batch:400 Log-loss0.043210744857788086\n",
      "Epoch:18 Batch:440 Log-loss0.0306980162858963\n",
      "Epoch:18 Batch:480 Log-loss0.04811590909957886\n",
      "Epoch:18 Batch:520 Log-loss0.028183894231915474\n",
      "Epoch:18 Batch:560 Log-loss0.04587263613939285\n",
      "Epoch average log loss:0.0387977229331487\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch18, val_loss:0.03724741371682446, best_val_loss:0.03724741371682446, best_auc:0.9893603398944103\n",
      "Epoch:19 Batch:0 Log-loss0.040433354675769806\n",
      "Epoch:19 Batch:40 Log-loss0.045817479491233826\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:19 Batch:80 Log-loss0.05360683426260948\n",
      "Epoch:19 Batch:120 Log-loss0.05828571319580078\n",
      "Epoch:19 Batch:160 Log-loss0.04043808579444885\n",
      "Epoch:19 Batch:200 Log-loss0.024384962394833565\n",
      "Epoch:19 Batch:240 Log-loss0.04166371375322342\n",
      "Epoch:19 Batch:280 Log-loss0.05821231007575989\n",
      "Epoch:19 Batch:320 Log-loss0.03888482227921486\n",
      "Epoch:19 Batch:360 Log-loss0.04209063574671745\n",
      "Epoch:19 Batch:400 Log-loss0.03534097224473953\n",
      "Epoch:19 Batch:440 Log-loss0.02912217192351818\n",
      "Epoch:19 Batch:480 Log-loss0.03045186959207058\n",
      "Epoch:19 Batch:520 Log-loss0.05266507342457771\n",
      "Epoch:19 Batch:560 Log-loss0.04180420935153961\n",
      "Epoch average log loss:0.03868482537634139\n",
      "In Epoch19, val_loss:0.03725522742800543, best_val_loss:0.03724741371682446, best_auc:0.9893603398944103\n",
      "Epoch:20 Batch:0 Log-loss0.034415099769830704\n",
      "Epoch:20 Batch:40 Log-loss0.03907474875450134\n",
      "Epoch:20 Batch:80 Log-loss0.03349713981151581\n",
      "Epoch:20 Batch:120 Log-loss0.038461823016405106\n",
      "Epoch:20 Batch:160 Log-loss0.038610685616731644\n",
      "Epoch:20 Batch:200 Log-loss0.040330883115530014\n",
      "Epoch:20 Batch:240 Log-loss0.04207126423716545\n",
      "Epoch:20 Batch:280 Log-loss0.03351474553346634\n",
      "Epoch:20 Batch:320 Log-loss0.04276834428310394\n",
      "Epoch:20 Batch:360 Log-loss0.04138711094856262\n",
      "Epoch:20 Batch:400 Log-loss0.02659069560468197\n",
      "Epoch:20 Batch:440 Log-loss0.028834976255893707\n",
      "Epoch:20 Batch:480 Log-loss0.030315140262246132\n",
      "Epoch:20 Batch:520 Log-loss0.034444455057382584\n",
      "Epoch:20 Batch:560 Log-loss0.04219679534435272\n",
      "Epoch average log loss:0.03847605109414352\n",
      "In Epoch20, val_loss:0.03727364605449896, best_val_loss:0.03724741371682446, best_auc:0.9893603398944103\n",
      "Epoch:21 Batch:0 Log-loss0.050627946853637695\n",
      "Epoch:21 Batch:40 Log-loss0.04329685866832733\n",
      "Epoch:21 Batch:80 Log-loss0.027221783995628357\n",
      "Epoch:21 Batch:120 Log-loss0.031688518822193146\n",
      "Epoch:21 Batch:160 Log-loss0.03505614027380943\n",
      "Epoch:21 Batch:200 Log-loss0.03058973141014576\n",
      "Epoch:21 Batch:240 Log-loss0.0388977937400341\n",
      "Epoch:21 Batch:280 Log-loss0.045375410467386246\n",
      "Epoch:21 Batch:320 Log-loss0.03246944025158882\n",
      "Epoch:21 Batch:360 Log-loss0.037186432629823685\n",
      "Epoch:21 Batch:400 Log-loss0.02860615961253643\n",
      "Epoch:21 Batch:440 Log-loss0.044276703149080276\n",
      "Epoch:21 Batch:480 Log-loss0.03197610005736351\n",
      "Epoch:21 Batch:520 Log-loss0.045383941382169724\n",
      "Epoch:21 Batch:560 Log-loss0.034185316413640976\n",
      "Epoch average log loss:0.03812314389007432\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch21, val_loss:0.0371718983965491, best_val_loss:0.0371718983965491, best_auc:0.9893133946687094\n",
      "Epoch:22 Batch:0 Log-loss0.028315536677837372\n",
      "Epoch:22 Batch:40 Log-loss0.027140600606799126\n",
      "Epoch:22 Batch:80 Log-loss0.03417358174920082\n",
      "Epoch:22 Batch:120 Log-loss0.03949957713484764\n",
      "Epoch:22 Batch:160 Log-loss0.040738221257925034\n",
      "Epoch:22 Batch:200 Log-loss0.04176154360175133\n",
      "Epoch:22 Batch:240 Log-loss0.035430703312158585\n",
      "Epoch:22 Batch:280 Log-loss0.03903495892882347\n",
      "Epoch:22 Batch:320 Log-loss0.04774244502186775\n",
      "Epoch:22 Batch:360 Log-loss0.042632121592760086\n",
      "Epoch:22 Batch:400 Log-loss0.031806278973817825\n",
      "Epoch:22 Batch:440 Log-loss0.052811335772275925\n",
      "Epoch:22 Batch:480 Log-loss0.049546193331480026\n",
      "Epoch:22 Batch:520 Log-loss0.030499249696731567\n",
      "Epoch:22 Batch:560 Log-loss0.046400729566812515\n",
      "Epoch average log loss:0.03817291741059827\n",
      "In Epoch22, val_loss:0.03727735940354575, best_val_loss:0.0371718983965491, best_auc:0.9893133946687094\n",
      "Epoch:23 Batch:0 Log-loss0.02294871397316456\n",
      "Epoch:23 Batch:40 Log-loss0.02319283038377762\n",
      "Epoch:23 Batch:80 Log-loss0.0313132219016552\n",
      "Epoch:23 Batch:120 Log-loss0.029488638043403625\n",
      "Epoch:23 Batch:160 Log-loss0.035033468157052994\n",
      "Epoch:23 Batch:200 Log-loss0.039489444345235825\n",
      "Epoch:23 Batch:240 Log-loss0.03596692532300949\n",
      "Epoch:23 Batch:280 Log-loss0.04450378939509392\n",
      "Epoch:23 Batch:320 Log-loss0.03608432412147522\n",
      "Epoch:23 Batch:360 Log-loss0.04406682774424553\n",
      "Epoch:23 Batch:400 Log-loss0.041250213980674744\n",
      "Epoch:23 Batch:440 Log-loss0.04606792330741882\n",
      "Epoch:23 Batch:480 Log-loss0.041651662439107895\n",
      "Epoch:23 Batch:520 Log-loss0.03261914476752281\n",
      "Epoch:23 Batch:560 Log-loss0.036399904638528824\n",
      "Epoch average log loss:0.03809540901233309\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch23, val_loss:0.03710159828232489, best_val_loss:0.03710159828232489, best_auc:0.9893839104341803\n",
      "Epoch:24 Batch:0 Log-loss0.020188087597489357\n",
      "Epoch:24 Batch:40 Log-loss0.029546143487095833\n",
      "Epoch:24 Batch:80 Log-loss0.03559206798672676\n",
      "Epoch:24 Batch:120 Log-loss0.035080354660749435\n",
      "Epoch:24 Batch:160 Log-loss0.04346700385212898\n",
      "Epoch:24 Batch:200 Log-loss0.033350542187690735\n",
      "Epoch:24 Batch:240 Log-loss0.04141024872660637\n",
      "Epoch:24 Batch:280 Log-loss0.036230068653821945\n",
      "Epoch:24 Batch:320 Log-loss0.04182161018252373\n",
      "Epoch:24 Batch:360 Log-loss0.05584610626101494\n",
      "Epoch:24 Batch:400 Log-loss0.04062120243906975\n",
      "Epoch:24 Batch:440 Log-loss0.05156521499156952\n",
      "Epoch:24 Batch:480 Log-loss0.04834919795393944\n",
      "Epoch:24 Batch:520 Log-loss0.03379335626959801\n",
      "Epoch:24 Batch:560 Log-loss0.033295001834630966\n",
      "Epoch average log loss:0.03789077893431698\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch24, val_loss:0.03702593379868117, best_val_loss:0.03702593379868117, best_auc:0.9892036964695868\n",
      "Epoch:25 Batch:0 Log-loss0.026977354660630226\n",
      "Epoch:25 Batch:40 Log-loss0.0368429534137249\n",
      "Epoch:25 Batch:80 Log-loss0.033793289214372635\n",
      "Epoch:25 Batch:120 Log-loss0.049247365444898605\n",
      "Epoch:25 Batch:160 Log-loss0.03754157945513725\n",
      "Epoch:25 Batch:200 Log-loss0.026966562494635582\n",
      "Epoch:25 Batch:240 Log-loss0.03238988667726517\n",
      "Epoch:25 Batch:280 Log-loss0.04154418781399727\n",
      "Epoch:25 Batch:320 Log-loss0.03437783196568489\n",
      "Epoch:25 Batch:360 Log-loss0.03344954177737236\n",
      "Epoch:25 Batch:400 Log-loss0.04158373549580574\n",
      "Epoch:25 Batch:440 Log-loss0.04108862206339836\n",
      "Epoch:25 Batch:480 Log-loss0.02522909641265869\n",
      "Epoch:25 Batch:520 Log-loss0.035650383681058884\n",
      "Epoch:25 Batch:560 Log-loss0.0408538319170475\n",
      "Epoch average log loss:0.037777050524683935\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch25, val_loss:0.03696104645055152, best_val_loss:0.03696104645055152, best_auc:0.9891425178489727\n",
      "Epoch:26 Batch:0 Log-loss0.05380634590983391\n",
      "Epoch:26 Batch:40 Log-loss0.025991901755332947\n",
      "Epoch:26 Batch:80 Log-loss0.032185498625040054\n",
      "Epoch:26 Batch:120 Log-loss0.042781099677085876\n",
      "Epoch:26 Batch:160 Log-loss0.03918003663420677\n",
      "Epoch:26 Batch:200 Log-loss0.037711191922426224\n",
      "Epoch:26 Batch:240 Log-loss0.03837515786290169\n",
      "Epoch:26 Batch:280 Log-loss0.02656410075724125\n",
      "Epoch:26 Batch:320 Log-loss0.04705999419093132\n",
      "Epoch:26 Batch:360 Log-loss0.0540807731449604\n",
      "Epoch:26 Batch:400 Log-loss0.04832648113369942\n",
      "Epoch:26 Batch:440 Log-loss0.027250604704022408\n",
      "Epoch:26 Batch:480 Log-loss0.029266655445098877\n",
      "Epoch:26 Batch:520 Log-loss0.0542508028447628\n",
      "Epoch:26 Batch:560 Log-loss0.03042042814195156\n",
      "Epoch average log loss:0.037758879173946165\n",
      "In Epoch26, val_loss:0.037000756418785495, best_val_loss:0.03696104645055152, best_auc:0.9891425178489727\n",
      "Epoch:27 Batch:0 Log-loss0.03724614158272743\n",
      "Epoch:27 Batch:40 Log-loss0.032132554799318314\n",
      "Epoch:27 Batch:80 Log-loss0.04139873385429382\n",
      "Epoch:27 Batch:120 Log-loss0.03571419045329094\n",
      "Epoch:27 Batch:160 Log-loss0.04403592273592949\n",
      "Epoch:27 Batch:200 Log-loss0.04969759285449982\n",
      "Epoch:27 Batch:240 Log-loss0.0329265296459198\n",
      "Epoch:27 Batch:280 Log-loss0.04414353892207146\n",
      "Epoch:27 Batch:320 Log-loss0.04011565446853638\n",
      "Epoch:27 Batch:360 Log-loss0.036212194710969925\n",
      "Epoch:27 Batch:400 Log-loss0.03839101269841194\n",
      "Epoch:27 Batch:440 Log-loss0.041472528129816055\n",
      "Epoch:27 Batch:480 Log-loss0.03777704015374184\n",
      "Epoch:27 Batch:520 Log-loss0.037891123443841934\n",
      "Epoch:27 Batch:560 Log-loss0.05033440887928009\n",
      "Epoch average log loss:0.03771877424525363\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch27, val_loss:0.036953444690154345, best_val_loss:0.036953444690154345, best_auc:0.9895186046708294\n",
      "Epoch:28 Batch:0 Log-loss0.03435344621539116\n",
      "Epoch:28 Batch:40 Log-loss0.03544742986559868\n",
      "Epoch:28 Batch:80 Log-loss0.027348199859261513\n",
      "Epoch:28 Batch:120 Log-loss0.05521097779273987\n",
      "Epoch:28 Batch:160 Log-loss0.036024272441864014\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:28 Batch:200 Log-loss0.031570933759212494\n",
      "Epoch:28 Batch:240 Log-loss0.06176721677184105\n",
      "Epoch:28 Batch:280 Log-loss0.038737863302230835\n",
      "Epoch:28 Batch:320 Log-loss0.04648162052035332\n",
      "Epoch:28 Batch:360 Log-loss0.03659859299659729\n",
      "Epoch:28 Batch:400 Log-loss0.03391778841614723\n",
      "Epoch:28 Batch:440 Log-loss0.03358413279056549\n",
      "Epoch:28 Batch:480 Log-loss0.03738055005669594\n",
      "Epoch:28 Batch:520 Log-loss0.030275478959083557\n",
      "Epoch:28 Batch:560 Log-loss0.031242234632372856\n",
      "Epoch average log loss:0.037374141542906206\n",
      "In Epoch28, val_loss:0.037056066856172847, best_val_loss:0.036953444690154345, best_auc:0.9895186046708294\n",
      "Epoch:29 Batch:0 Log-loss0.0413806326687336\n",
      "Epoch:29 Batch:40 Log-loss0.0488111674785614\n",
      "Epoch:29 Batch:80 Log-loss0.03226311132311821\n",
      "Epoch:29 Batch:120 Log-loss0.060494571924209595\n",
      "Epoch:29 Batch:160 Log-loss0.030656667426228523\n",
      "Epoch:29 Batch:200 Log-loss0.03970308601856232\n",
      "Epoch:29 Batch:240 Log-loss0.03575023636221886\n",
      "Epoch:29 Batch:280 Log-loss0.03034629113972187\n",
      "Epoch:29 Batch:320 Log-loss0.024671025574207306\n",
      "Epoch:29 Batch:360 Log-loss0.024531090632081032\n",
      "Epoch:29 Batch:400 Log-loss0.04132862016558647\n",
      "Epoch:29 Batch:440 Log-loss0.03334791958332062\n",
      "Epoch:29 Batch:480 Log-loss0.02805553376674652\n",
      "Epoch:29 Batch:520 Log-loss0.02885289490222931\n",
      "Epoch:29 Batch:560 Log-loss0.0316438302397728\n",
      "Epoch average log loss:0.03750020012126437\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch29, val_loss:0.036946430352045465, best_val_loss:0.036946430352045465, best_auc:0.9895059606832097\n",
      "Epoch:30 Batch:0 Log-loss0.05984675511717796\n",
      "Epoch:30 Batch:40 Log-loss0.04036632180213928\n",
      "Epoch:30 Batch:80 Log-loss0.034646015614271164\n",
      "Epoch:30 Batch:120 Log-loss0.03820523992180824\n",
      "Epoch:30 Batch:160 Log-loss0.03649530187249184\n",
      "Epoch:30 Batch:200 Log-loss0.04038769379258156\n",
      "Epoch:30 Batch:240 Log-loss0.029921114444732666\n",
      "Epoch:30 Batch:280 Log-loss0.034326136112213135\n",
      "Epoch:30 Batch:320 Log-loss0.028550120070576668\n",
      "Epoch:30 Batch:360 Log-loss0.04097624495625496\n",
      "Epoch:30 Batch:400 Log-loss0.017992248758673668\n",
      "Epoch:30 Batch:440 Log-loss0.03207756578922272\n",
      "Epoch:30 Batch:480 Log-loss0.04349773749709129\n",
      "Epoch:30 Batch:520 Log-loss0.037456002086400986\n",
      "Epoch:30 Batch:560 Log-loss0.033697545528411865\n",
      "Epoch average log loss:0.037322685643032726\n",
      "In Epoch30, val_loss:0.037249935469444, best_val_loss:0.036946430352045465, best_auc:0.9895059606832097\n",
      "Epoch:31 Batch:0 Log-loss0.03137374296784401\n",
      "Epoch:31 Batch:40 Log-loss0.03188662603497505\n",
      "Epoch:31 Batch:80 Log-loss0.03696281835436821\n",
      "Epoch:31 Batch:120 Log-loss0.03559885919094086\n",
      "Epoch:31 Batch:160 Log-loss0.050297532230615616\n",
      "Epoch:31 Batch:200 Log-loss0.043092016130685806\n",
      "Epoch:31 Batch:240 Log-loss0.04278978705406189\n",
      "Epoch:31 Batch:280 Log-loss0.03934822231531143\n",
      "Epoch:31 Batch:320 Log-loss0.03354959562420845\n",
      "Epoch:31 Batch:360 Log-loss0.04686437547206879\n",
      "Epoch:31 Batch:400 Log-loss0.024503132328391075\n",
      "Epoch:31 Batch:440 Log-loss0.04252990707755089\n",
      "Epoch:31 Batch:480 Log-loss0.06017795577645302\n",
      "Epoch:31 Batch:520 Log-loss0.03907265514135361\n",
      "Epoch:31 Batch:560 Log-loss0.03912781924009323\n",
      "Epoch average log loss:0.037238247958677154\n",
      "In Epoch31, val_loss:0.03694850825224764, best_val_loss:0.036946430352045465, best_auc:0.9895059606832097\n",
      "Epoch:32 Batch:0 Log-loss0.024550022557377815\n",
      "Epoch:32 Batch:40 Log-loss0.03845107555389404\n",
      "Epoch:32 Batch:80 Log-loss0.02943115495145321\n",
      "Epoch:32 Batch:120 Log-loss0.030277058482170105\n",
      "Epoch:32 Batch:160 Log-loss0.03996164724230766\n",
      "Epoch:32 Batch:200 Log-loss0.039609313011169434\n",
      "Epoch:32 Batch:240 Log-loss0.04094173386693001\n",
      "Epoch:32 Batch:280 Log-loss0.024875901639461517\n",
      "Epoch:32 Batch:320 Log-loss0.04521198570728302\n",
      "Epoch:32 Batch:360 Log-loss0.04523533210158348\n",
      "Epoch:32 Batch:400 Log-loss0.03908118978142738\n",
      "Epoch:32 Batch:440 Log-loss0.036674533039331436\n",
      "Epoch:32 Batch:480 Log-loss0.04959762468934059\n",
      "Epoch:32 Batch:520 Log-loss0.03626297786831856\n",
      "Epoch:32 Batch:560 Log-loss0.0260720644146204\n",
      "Epoch average log loss:0.037177649117074905\n",
      "In Epoch32, val_loss:0.036954077171879686, best_val_loss:0.036946430352045465, best_auc:0.9895059606832097\n",
      "Epoch:33 Batch:0 Log-loss0.045042458921670914\n",
      "Epoch:33 Batch:40 Log-loss0.037948690354824066\n",
      "Epoch:33 Batch:80 Log-loss0.031077692285180092\n",
      "Epoch:33 Batch:120 Log-loss0.02989860810339451\n",
      "Epoch:33 Batch:160 Log-loss0.024002818390727043\n",
      "Epoch:33 Batch:200 Log-loss0.024053893983364105\n",
      "Epoch:33 Batch:240 Log-loss0.035787809640169144\n",
      "Epoch:33 Batch:280 Log-loss0.03394898772239685\n",
      "Epoch:33 Batch:320 Log-loss0.026800906285643578\n",
      "Epoch:33 Batch:360 Log-loss0.03313779458403587\n",
      "Epoch:33 Batch:400 Log-loss0.02219502627849579\n",
      "Epoch:33 Batch:440 Log-loss0.050508931279182434\n",
      "Epoch:33 Batch:480 Log-loss0.037076812237501144\n",
      "Epoch:33 Batch:520 Log-loss0.024640297517180443\n",
      "Epoch:33 Batch:560 Log-loss0.03511400893330574\n",
      "Epoch average log loss:0.03723317169185195\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch33, val_loss:0.03685045472166809, best_val_loss:0.03685045472166809, best_auc:0.9894506123515687\n",
      "Epoch:34 Batch:0 Log-loss0.03990698978304863\n",
      "Epoch:34 Batch:40 Log-loss0.02200227789580822\n",
      "Epoch:34 Batch:80 Log-loss0.03762834146618843\n",
      "Epoch:34 Batch:120 Log-loss0.026852922514081\n",
      "Epoch:34 Batch:160 Log-loss0.03547540307044983\n",
      "Epoch:34 Batch:200 Log-loss0.03262751176953316\n",
      "Epoch:34 Batch:240 Log-loss0.037400633096694946\n",
      "Epoch:34 Batch:280 Log-loss0.025090934708714485\n",
      "Epoch:34 Batch:320 Log-loss0.03161204233765602\n",
      "Epoch:34 Batch:360 Log-loss0.043744370341300964\n",
      "Epoch:34 Batch:400 Log-loss0.0383477509021759\n",
      "Epoch:34 Batch:440 Log-loss0.02681908756494522\n",
      "Epoch:34 Batch:480 Log-loss0.031787145882844925\n",
      "Epoch:34 Batch:520 Log-loss0.03384857252240181\n",
      "Epoch:34 Batch:560 Log-loss0.04718747362494469\n",
      "Epoch average log loss:0.03707045710179955\n",
      "In Epoch34, val_loss:0.036882586302734215, best_val_loss:0.03685045472166809, best_auc:0.9894506123515687\n",
      "Epoch:35 Batch:0 Log-loss0.05462096258997917\n",
      "Epoch:35 Batch:40 Log-loss0.0463414341211319\n",
      "Epoch:35 Batch:80 Log-loss0.03695293515920639\n",
      "Epoch:35 Batch:120 Log-loss0.03324008360505104\n",
      "Epoch:35 Batch:160 Log-loss0.0368196964263916\n",
      "Epoch:35 Batch:200 Log-loss0.03477618470788002\n",
      "Epoch:35 Batch:240 Log-loss0.03512202575802803\n",
      "Epoch:35 Batch:280 Log-loss0.03568184748291969\n",
      "Epoch:35 Batch:320 Log-loss0.039729125797748566\n",
      "Epoch:35 Batch:360 Log-loss0.030913082882761955\n",
      "Epoch:35 Batch:400 Log-loss0.035580527037382126\n",
      "Epoch:35 Batch:440 Log-loss0.03301239758729935\n",
      "Epoch:35 Batch:480 Log-loss0.06218094751238823\n",
      "Epoch:35 Batch:520 Log-loss0.04542786255478859\n",
      "Epoch:35 Batch:560 Log-loss0.03753327578306198\n",
      "Epoch average log loss:0.03713505689187774\n",
      "In Epoch35, val_loss:0.03702416903427235, best_val_loss:0.03685045472166809, best_auc:0.9894506123515687\n",
      "Epoch:36 Batch:0 Log-loss0.04245641455054283\n",
      "Epoch:36 Batch:40 Log-loss0.04110080003738403\n",
      "Epoch:36 Batch:80 Log-loss0.028409035876393318\n",
      "Epoch:36 Batch:120 Log-loss0.0379188247025013\n",
      "Epoch:36 Batch:160 Log-loss0.04331878200173378\n",
      "Epoch:36 Batch:200 Log-loss0.04239397868514061\n",
      "Epoch:36 Batch:240 Log-loss0.03958999738097191\n",
      "Epoch:36 Batch:280 Log-loss0.03897681459784508\n",
      "Epoch:36 Batch:320 Log-loss0.03237045183777809\n",
      "Epoch:36 Batch:360 Log-loss0.03319082036614418\n",
      "Epoch:36 Batch:400 Log-loss0.029043259099125862\n",
      "Epoch:36 Batch:440 Log-loss0.034833163022994995\n",
      "Epoch:36 Batch:480 Log-loss0.03226916864514351\n",
      "Epoch:36 Batch:520 Log-loss0.03404688090085983\n",
      "Epoch:36 Batch:560 Log-loss0.03502083197236061\n",
      "Epoch average log loss:0.03716790107178635\n",
      "In Epoch36, val_loss:0.03700696661175414, best_val_loss:0.03685045472166809, best_auc:0.9894506123515687\n",
      "Epoch:37 Batch:0 Log-loss0.04181408882141113\n",
      "Epoch:37 Batch:40 Log-loss0.03812292963266373\n",
      "Epoch:37 Batch:80 Log-loss0.027309710159897804\n",
      "Epoch:37 Batch:120 Log-loss0.0445278026163578\n",
      "Epoch:37 Batch:160 Log-loss0.024697592481970787\n",
      "Epoch:37 Batch:200 Log-loss0.042861517518758774\n",
      "Epoch:37 Batch:240 Log-loss0.028210438787937164\n",
      "Epoch:37 Batch:280 Log-loss0.03396926447749138\n",
      "Epoch:37 Batch:320 Log-loss0.03776480257511139\n",
      "Epoch:37 Batch:360 Log-loss0.03636036440730095\n",
      "Epoch:37 Batch:400 Log-loss0.03484584763646126\n",
      "Epoch:37 Batch:440 Log-loss0.04358665272593498\n",
      "Epoch:37 Batch:480 Log-loss0.039557263255119324\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:37 Batch:520 Log-loss0.0317780040204525\n",
      "Epoch:37 Batch:560 Log-loss0.036394987255334854\n",
      "Epoch average log loss:0.03683391263974564\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch37, val_loss:0.036779424331287926, best_val_loss:0.036779424331287926, best_auc:0.9895815556650747\n",
      "Epoch:38 Batch:0 Log-loss0.040687840431928635\n",
      "Epoch:38 Batch:40 Log-loss0.04156241938471794\n",
      "Epoch:38 Batch:80 Log-loss0.040576811879873276\n",
      "Epoch:38 Batch:120 Log-loss0.031041691079735756\n",
      "Epoch:38 Batch:160 Log-loss0.04214386269450188\n",
      "Epoch:38 Batch:200 Log-loss0.045967068523168564\n",
      "Epoch:38 Batch:240 Log-loss0.031998272985219955\n",
      "Epoch:38 Batch:280 Log-loss0.03450891748070717\n",
      "Epoch:38 Batch:320 Log-loss0.03579799458384514\n",
      "Epoch:38 Batch:360 Log-loss0.048357706516981125\n",
      "Epoch:38 Batch:400 Log-loss0.03253639116883278\n",
      "Epoch:38 Batch:440 Log-loss0.035472165793180466\n",
      "Epoch:38 Batch:480 Log-loss0.022564785555005074\n",
      "Epoch:38 Batch:520 Log-loss0.026673145592212677\n",
      "Epoch:38 Batch:560 Log-loss0.0351593941450119\n",
      "Epoch average log loss:0.03698263983907444\n",
      "In Epoch38, val_loss:0.03690475596633406, best_val_loss:0.036779424331287926, best_auc:0.9895815556650747\n",
      "Epoch:39 Batch:0 Log-loss0.03807559981942177\n",
      "Epoch:39 Batch:40 Log-loss0.05057978630065918\n",
      "Epoch:39 Batch:80 Log-loss0.045773785561323166\n",
      "Epoch:39 Batch:120 Log-loss0.04727158322930336\n",
      "Epoch:39 Batch:160 Log-loss0.03808673098683357\n",
      "Epoch:39 Batch:200 Log-loss0.025789594277739525\n",
      "Epoch:39 Batch:240 Log-loss0.037169378250837326\n",
      "Epoch:39 Batch:280 Log-loss0.035156041383743286\n",
      "Epoch:39 Batch:320 Log-loss0.05295984074473381\n",
      "Epoch:39 Batch:360 Log-loss0.038340236991643906\n",
      "Epoch:39 Batch:400 Log-loss0.042638540267944336\n",
      "Epoch:39 Batch:440 Log-loss0.027680404484272003\n",
      "Epoch:39 Batch:480 Log-loss0.04154578596353531\n",
      "Epoch:39 Batch:520 Log-loss0.03126930817961693\n",
      "Epoch:39 Batch:560 Log-loss0.04197578877210617\n",
      "Epoch average log loss:0.036955274644839976\n",
      "In Epoch39, val_loss:0.036881704161292855, best_val_loss:0.036779424331287926, best_auc:0.9895815556650747\n",
      "Epoch:40 Batch:0 Log-loss0.03653962165117264\n",
      "Epoch:40 Batch:40 Log-loss0.035733941942453384\n",
      "Epoch:40 Batch:80 Log-loss0.04277871176600456\n",
      "Epoch:40 Batch:120 Log-loss0.046780142933130264\n",
      "Epoch:40 Batch:160 Log-loss0.034213826060295105\n",
      "Epoch:40 Batch:200 Log-loss0.029059937223792076\n",
      "Epoch:40 Batch:240 Log-loss0.031225822865962982\n",
      "Epoch:40 Batch:280 Log-loss0.024687020108103752\n",
      "Epoch:40 Batch:320 Log-loss0.02799355424940586\n",
      "Epoch:40 Batch:360 Log-loss0.03158964961767197\n",
      "Epoch:40 Batch:400 Log-loss0.0544680655002594\n",
      "Epoch:40 Batch:440 Log-loss0.04156803712248802\n",
      "Epoch:40 Batch:480 Log-loss0.036043547093868256\n",
      "Epoch:40 Batch:520 Log-loss0.05093027278780937\n",
      "Epoch:40 Batch:560 Log-loss0.034555140882730484\n",
      "Epoch average log loss:0.036789428060209114\n",
      "In Epoch40, val_loss:0.03682717914139827, best_val_loss:0.036779424331287926, best_auc:0.9895815556650747\n",
      "Epoch:41 Batch:0 Log-loss0.02822846733033657\n",
      "Epoch:41 Batch:40 Log-loss0.032350655645132065\n",
      "Epoch:41 Batch:80 Log-loss0.02900460921227932\n",
      "Epoch:41 Batch:120 Log-loss0.01946212537586689\n",
      "Epoch:41 Batch:160 Log-loss0.025386126711964607\n",
      "Epoch:41 Batch:200 Log-loss0.029901819303631783\n",
      "Epoch:41 Batch:240 Log-loss0.036139145493507385\n",
      "Epoch:41 Batch:280 Log-loss0.04835100844502449\n",
      "Epoch:41 Batch:320 Log-loss0.029775559902191162\n",
      "Epoch:41 Batch:360 Log-loss0.04188448563218117\n",
      "Epoch:41 Batch:400 Log-loss0.03215942904353142\n",
      "Epoch:41 Batch:440 Log-loss0.04994639754295349\n",
      "Epoch:41 Batch:480 Log-loss0.027630405500531197\n",
      "Epoch:41 Batch:520 Log-loss0.031207725405693054\n",
      "Epoch:41 Batch:560 Log-loss0.06173643842339516\n",
      "Epoch average log loss:0.03692918002073254\n",
      "In Epoch41, val_loss:0.03685726816704996, best_val_loss:0.036779424331287926, best_auc:0.9895815556650747\n",
      "Epoch:42 Batch:0 Log-loss0.029008449986577034\n",
      "Epoch:42 Batch:40 Log-loss0.03451765701174736\n",
      "Epoch:42 Batch:80 Log-loss0.03319038823246956\n",
      "Epoch:42 Batch:120 Log-loss0.0538812130689621\n",
      "Epoch:42 Batch:160 Log-loss0.06701799482107162\n",
      "Epoch:42 Batch:200 Log-loss0.03753529489040375\n",
      "Epoch:42 Batch:240 Log-loss0.026587555184960365\n",
      "Epoch:42 Batch:280 Log-loss0.034241486340761185\n",
      "Epoch:42 Batch:320 Log-loss0.037418536841869354\n",
      "Epoch:42 Batch:360 Log-loss0.03486824408173561\n",
      "Epoch:42 Batch:400 Log-loss0.031203120946884155\n",
      "Epoch:42 Batch:440 Log-loss0.03365864232182503\n",
      "Epoch:42 Batch:480 Log-loss0.038846664130687714\n",
      "Epoch:42 Batch:520 Log-loss0.03252311795949936\n",
      "Epoch:42 Batch:560 Log-loss0.04471322149038315\n",
      "Epoch average log loss:0.03698213135357946\n",
      "In Epoch42, val_loss:0.03689074588253406, best_val_loss:0.036779424331287926, best_auc:0.9895815556650747\n",
      "Epoch:43 Batch:0 Log-loss0.04382934048771858\n",
      "Epoch:43 Batch:40 Log-loss0.044601913541555405\n",
      "Epoch:43 Batch:80 Log-loss0.038064148277044296\n",
      "Epoch:43 Batch:120 Log-loss0.030738862231373787\n",
      "Epoch:43 Batch:160 Log-loss0.03664319962263107\n",
      "Epoch:43 Batch:200 Log-loss0.03386829420924187\n",
      "Epoch:43 Batch:240 Log-loss0.05448826774954796\n",
      "Epoch:43 Batch:280 Log-loss0.03649897500872612\n",
      "Epoch:43 Batch:320 Log-loss0.0428546667098999\n",
      "Epoch:43 Batch:360 Log-loss0.035165127366781235\n",
      "Epoch:43 Batch:400 Log-loss0.036113426089286804\n",
      "Epoch:43 Batch:440 Log-loss0.032082267105579376\n",
      "Epoch:43 Batch:480 Log-loss0.04871675372123718\n",
      "Epoch:43 Batch:520 Log-loss0.03376783803105354\n",
      "Epoch:43 Batch:560 Log-loss0.03364066034555435\n",
      "Epoch average log loss:0.03693489140631365\n",
      "In Epoch43, val_loss:0.03684379164203771, best_val_loss:0.036779424331287926, best_auc:0.9895815556650747\n",
      "Epoch:44 Batch:0 Log-loss0.0368502251803875\n",
      "Epoch:44 Batch:40 Log-loss0.04976074397563934\n",
      "Epoch:44 Batch:80 Log-loss0.05330665409564972\n",
      "Epoch:44 Batch:120 Log-loss0.03899279981851578\n",
      "Epoch:44 Batch:160 Log-loss0.052394405007362366\n",
      "Epoch:44 Batch:200 Log-loss0.05981970950961113\n",
      "Epoch:44 Batch:240 Log-loss0.03944256156682968\n",
      "Epoch:44 Batch:280 Log-loss0.04858657345175743\n",
      "Epoch:44 Batch:320 Log-loss0.03754880651831627\n",
      "Epoch:44 Batch:360 Log-loss0.033855073153972626\n",
      "Epoch:44 Batch:400 Log-loss0.0374278761446476\n",
      "Epoch:44 Batch:440 Log-loss0.03404397889971733\n",
      "Epoch:44 Batch:480 Log-loss0.03144681826233864\n",
      "Epoch:44 Batch:520 Log-loss0.038486067205667496\n",
      "Epoch:44 Batch:560 Log-loss0.04075628146529198\n",
      "Epoch average log loss:0.03681212373942669\n",
      "In Epoch44, val_loss:0.03684656920218106, best_val_loss:0.036779424331287926, best_auc:0.9895815556650747\n",
      "Epoch:45 Batch:0 Log-loss0.037298642098903656\n",
      "Epoch:45 Batch:40 Log-loss0.028539815917611122\n",
      "Epoch:45 Batch:80 Log-loss0.03445662930607796\n",
      "Epoch:45 Batch:120 Log-loss0.036051880568265915\n",
      "Epoch:45 Batch:160 Log-loss0.04764030873775482\n",
      "Epoch:45 Batch:200 Log-loss0.0414140410721302\n",
      "Epoch:45 Batch:240 Log-loss0.03733081743121147\n",
      "Epoch:45 Batch:280 Log-loss0.032368578016757965\n",
      "Epoch:45 Batch:320 Log-loss0.055764731019735336\n",
      "Epoch:45 Batch:360 Log-loss0.024579323828220367\n",
      "Epoch:45 Batch:400 Log-loss0.035887960344552994\n",
      "Epoch:45 Batch:440 Log-loss0.020738080143928528\n",
      "Epoch:45 Batch:480 Log-loss0.03682814911007881\n",
      "Epoch:45 Batch:520 Log-loss0.04296993091702461\n",
      "Epoch:45 Batch:560 Log-loss0.04260142147541046\n",
      "Epoch average log loss:0.0365080184935193\n",
      "In Epoch45, val_loss:0.036841663170472806, best_val_loss:0.036779424331287926, best_auc:0.9895815556650747\n",
      "Epoch:46 Batch:0 Log-loss0.0407344214618206\n",
      "Epoch:46 Batch:40 Log-loss0.022203238680958748\n",
      "Epoch:46 Batch:80 Log-loss0.03187166899442673\n",
      "Epoch:46 Batch:120 Log-loss0.031087161973118782\n",
      "Epoch:46 Batch:160 Log-loss0.038957629352808\n",
      "Epoch:46 Batch:200 Log-loss0.02837744913995266\n",
      "Epoch:46 Batch:240 Log-loss0.06378493458032608\n",
      "Epoch:46 Batch:280 Log-loss0.038088176399469376\n",
      "Epoch:46 Batch:320 Log-loss0.03619389981031418\n",
      "Epoch:46 Batch:360 Log-loss0.02844088524580002\n",
      "Epoch:46 Batch:400 Log-loss0.02758813463151455\n",
      "Epoch:46 Batch:440 Log-loss0.02589370310306549\n",
      "Epoch:46 Batch:480 Log-loss0.030248790979385376\n",
      "Epoch:46 Batch:520 Log-loss0.03618088364601135\n",
      "Epoch:46 Batch:560 Log-loss0.0379677377641201\n",
      "Epoch average log loss:0.03664473699006651\n",
      "In Epoch46, val_loss:0.036842450227374604, best_val_loss:0.036779424331287926, best_auc:0.9895815556650747\n",
      "Epoch:47 Batch:0 Log-loss0.05392705276608467\n",
      "Epoch:47 Batch:40 Log-loss0.04010666906833649\n",
      "Epoch:47 Batch:80 Log-loss0.03924063220620155\n",
      "Epoch:47 Batch:120 Log-loss0.031908068805933\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:47 Batch:160 Log-loss0.04869798943400383\n",
      "Epoch:47 Batch:200 Log-loss0.03346498683094978\n",
      "Epoch:47 Batch:240 Log-loss0.03800887614488602\n",
      "Epoch:47 Batch:280 Log-loss0.043155938386917114\n",
      "Epoch:47 Batch:320 Log-loss0.03454754129052162\n",
      "Epoch:47 Batch:360 Log-loss0.0354398675262928\n",
      "Epoch:47 Batch:400 Log-loss0.035380031913518906\n",
      "Epoch:47 Batch:440 Log-loss0.044184986501932144\n",
      "Epoch:47 Batch:480 Log-loss0.03441204875707626\n",
      "Epoch:47 Batch:520 Log-loss0.04677858576178551\n",
      "Epoch:47 Batch:560 Log-loss0.03860176354646683\n",
      "Epoch average log loss:0.03657726087341351\n",
      "A pre-trained model at temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt has been loaded.\n",
      "\n",
      "Model has been saved as checkpoints/FASTTXT_BGRU_64_640.pt.\n",
      "\n",
      "Choose the torch base model.\n",
      "Training on fold 1\n",
      "Epoch:1 Batch:0 Log-loss0.6727760434150696\n",
      "Epoch:1 Batch:40 Log-loss0.1336449235677719\n",
      "Epoch:1 Batch:80 Log-loss0.10393470525741577\n",
      "Epoch:1 Batch:120 Log-loss0.057671431452035904\n",
      "Epoch:1 Batch:160 Log-loss0.06155194714665413\n",
      "Epoch:1 Batch:200 Log-loss0.05471082404255867\n",
      "Epoch:1 Batch:240 Log-loss0.057089079171419144\n",
      "Epoch:1 Batch:280 Log-loss0.06460430473089218\n",
      "Epoch:1 Batch:320 Log-loss0.06532611697912216\n",
      "Epoch:1 Batch:360 Log-loss0.04995075985789299\n",
      "Epoch:1 Batch:400 Log-loss0.04205774888396263\n",
      "Epoch:1 Batch:440 Log-loss0.05861636996269226\n",
      "Epoch:1 Batch:480 Log-loss0.06474127620458603\n",
      "Epoch:1 Batch:520 Log-loss0.028494814410805702\n",
      "Epoch:1 Batch:560 Log-loss0.04608842730522156\n",
      "Epoch average log loss:0.08103964133520744\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch1, val_loss:0.04809863237672146, best_val_loss:0.04809863237672146, best_auc:0.9746432936476846\n",
      "Epoch:2 Batch:0 Log-loss0.059111401438713074\n",
      "Epoch:2 Batch:40 Log-loss0.0558774508535862\n",
      "Epoch:2 Batch:80 Log-loss0.054683148860931396\n",
      "Epoch:2 Batch:120 Log-loss0.06346443295478821\n",
      "Epoch:2 Batch:160 Log-loss0.06967416405677795\n",
      "Epoch:2 Batch:200 Log-loss0.037251055240631104\n",
      "Epoch:2 Batch:240 Log-loss0.02678484283387661\n",
      "Epoch:2 Batch:280 Log-loss0.04401285946369171\n",
      "Epoch:2 Batch:320 Log-loss0.0503217913210392\n",
      "Epoch:2 Batch:360 Log-loss0.04743117466568947\n",
      "Epoch:2 Batch:400 Log-loss0.03669824078679085\n",
      "Epoch:2 Batch:440 Log-loss0.05245361104607582\n",
      "Epoch:2 Batch:480 Log-loss0.05507412180304527\n",
      "Epoch:2 Batch:520 Log-loss0.04662197828292847\n",
      "Epoch:2 Batch:560 Log-loss0.05268581584095955\n",
      "Epoch average log loss:0.049748783694979336\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch2, val_loss:0.04534472269602024, best_val_loss:0.04534472269602024, best_auc:0.979075226274511\n",
      "Epoch:3 Batch:0 Log-loss0.04036319628357887\n",
      "Epoch:3 Batch:40 Log-loss0.04105386883020401\n",
      "Epoch:3 Batch:80 Log-loss0.04229682683944702\n",
      "Epoch:3 Batch:120 Log-loss0.04708269238471985\n",
      "Epoch:3 Batch:160 Log-loss0.0401812382042408\n",
      "Epoch:3 Batch:200 Log-loss0.03317039832472801\n",
      "Epoch:3 Batch:240 Log-loss0.05819818377494812\n",
      "Epoch:3 Batch:280 Log-loss0.045551881194114685\n",
      "Epoch:3 Batch:320 Log-loss0.034480880945920944\n",
      "Epoch:3 Batch:360 Log-loss0.042917266488075256\n",
      "Epoch:3 Batch:400 Log-loss0.04687904939055443\n",
      "Epoch:3 Batch:440 Log-loss0.036569371819496155\n",
      "Epoch:3 Batch:480 Log-loss0.039296746253967285\n",
      "Epoch:3 Batch:520 Log-loss0.052068423479795456\n",
      "Epoch:3 Batch:560 Log-loss0.05208706110715866\n",
      "Epoch average log loss:0.04730017806869\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch3, val_loss:0.04439769968135957, best_val_loss:0.04439769968135957, best_auc:0.98206761775662\n",
      "Epoch:4 Batch:0 Log-loss0.04856358841061592\n",
      "Epoch:4 Batch:40 Log-loss0.03585114702582359\n",
      "Epoch:4 Batch:80 Log-loss0.04838089644908905\n",
      "Epoch:4 Batch:120 Log-loss0.03710492327809334\n",
      "Epoch:4 Batch:160 Log-loss0.05734674260020256\n",
      "Epoch:4 Batch:200 Log-loss0.04462427273392677\n",
      "Epoch:4 Batch:240 Log-loss0.03820354864001274\n",
      "Epoch:4 Batch:280 Log-loss0.050293225795030594\n",
      "Epoch:4 Batch:320 Log-loss0.052235040813684464\n",
      "Epoch:4 Batch:360 Log-loss0.064984031021595\n",
      "Epoch:4 Batch:400 Log-loss0.029629310593008995\n",
      "Epoch:4 Batch:440 Log-loss0.039353467524051666\n",
      "Epoch:4 Batch:480 Log-loss0.0680144801735878\n",
      "Epoch:4 Batch:520 Log-loss0.057949602603912354\n",
      "Epoch:4 Batch:560 Log-loss0.031205227598547935\n",
      "Epoch average log loss:0.04585059221634375\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch4, val_loss:0.04299766561462292, best_val_loss:0.04299766561462292, best_auc:0.9841325864710758\n",
      "Epoch:5 Batch:0 Log-loss0.056753844022750854\n",
      "Epoch:5 Batch:40 Log-loss0.032236773520708084\n",
      "Epoch:5 Batch:80 Log-loss0.054152365773916245\n",
      "Epoch:5 Batch:120 Log-loss0.04483119025826454\n",
      "Epoch:5 Batch:160 Log-loss0.04339274764060974\n",
      "Epoch:5 Batch:200 Log-loss0.03879684954881668\n",
      "Epoch:5 Batch:240 Log-loss0.04107380658388138\n",
      "Epoch:5 Batch:280 Log-loss0.04653553664684296\n",
      "Epoch:5 Batch:320 Log-loss0.05834229290485382\n",
      "Epoch:5 Batch:360 Log-loss0.060627054423093796\n",
      "Epoch:5 Batch:400 Log-loss0.052543770521879196\n",
      "Epoch:5 Batch:440 Log-loss0.053448159247636795\n",
      "Epoch:5 Batch:480 Log-loss0.047275345772504807\n",
      "Epoch:5 Batch:520 Log-loss0.046732138842344284\n",
      "Epoch:5 Batch:560 Log-loss0.0499957874417305\n",
      "Epoch average log loss:0.04426202046951013\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch5, val_loss:0.04278488127899116, best_val_loss:0.04278488127899116, best_auc:0.9861511124736082\n",
      "Epoch:6 Batch:0 Log-loss0.032748375087976456\n",
      "Epoch:6 Batch:40 Log-loss0.05271555855870247\n",
      "Epoch:6 Batch:80 Log-loss0.052354250103235245\n",
      "Epoch:6 Batch:120 Log-loss0.05151710286736488\n",
      "Epoch:6 Batch:160 Log-loss0.03506479039788246\n",
      "Epoch:6 Batch:200 Log-loss0.044160258024930954\n",
      "Epoch:6 Batch:240 Log-loss0.06004012003540993\n",
      "Epoch:6 Batch:280 Log-loss0.0337318517267704\n",
      "Epoch:6 Batch:320 Log-loss0.027385303750634193\n",
      "Epoch:6 Batch:360 Log-loss0.0373309850692749\n",
      "Epoch:6 Batch:400 Log-loss0.04664253816008568\n",
      "Epoch:6 Batch:440 Log-loss0.05172867700457573\n",
      "Epoch:6 Batch:480 Log-loss0.05077628418803215\n",
      "Epoch:6 Batch:520 Log-loss0.030392156913876534\n",
      "Epoch:6 Batch:560 Log-loss0.040296949446201324\n",
      "Epoch average log loss:0.043428950597132955\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch6, val_loss:0.04187963891261338, best_val_loss:0.04187963891261338, best_auc:0.9875733376591457\n",
      "Epoch:7 Batch:0 Log-loss0.04411400482058525\n",
      "Epoch:7 Batch:40 Log-loss0.028560696169734\n",
      "Epoch:7 Batch:80 Log-loss0.02843683399260044\n",
      "Epoch:7 Batch:120 Log-loss0.03893585130572319\n",
      "Epoch:7 Batch:160 Log-loss0.04648426175117493\n",
      "Epoch:7 Batch:200 Log-loss0.04563752934336662\n",
      "Epoch:7 Batch:240 Log-loss0.04917936027050018\n",
      "Epoch:7 Batch:280 Log-loss0.0383734367787838\n",
      "Epoch:7 Batch:320 Log-loss0.03244008496403694\n",
      "Epoch:7 Batch:360 Log-loss0.04008994996547699\n",
      "Epoch:7 Batch:400 Log-loss0.058085694909095764\n",
      "Epoch:7 Batch:440 Log-loss0.04336382821202278\n",
      "Epoch:7 Batch:480 Log-loss0.0375632680952549\n",
      "Epoch:7 Batch:520 Log-loss0.04676268622279167\n",
      "Epoch:7 Batch:560 Log-loss0.05872436240315437\n",
      "Epoch average log loss:0.0425560999661684\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch7, val_loss:0.040984598798752596, best_val_loss:0.040984598798752596, best_auc:0.9880308111995241\n",
      "Epoch:8 Batch:0 Log-loss0.04058793559670448\n",
      "Epoch:8 Batch:40 Log-loss0.04498807713389397\n",
      "Epoch:8 Batch:80 Log-loss0.031758055090904236\n",
      "Epoch:8 Batch:120 Log-loss0.04818152263760567\n",
      "Epoch:8 Batch:160 Log-loss0.04897372052073479\n",
      "Epoch:8 Batch:200 Log-loss0.04950004443526268\n",
      "Epoch:8 Batch:240 Log-loss0.051625121384859085\n",
      "Epoch:8 Batch:280 Log-loss0.03736817464232445\n",
      "Epoch:8 Batch:320 Log-loss0.037978921085596085\n",
      "Epoch:8 Batch:360 Log-loss0.042276784777641296\n",
      "Epoch:8 Batch:400 Log-loss0.051271289587020874\n",
      "Epoch:8 Batch:440 Log-loss0.04174366220831871\n",
      "Epoch:8 Batch:480 Log-loss0.04907180741429329\n",
      "Epoch:8 Batch:520 Log-loss0.04264471307396889\n",
      "Epoch:8 Batch:560 Log-loss0.04773768410086632\n",
      "Epoch average log loss:0.04192388885587986\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch8, val_loss:0.04030516454867792, best_val_loss:0.04030516454867792, best_auc:0.9887097250687139\n",
      "Epoch:9 Batch:0 Log-loss0.050151098519563675\n",
      "Epoch:9 Batch:40 Log-loss0.04545117914676666\n",
      "Epoch:9 Batch:80 Log-loss0.045369941741228104\n",
      "Epoch:9 Batch:120 Log-loss0.03782022371888161\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:9 Batch:160 Log-loss0.04311126098036766\n",
      "Epoch:9 Batch:200 Log-loss0.03977668657898903\n",
      "Epoch:9 Batch:240 Log-loss0.04928787052631378\n",
      "Epoch:9 Batch:280 Log-loss0.04343109950423241\n",
      "Epoch:9 Batch:320 Log-loss0.045554351061582565\n",
      "Epoch:9 Batch:360 Log-loss0.028222650289535522\n",
      "Epoch:9 Batch:400 Log-loss0.02884092926979065\n",
      "Epoch:9 Batch:440 Log-loss0.04377659782767296\n",
      "Epoch:9 Batch:480 Log-loss0.0330808125436306\n",
      "Epoch:9 Batch:520 Log-loss0.04353070631623268\n",
      "Epoch:9 Batch:560 Log-loss0.03304726257920265\n",
      "Epoch average log loss:0.041231107442373675\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch9, val_loss:0.040261446192419954, best_val_loss:0.040261446192419954, best_auc:0.9891183657826991\n",
      "Epoch:10 Batch:0 Log-loss0.045131534337997437\n",
      "Epoch:10 Batch:40 Log-loss0.047467898577451706\n",
      "Epoch:10 Batch:80 Log-loss0.034875765442848206\n",
      "Epoch:10 Batch:120 Log-loss0.04786363244056702\n",
      "Epoch:10 Batch:160 Log-loss0.04445711150765419\n",
      "Epoch:10 Batch:200 Log-loss0.04032304883003235\n",
      "Epoch:10 Batch:240 Log-loss0.032879192382097244\n",
      "Epoch:10 Batch:280 Log-loss0.05887238681316376\n",
      "Epoch:10 Batch:320 Log-loss0.0532517172396183\n",
      "Epoch:10 Batch:360 Log-loss0.03857265040278435\n",
      "Epoch:10 Batch:400 Log-loss0.03881845623254776\n",
      "Epoch:10 Batch:440 Log-loss0.03307285159826279\n",
      "Epoch:10 Batch:480 Log-loss0.034376680850982666\n",
      "Epoch:10 Batch:520 Log-loss0.04055144637823105\n",
      "Epoch:10 Batch:560 Log-loss0.05012212321162224\n",
      "Epoch average log loss:0.04085008029360324\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch10, val_loss:0.03977368727117188, best_val_loss:0.03977368727117188, best_auc:0.989386164883786\n",
      "Epoch:11 Batch:0 Log-loss0.03779646381735802\n",
      "Epoch:11 Batch:40 Log-loss0.0454878956079483\n",
      "Epoch:11 Batch:80 Log-loss0.029095863923430443\n",
      "Epoch:11 Batch:120 Log-loss0.030788838863372803\n",
      "Epoch:11 Batch:160 Log-loss0.03624343499541283\n",
      "Epoch:11 Batch:200 Log-loss0.040772344917058945\n",
      "Epoch:11 Batch:240 Log-loss0.04999774321913719\n",
      "Epoch:11 Batch:280 Log-loss0.04821490868926048\n",
      "Epoch:11 Batch:320 Log-loss0.038747381418943405\n",
      "Epoch:11 Batch:360 Log-loss0.02571268565952778\n",
      "Epoch:11 Batch:400 Log-loss0.06094163656234741\n",
      "Epoch:11 Batch:440 Log-loss0.03738926723599434\n",
      "Epoch:11 Batch:480 Log-loss0.05038358271121979\n",
      "Epoch:11 Batch:520 Log-loss0.027241384610533714\n",
      "Epoch:11 Batch:560 Log-loss0.022854972630739212\n",
      "Epoch average log loss:0.04041438208660111\n",
      "In Epoch11, val_loss:0.039829531514011106, best_val_loss:0.03977368727117188, best_auc:0.989386164883786\n",
      "Epoch:12 Batch:0 Log-loss0.04677221551537514\n",
      "Epoch:12 Batch:40 Log-loss0.029183773323893547\n",
      "Epoch:12 Batch:80 Log-loss0.04931974783539772\n",
      "Epoch:12 Batch:120 Log-loss0.0399661585688591\n",
      "Epoch:12 Batch:160 Log-loss0.026339424774050713\n",
      "Epoch:12 Batch:200 Log-loss0.06050260737538338\n",
      "Epoch:12 Batch:240 Log-loss0.04522104933857918\n",
      "Epoch:12 Batch:280 Log-loss0.03312545269727707\n",
      "Epoch:12 Batch:320 Log-loss0.04461757466197014\n",
      "Epoch:12 Batch:360 Log-loss0.03456433117389679\n",
      "Epoch:12 Batch:400 Log-loss0.03689216449856758\n",
      "Epoch:12 Batch:440 Log-loss0.0341433547437191\n",
      "Epoch:12 Batch:480 Log-loss0.03078398108482361\n",
      "Epoch:12 Batch:520 Log-loss0.05249645933508873\n",
      "Epoch:12 Batch:560 Log-loss0.05256284400820732\n",
      "Epoch average log loss:0.04014662106055766\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch12, val_loss:0.039222289130702886, best_val_loss:0.039222289130702886, best_auc:0.9899842938112725\n",
      "Epoch:13 Batch:0 Log-loss0.04718777909874916\n",
      "Epoch:13 Batch:40 Log-loss0.03676667809486389\n",
      "Epoch:13 Batch:80 Log-loss0.02106073498725891\n",
      "Epoch:13 Batch:120 Log-loss0.03314919397234917\n",
      "Epoch:13 Batch:160 Log-loss0.04859912768006325\n",
      "Epoch:13 Batch:200 Log-loss0.02355322055518627\n",
      "Epoch:13 Batch:240 Log-loss0.048086777329444885\n",
      "Epoch:13 Batch:280 Log-loss0.035832393914461136\n",
      "Epoch:13 Batch:320 Log-loss0.041800398379564285\n",
      "Epoch:13 Batch:360 Log-loss0.032960113137960434\n",
      "Epoch:13 Batch:400 Log-loss0.04276496544480324\n",
      "Epoch:13 Batch:440 Log-loss0.04123242571949959\n",
      "Epoch:13 Batch:480 Log-loss0.04699636623263359\n",
      "Epoch:13 Batch:520 Log-loss0.05051588639616966\n",
      "Epoch:13 Batch:560 Log-loss0.05402117595076561\n",
      "Epoch average log loss:0.03967555497240807\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch13, val_loss:0.039198693807849926, best_val_loss:0.039198693807849926, best_auc:0.9901867392386811\n",
      "Epoch:14 Batch:0 Log-loss0.03454236313700676\n",
      "Epoch:14 Batch:40 Log-loss0.03618853911757469\n",
      "Epoch:14 Batch:80 Log-loss0.047941144555807114\n",
      "Epoch:14 Batch:120 Log-loss0.04009494557976723\n",
      "Epoch:14 Batch:160 Log-loss0.028641141951084137\n",
      "Epoch:14 Batch:200 Log-loss0.040015336126089096\n",
      "Epoch:14 Batch:240 Log-loss0.023585284128785133\n",
      "Epoch:14 Batch:280 Log-loss0.047355446964502335\n",
      "Epoch:14 Batch:320 Log-loss0.034913189709186554\n",
      "Epoch:14 Batch:360 Log-loss0.034299641847610474\n",
      "Epoch:14 Batch:400 Log-loss0.05078236386179924\n",
      "Epoch:14 Batch:440 Log-loss0.029719151556491852\n",
      "Epoch:14 Batch:480 Log-loss0.0384211428463459\n",
      "Epoch:14 Batch:520 Log-loss0.027127213776111603\n",
      "Epoch:14 Batch:560 Log-loss0.033772923052310944\n",
      "Epoch average log loss:0.039251300263484674\n",
      "In Epoch14, val_loss:0.04033961500497752, best_val_loss:0.039198693807849926, best_auc:0.9901867392386811\n",
      "Epoch:15 Batch:0 Log-loss0.03965131565928459\n",
      "Epoch:15 Batch:40 Log-loss0.039349693804979324\n",
      "Epoch:15 Batch:80 Log-loss0.03885135427117348\n",
      "Epoch:15 Batch:120 Log-loss0.04208354279398918\n",
      "Epoch:15 Batch:160 Log-loss0.041844308376312256\n",
      "Epoch:15 Batch:200 Log-loss0.03189607709646225\n",
      "Epoch:15 Batch:240 Log-loss0.03427660092711449\n",
      "Epoch:15 Batch:280 Log-loss0.03938312456011772\n",
      "Epoch:15 Batch:320 Log-loss0.039807893335819244\n",
      "Epoch:15 Batch:360 Log-loss0.037440165877342224\n",
      "Epoch:15 Batch:400 Log-loss0.030837414786219597\n",
      "Epoch:15 Batch:440 Log-loss0.06343341618776321\n",
      "Epoch:15 Batch:480 Log-loss0.04550078138709068\n",
      "Epoch:15 Batch:520 Log-loss0.03794003650546074\n",
      "Epoch:15 Batch:560 Log-loss0.03551972284913063\n",
      "Epoch average log loss:0.03899067771687571\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch15, val_loss:0.038929982527864705, best_val_loss:0.038929982527864705, best_auc:0.9901347726813053\n",
      "Epoch:16 Batch:0 Log-loss0.045414555817842484\n",
      "Epoch:16 Batch:40 Log-loss0.03844379261136055\n",
      "Epoch:16 Batch:80 Log-loss0.03842873498797417\n",
      "Epoch:16 Batch:120 Log-loss0.03954574093222618\n",
      "Epoch:16 Batch:160 Log-loss0.04263973608613014\n",
      "Epoch:16 Batch:200 Log-loss0.045617058873176575\n",
      "Epoch:16 Batch:240 Log-loss0.03223039209842682\n",
      "Epoch:16 Batch:280 Log-loss0.06754780560731888\n",
      "Epoch:16 Batch:320 Log-loss0.029176821932196617\n",
      "Epoch:16 Batch:360 Log-loss0.05417674407362938\n",
      "Epoch:16 Batch:400 Log-loss0.03184359148144722\n",
      "Epoch:16 Batch:440 Log-loss0.03165509179234505\n",
      "Epoch:16 Batch:480 Log-loss0.033398136496543884\n",
      "Epoch:16 Batch:520 Log-loss0.03569011762738228\n",
      "Epoch:16 Batch:560 Log-loss0.0414452962577343\n",
      "Epoch average log loss:0.038907278907884445\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch16, val_loss:0.03888889618450993, best_val_loss:0.03888889618450993, best_auc:0.9903106295649841\n",
      "Epoch:17 Batch:0 Log-loss0.04507075622677803\n",
      "Epoch:17 Batch:40 Log-loss0.03674500063061714\n",
      "Epoch:17 Batch:80 Log-loss0.05949462950229645\n",
      "Epoch:17 Batch:120 Log-loss0.023588085547089577\n",
      "Epoch:17 Batch:160 Log-loss0.04266424477100372\n",
      "Epoch:17 Batch:200 Log-loss0.020973024889826775\n",
      "Epoch:17 Batch:240 Log-loss0.025667110458016396\n",
      "Epoch:17 Batch:280 Log-loss0.045229192823171616\n",
      "Epoch:17 Batch:320 Log-loss0.03583383187651634\n",
      "Epoch:17 Batch:360 Log-loss0.04195235297083855\n",
      "Epoch:17 Batch:400 Log-loss0.033191293478012085\n",
      "Epoch:17 Batch:440 Log-loss0.03989003971219063\n",
      "Epoch:17 Batch:480 Log-loss0.03054330311715603\n",
      "Epoch:17 Batch:520 Log-loss0.05054328963160515\n",
      "Epoch:17 Batch:560 Log-loss0.027951808646321297\n",
      "Epoch average log loss:0.0386640767965998\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch17, val_loss:0.03884987115396534, best_val_loss:0.03884987115396534, best_auc:0.9903088861354467\n",
      "Epoch:18 Batch:0 Log-loss0.043490126729011536\n",
      "Epoch:18 Batch:40 Log-loss0.03948910906910896\n",
      "Epoch:18 Batch:80 Log-loss0.037463318556547165\n",
      "Epoch:18 Batch:120 Log-loss0.0439317412674427\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:18 Batch:160 Log-loss0.035658031702041626\n",
      "Epoch:18 Batch:200 Log-loss0.04560533165931702\n",
      "Epoch:18 Batch:240 Log-loss0.06093995273113251\n",
      "Epoch:18 Batch:280 Log-loss0.04654531553387642\n",
      "Epoch:18 Batch:320 Log-loss0.03224104270339012\n",
      "Epoch:18 Batch:360 Log-loss0.04050925374031067\n",
      "Epoch:18 Batch:400 Log-loss0.037411078810691833\n",
      "Epoch:18 Batch:440 Log-loss0.02660735882818699\n",
      "Epoch:18 Batch:480 Log-loss0.048006296157836914\n",
      "Epoch:18 Batch:520 Log-loss0.03775300830602646\n",
      "Epoch:18 Batch:560 Log-loss0.04408837482333183\n",
      "Epoch average log loss:0.03856381814527724\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch18, val_loss:0.038620384849044447, best_val_loss:0.038620384849044447, best_auc:0.9905910215615558\n",
      "Epoch:19 Batch:0 Log-loss0.03824811056256294\n",
      "Epoch:19 Batch:40 Log-loss0.016808228567242622\n",
      "Epoch:19 Batch:80 Log-loss0.035543378442525864\n",
      "Epoch:19 Batch:120 Log-loss0.03020019829273224\n",
      "Epoch:19 Batch:160 Log-loss0.03790932148694992\n",
      "Epoch:19 Batch:200 Log-loss0.024756601080298424\n",
      "Epoch:19 Batch:240 Log-loss0.04928004741668701\n",
      "Epoch:19 Batch:280 Log-loss0.05283847823739052\n",
      "Epoch:19 Batch:320 Log-loss0.04006388783454895\n",
      "Epoch:19 Batch:360 Log-loss0.04206452891230583\n",
      "Epoch:19 Batch:400 Log-loss0.038195665925741196\n",
      "Epoch:19 Batch:440 Log-loss0.02697364054620266\n",
      "Epoch:19 Batch:480 Log-loss0.026489147916436195\n",
      "Epoch:19 Batch:520 Log-loss0.052806999534368515\n",
      "Epoch:19 Batch:560 Log-loss0.04286741837859154\n",
      "Epoch average log loss:0.03834231591650418\n",
      "In Epoch19, val_loss:0.038787833925688395, best_val_loss:0.038620384849044447, best_auc:0.9905910215615558\n",
      "Epoch:20 Batch:0 Log-loss0.037665434181690216\n",
      "Epoch:20 Batch:40 Log-loss0.03730902820825577\n",
      "Epoch:20 Batch:80 Log-loss0.04601161554455757\n",
      "Epoch:20 Batch:120 Log-loss0.031297508627176285\n",
      "Epoch:20 Batch:160 Log-loss0.03481985628604889\n",
      "Epoch:20 Batch:200 Log-loss0.03356829285621643\n",
      "Epoch:20 Batch:240 Log-loss0.03212932124733925\n",
      "Epoch:20 Batch:280 Log-loss0.03495281934738159\n",
      "Epoch:20 Batch:320 Log-loss0.036505430936813354\n",
      "Epoch:20 Batch:360 Log-loss0.05246586725115776\n",
      "Epoch:20 Batch:400 Log-loss0.03553833067417145\n",
      "Epoch:20 Batch:440 Log-loss0.02698281966149807\n",
      "Epoch:20 Batch:480 Log-loss0.04356405511498451\n",
      "Epoch:20 Batch:520 Log-loss0.034976013004779816\n",
      "Epoch:20 Batch:560 Log-loss0.03769974038004875\n",
      "Epoch average log loss:0.038121999883359035\n",
      "In Epoch20, val_loss:0.03873702317610823, best_val_loss:0.038620384849044447, best_auc:0.9905910215615558\n",
      "Epoch:21 Batch:0 Log-loss0.0518709272146225\n",
      "Epoch:21 Batch:40 Log-loss0.03208162263035774\n",
      "Epoch:21 Batch:80 Log-loss0.04099683836102486\n",
      "Epoch:21 Batch:120 Log-loss0.027261776849627495\n",
      "Epoch:21 Batch:160 Log-loss0.052843887358903885\n",
      "Epoch:21 Batch:200 Log-loss0.030533133074641228\n",
      "Epoch:21 Batch:240 Log-loss0.04123115539550781\n",
      "Epoch:21 Batch:280 Log-loss0.04420395568013191\n",
      "Epoch:21 Batch:320 Log-loss0.045468881726264954\n",
      "Epoch:21 Batch:360 Log-loss0.031339049339294434\n",
      "Epoch:21 Batch:400 Log-loss0.040993642061948776\n",
      "Epoch:21 Batch:440 Log-loss0.03609446808695793\n",
      "Epoch:21 Batch:480 Log-loss0.04449297860264778\n",
      "Epoch:21 Batch:520 Log-loss0.04902457818388939\n",
      "Epoch:21 Batch:560 Log-loss0.04019046574831009\n",
      "Epoch average log loss:0.03807020116863506\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch21, val_loss:0.03860914148535799, best_val_loss:0.03860914148535799, best_auc:0.9906862263412717\n",
      "Epoch:22 Batch:0 Log-loss0.04125721752643585\n",
      "Epoch:22 Batch:40 Log-loss0.036646194756031036\n",
      "Epoch:22 Batch:80 Log-loss0.033578500151634216\n",
      "Epoch:22 Batch:120 Log-loss0.045598868280649185\n",
      "Epoch:22 Batch:160 Log-loss0.03308936953544617\n",
      "Epoch:22 Batch:200 Log-loss0.03228260949254036\n",
      "Epoch:22 Batch:240 Log-loss0.019516972824931145\n",
      "Epoch:22 Batch:280 Log-loss0.03570167347788811\n",
      "Epoch:22 Batch:320 Log-loss0.04614843428134918\n",
      "Epoch:22 Batch:360 Log-loss0.03747442364692688\n",
      "Epoch:22 Batch:400 Log-loss0.03532399237155914\n",
      "Epoch:22 Batch:440 Log-loss0.028532812371850014\n",
      "Epoch:22 Batch:480 Log-loss0.0471409447491169\n",
      "Epoch:22 Batch:520 Log-loss0.04150479659438133\n",
      "Epoch:22 Batch:560 Log-loss0.024473773315548897\n",
      "Epoch average log loss:0.037857139612814145\n",
      "In Epoch22, val_loss:0.038773913750438954, best_val_loss:0.03860914148535799, best_auc:0.9906862263412717\n",
      "Epoch:23 Batch:0 Log-loss0.03435398265719414\n",
      "Epoch:23 Batch:40 Log-loss0.031508397310972214\n",
      "Epoch:23 Batch:80 Log-loss0.052202384918928146\n",
      "Epoch:23 Batch:120 Log-loss0.039687562733888626\n",
      "Epoch:23 Batch:160 Log-loss0.020815489813685417\n",
      "Epoch:23 Batch:200 Log-loss0.037799376994371414\n",
      "Epoch:23 Batch:240 Log-loss0.040256623178720474\n",
      "Epoch:23 Batch:280 Log-loss0.04293017089366913\n",
      "Epoch:23 Batch:320 Log-loss0.06483500450849533\n",
      "Epoch:23 Batch:360 Log-loss0.03622003272175789\n",
      "Epoch:23 Batch:400 Log-loss0.03607969358563423\n",
      "Epoch:23 Batch:440 Log-loss0.03892744332551956\n",
      "Epoch:23 Batch:480 Log-loss0.03696915879845619\n",
      "Epoch:23 Batch:520 Log-loss0.045494239777326584\n",
      "Epoch:23 Batch:560 Log-loss0.03866804763674736\n",
      "Epoch average log loss:0.03778487664447831\n",
      "In Epoch23, val_loss:0.038827857346349956, best_val_loss:0.03860914148535799, best_auc:0.9906862263412717\n",
      "Epoch:24 Batch:0 Log-loss0.03570084646344185\n",
      "Epoch:24 Batch:40 Log-loss0.025821663439273834\n",
      "Epoch:24 Batch:80 Log-loss0.03048795275390148\n",
      "Epoch:24 Batch:120 Log-loss0.03638641908764839\n",
      "Epoch:24 Batch:160 Log-loss0.036644432693719864\n",
      "Epoch:24 Batch:200 Log-loss0.046839918941259384\n",
      "Epoch:24 Batch:240 Log-loss0.03636566177010536\n",
      "Epoch:24 Batch:280 Log-loss0.032241448760032654\n",
      "Epoch:24 Batch:320 Log-loss0.04180486872792244\n",
      "Epoch:24 Batch:360 Log-loss0.03172941505908966\n",
      "Epoch:24 Batch:400 Log-loss0.03389356657862663\n",
      "Epoch:24 Batch:440 Log-loss0.0296777356415987\n",
      "Epoch:24 Batch:480 Log-loss0.03313636779785156\n",
      "Epoch:24 Batch:520 Log-loss0.0388638898730278\n",
      "Epoch:24 Batch:560 Log-loss0.035236094146966934\n",
      "Epoch average log loss:0.03762112796373133\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch24, val_loss:0.03849574537201767, best_val_loss:0.03849574537201767, best_auc:0.9907907084918977\n",
      "Epoch:25 Batch:0 Log-loss0.031161269173026085\n",
      "Epoch:25 Batch:40 Log-loss0.038337770849466324\n",
      "Epoch:25 Batch:80 Log-loss0.032460059970617294\n",
      "Epoch:25 Batch:120 Log-loss0.04323480650782585\n",
      "Epoch:25 Batch:160 Log-loss0.031995631754398346\n",
      "Epoch:25 Batch:200 Log-loss0.03215109929442406\n",
      "Epoch:25 Batch:240 Log-loss0.04427240416407585\n",
      "Epoch:25 Batch:280 Log-loss0.047164905816316605\n",
      "Epoch:25 Batch:320 Log-loss0.039971332997083664\n",
      "Epoch:25 Batch:360 Log-loss0.044966887682676315\n",
      "Epoch:25 Batch:400 Log-loss0.04047926887869835\n",
      "Epoch:25 Batch:440 Log-loss0.03950197622179985\n",
      "Epoch:25 Batch:480 Log-loss0.037716399878263474\n",
      "Epoch:25 Batch:520 Log-loss0.03896959125995636\n",
      "Epoch:25 Batch:560 Log-loss0.03891372308135033\n",
      "Epoch average log loss:0.037471299738224065\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch25, val_loss:0.03844808302013479, best_val_loss:0.03844808302013479, best_auc:0.990704355678504\n",
      "Epoch:26 Batch:0 Log-loss0.044050801545381546\n",
      "Epoch:26 Batch:40 Log-loss0.0463310182094574\n",
      "Epoch:26 Batch:80 Log-loss0.03189026191830635\n",
      "Epoch:26 Batch:120 Log-loss0.04140562191605568\n",
      "Epoch:26 Batch:160 Log-loss0.047172579914331436\n",
      "Epoch:26 Batch:200 Log-loss0.02748330868780613\n",
      "Epoch:26 Batch:240 Log-loss0.03763885796070099\n",
      "Epoch:26 Batch:280 Log-loss0.03944234922528267\n",
      "Epoch:26 Batch:320 Log-loss0.04940888285636902\n",
      "Epoch:26 Batch:360 Log-loss0.0373309962451458\n",
      "Epoch:26 Batch:400 Log-loss0.026429444551467896\n",
      "Epoch:26 Batch:440 Log-loss0.024788638576865196\n",
      "Epoch:26 Batch:480 Log-loss0.03926298767328262\n",
      "Epoch:26 Batch:520 Log-loss0.027615239843726158\n",
      "Epoch:26 Batch:560 Log-loss0.029569333419203758\n",
      "Epoch average log loss:0.037461761732785294\n",
      "In Epoch26, val_loss:0.038568641844131966, best_val_loss:0.03844808302013479, best_auc:0.990704355678504\n",
      "Epoch:27 Batch:0 Log-loss0.037600722163915634\n",
      "Epoch:27 Batch:40 Log-loss0.0622781477868557\n",
      "Epoch:27 Batch:80 Log-loss0.04079475626349449\n",
      "Epoch:27 Batch:120 Log-loss0.0368555448949337\n",
      "Epoch:27 Batch:160 Log-loss0.027403339743614197\n",
      "Epoch:27 Batch:200 Log-loss0.058875199407339096\n",
      "Epoch:27 Batch:240 Log-loss0.03356587514281273\n",
      "Epoch:27 Batch:280 Log-loss0.04109354689717293\n",
      "Epoch:27 Batch:320 Log-loss0.04953471198678017\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:27 Batch:360 Log-loss0.03544208034873009\n",
      "Epoch:27 Batch:400 Log-loss0.031754668802022934\n",
      "Epoch:27 Batch:440 Log-loss0.04824363812804222\n",
      "Epoch:27 Batch:480 Log-loss0.048651132732629776\n",
      "Epoch:27 Batch:520 Log-loss0.024600772187113762\n",
      "Epoch:27 Batch:560 Log-loss0.04342334344983101\n",
      "Epoch average log loss:0.03731697032095066\n",
      "In Epoch27, val_loss:0.03860016198210927, best_val_loss:0.03844808302013479, best_auc:0.990704355678504\n",
      "Epoch:28 Batch:0 Log-loss0.028779620304703712\n",
      "Epoch:28 Batch:40 Log-loss0.03864942863583565\n",
      "Epoch:28 Batch:80 Log-loss0.03994976356625557\n",
      "Epoch:28 Batch:120 Log-loss0.05983389914035797\n",
      "Epoch:28 Batch:160 Log-loss0.0371534563601017\n",
      "Epoch:28 Batch:200 Log-loss0.03952157124876976\n",
      "Epoch:28 Batch:240 Log-loss0.023651396855711937\n",
      "Epoch:28 Batch:280 Log-loss0.022818179801106453\n",
      "Epoch:28 Batch:320 Log-loss0.05143021047115326\n",
      "Epoch:28 Batch:360 Log-loss0.04304340109229088\n",
      "Epoch:28 Batch:400 Log-loss0.04066958650946617\n",
      "Epoch:28 Batch:440 Log-loss0.035157933831214905\n",
      "Epoch:28 Batch:480 Log-loss0.06338252872228622\n",
      "Epoch:28 Batch:520 Log-loss0.04667341336607933\n",
      "Epoch:28 Batch:560 Log-loss0.03921434283256531\n",
      "Epoch average log loss:0.03730056378325181\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch28, val_loss:0.03840601453386478, best_val_loss:0.03840601453386478, best_auc:0.9906998048446368\n",
      "Epoch:29 Batch:0 Log-loss0.030716976150870323\n",
      "Epoch:29 Batch:40 Log-loss0.031993985176086426\n",
      "Epoch:29 Batch:80 Log-loss0.04391564801335335\n",
      "Epoch:29 Batch:120 Log-loss0.041894182562828064\n",
      "Epoch:29 Batch:160 Log-loss0.03747974708676338\n",
      "Epoch:29 Batch:200 Log-loss0.04364881291985512\n",
      "Epoch:29 Batch:240 Log-loss0.03627510741353035\n",
      "Epoch:29 Batch:280 Log-loss0.03666702285408974\n",
      "Epoch:29 Batch:320 Log-loss0.0438711978495121\n",
      "Epoch:29 Batch:360 Log-loss0.04102357476949692\n",
      "Epoch:29 Batch:400 Log-loss0.033269252628088\n",
      "Epoch:29 Batch:440 Log-loss0.038755595684051514\n",
      "Epoch:29 Batch:480 Log-loss0.03221834450960159\n",
      "Epoch:29 Batch:520 Log-loss0.028882311657071114\n",
      "Epoch:29 Batch:560 Log-loss0.03626720979809761\n",
      "Epoch average log loss:0.03727523689158261\n",
      "In Epoch29, val_loss:0.03842357302188977, best_val_loss:0.03840601453386478, best_auc:0.9906998048446368\n",
      "Epoch:30 Batch:0 Log-loss0.03754932060837746\n",
      "Epoch:30 Batch:40 Log-loss0.03773937374353409\n",
      "Epoch:30 Batch:80 Log-loss0.04163781926035881\n",
      "Epoch:30 Batch:120 Log-loss0.03706894442439079\n",
      "Epoch:30 Batch:160 Log-loss0.02271272800862789\n",
      "Epoch:30 Batch:200 Log-loss0.02983774058520794\n",
      "Epoch:30 Batch:240 Log-loss0.034047454595565796\n",
      "Epoch:30 Batch:280 Log-loss0.027896417304873466\n",
      "Epoch:30 Batch:320 Log-loss0.04202241078019142\n",
      "Epoch:30 Batch:360 Log-loss0.02835424244403839\n",
      "Epoch:30 Batch:400 Log-loss0.04495355859398842\n",
      "Epoch:30 Batch:440 Log-loss0.03950143977999687\n",
      "Epoch:30 Batch:480 Log-loss0.03773651644587517\n",
      "Epoch:30 Batch:520 Log-loss0.039787475019693375\n",
      "Epoch:30 Batch:560 Log-loss0.031885456293821335\n",
      "Epoch average log loss:0.03719664260445695\n",
      "In Epoch30, val_loss:0.038552171551959245, best_val_loss:0.03840601453386478, best_auc:0.9906998048446368\n",
      "Epoch:31 Batch:0 Log-loss0.033192191272974014\n",
      "Epoch:31 Batch:40 Log-loss0.024771980941295624\n",
      "Epoch:31 Batch:80 Log-loss0.03626633808016777\n",
      "Epoch:31 Batch:120 Log-loss0.03491505607962608\n",
      "Epoch:31 Batch:160 Log-loss0.033888738602399826\n",
      "Epoch:31 Batch:200 Log-loss0.030451186001300812\n",
      "Epoch:31 Batch:240 Log-loss0.030876556411385536\n",
      "Epoch:31 Batch:280 Log-loss0.04344412311911583\n",
      "Epoch:31 Batch:320 Log-loss0.03118606097996235\n",
      "Epoch:31 Batch:360 Log-loss0.03209656849503517\n",
      "Epoch:31 Batch:400 Log-loss0.030244870111346245\n",
      "Epoch:31 Batch:440 Log-loss0.0444624237716198\n",
      "Epoch:31 Batch:480 Log-loss0.048751283437013626\n",
      "Epoch:31 Batch:520 Log-loss0.04406009241938591\n",
      "Epoch:31 Batch:560 Log-loss0.036622244864702225\n",
      "Epoch average log loss:0.03734571583835142\n",
      "In Epoch31, val_loss:0.038612387366224486, best_val_loss:0.03840601453386478, best_auc:0.9906998048446368\n",
      "Epoch:32 Batch:0 Log-loss0.04348750412464142\n",
      "Epoch:32 Batch:40 Log-loss0.037636298686265945\n",
      "Epoch:32 Batch:80 Log-loss0.026739375665783882\n",
      "Epoch:32 Batch:120 Log-loss0.03105720691382885\n",
      "Epoch:32 Batch:160 Log-loss0.03576705977320671\n",
      "Epoch:32 Batch:200 Log-loss0.04317277669906616\n",
      "Epoch:32 Batch:240 Log-loss0.03142328932881355\n",
      "Epoch:32 Batch:280 Log-loss0.026663629338145256\n",
      "Epoch:32 Batch:320 Log-loss0.04647092521190643\n",
      "Epoch:32 Batch:360 Log-loss0.038113050162792206\n",
      "Epoch:32 Batch:400 Log-loss0.04061716049909592\n",
      "Epoch:32 Batch:440 Log-loss0.044126879423856735\n",
      "Epoch:32 Batch:480 Log-loss0.03520957753062248\n",
      "Epoch:32 Batch:520 Log-loss0.02731192857027054\n",
      "Epoch:32 Batch:560 Log-loss0.04557904601097107\n",
      "Epoch average log loss:0.03701077942137739\n",
      "In Epoch32, val_loss:0.03863865068520473, best_val_loss:0.03840601453386478, best_auc:0.9906998048446368\n",
      "Epoch:33 Batch:0 Log-loss0.03793955594301224\n",
      "Epoch:33 Batch:40 Log-loss0.026874035596847534\n",
      "Epoch:33 Batch:80 Log-loss0.02976515144109726\n",
      "Epoch:33 Batch:120 Log-loss0.03843171149492264\n",
      "Epoch:33 Batch:160 Log-loss0.03389664366841316\n",
      "Epoch:33 Batch:200 Log-loss0.06130934879183769\n",
      "Epoch:33 Batch:240 Log-loss0.0499783419072628\n",
      "Epoch:33 Batch:280 Log-loss0.04718082770705223\n",
      "Epoch:33 Batch:320 Log-loss0.02996784634888172\n",
      "Epoch:33 Batch:360 Log-loss0.023934124037623405\n",
      "Epoch:33 Batch:400 Log-loss0.03934333473443985\n",
      "Epoch:33 Batch:440 Log-loss0.029566390439867973\n",
      "Epoch:33 Batch:480 Log-loss0.03091312386095524\n",
      "Epoch:33 Batch:520 Log-loss0.04483003541827202\n",
      "Epoch:33 Batch:560 Log-loss0.03527807071805\n",
      "Epoch average log loss:0.03686804615759424\n",
      "In Epoch33, val_loss:0.03843577219404854, best_val_loss:0.03840601453386478, best_auc:0.9906998048446368\n",
      "Epoch:34 Batch:0 Log-loss0.023326819762587547\n",
      "Epoch:34 Batch:40 Log-loss0.030296437442302704\n",
      "Epoch:34 Batch:80 Log-loss0.03251088783144951\n",
      "Epoch:34 Batch:120 Log-loss0.04086371883749962\n",
      "Epoch:34 Batch:160 Log-loss0.028152748942375183\n",
      "Epoch:34 Batch:200 Log-loss0.051179107278585434\n",
      "Epoch:34 Batch:240 Log-loss0.037547752261161804\n",
      "Epoch:34 Batch:280 Log-loss0.02845984697341919\n",
      "Epoch:34 Batch:320 Log-loss0.0370509959757328\n",
      "Epoch:34 Batch:360 Log-loss0.028256280347704887\n",
      "Epoch:34 Batch:400 Log-loss0.024415135383605957\n",
      "Epoch:34 Batch:440 Log-loss0.031306199729442596\n",
      "Epoch:34 Batch:480 Log-loss0.03698015585541725\n",
      "Epoch:34 Batch:520 Log-loss0.0305596012622118\n",
      "Epoch:34 Batch:560 Log-loss0.025625402107834816\n",
      "Epoch average log loss:0.03689587689670069\n",
      "In Epoch34, val_loss:0.03849751766373552, best_val_loss:0.03840601453386478, best_auc:0.9906998048446368\n",
      "Epoch:35 Batch:0 Log-loss0.039904773235321045\n",
      "Epoch:35 Batch:40 Log-loss0.035961542278528214\n",
      "Epoch:35 Batch:80 Log-loss0.03595263510942459\n",
      "Epoch:35 Batch:120 Log-loss0.047597382217645645\n",
      "Epoch:35 Batch:160 Log-loss0.026431681588292122\n",
      "Epoch:35 Batch:200 Log-loss0.026773296296596527\n",
      "Epoch:35 Batch:240 Log-loss0.04337514936923981\n",
      "Epoch:35 Batch:280 Log-loss0.031872641295194626\n",
      "Epoch:35 Batch:320 Log-loss0.021754881367087364\n",
      "Epoch:35 Batch:360 Log-loss0.03169238939881325\n",
      "Epoch:35 Batch:400 Log-loss0.03323803469538689\n",
      "Epoch:35 Batch:440 Log-loss0.040912747383117676\n",
      "Epoch:35 Batch:480 Log-loss0.03687291219830513\n",
      "Epoch:35 Batch:520 Log-loss0.031442031264305115\n",
      "Epoch:35 Batch:560 Log-loss0.028161093592643738\n",
      "Epoch average log loss:0.036895236436144584\n",
      "In Epoch35, val_loss:0.038443186886709134, best_val_loss:0.03840601453386478, best_auc:0.9906998048446368\n",
      "Epoch:36 Batch:0 Log-loss0.03214505314826965\n",
      "Epoch:36 Batch:40 Log-loss0.045181263238191605\n",
      "Epoch:36 Batch:80 Log-loss0.03401315212249756\n",
      "Epoch:36 Batch:120 Log-loss0.03450533375144005\n",
      "Epoch:36 Batch:160 Log-loss0.04220523312687874\n",
      "Epoch:36 Batch:200 Log-loss0.03802959620952606\n",
      "Epoch:36 Batch:240 Log-loss0.0317801758646965\n",
      "Epoch:36 Batch:280 Log-loss0.03484896942973137\n",
      "Epoch:36 Batch:320 Log-loss0.04205971583724022\n",
      "Epoch:36 Batch:360 Log-loss0.0322701632976532\n",
      "Epoch:36 Batch:400 Log-loss0.03735718876123428\n",
      "Epoch:36 Batch:440 Log-loss0.032419826835393906\n",
      "Epoch:36 Batch:480 Log-loss0.03437837213277817\n",
      "Epoch:36 Batch:520 Log-loss0.03026527911424637\n",
      "Epoch:36 Batch:560 Log-loss0.030183492228388786\n",
      "Epoch average log loss:0.03675426099050258\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch36, val_loss:0.038372506614270004, best_val_loss:0.038372506614270004, best_auc:0.9907789462037643\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:37 Batch:0 Log-loss0.03646939992904663\n",
      "Epoch:37 Batch:40 Log-loss0.029216179624199867\n",
      "Epoch:37 Batch:80 Log-loss0.036086030304431915\n",
      "Epoch:37 Batch:120 Log-loss0.03187054023146629\n",
      "Epoch:37 Batch:160 Log-loss0.03597589209675789\n",
      "Epoch:37 Batch:200 Log-loss0.040180642157793045\n",
      "Epoch:37 Batch:240 Log-loss0.0369984395802021\n",
      "Epoch:37 Batch:280 Log-loss0.032506559044122696\n",
      "Epoch:37 Batch:320 Log-loss0.04712832346558571\n",
      "Epoch:37 Batch:360 Log-loss0.03497335687279701\n",
      "Epoch:37 Batch:400 Log-loss0.0449063815176487\n",
      "Epoch:37 Batch:440 Log-loss0.0633268728852272\n",
      "Epoch:37 Batch:480 Log-loss0.027778228744864464\n",
      "Epoch:37 Batch:520 Log-loss0.031030038371682167\n",
      "Epoch:37 Batch:560 Log-loss0.027357565239071846\n",
      "Epoch average log loss:0.03663257978457425\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch37, val_loss:0.038327332179147876, best_val_loss:0.038327332179147876, best_auc:0.9907846981884747\n",
      "Epoch:38 Batch:0 Log-loss0.0414077527821064\n",
      "Epoch:38 Batch:40 Log-loss0.035441018640995026\n",
      "Epoch:38 Batch:80 Log-loss0.018712541088461876\n",
      "Epoch:38 Batch:120 Log-loss0.03802482411265373\n",
      "Epoch:38 Batch:160 Log-loss0.02863258868455887\n",
      "Epoch:38 Batch:200 Log-loss0.02431757003068924\n",
      "Epoch:38 Batch:240 Log-loss0.02715292014181614\n",
      "Epoch:38 Batch:280 Log-loss0.0347268246114254\n",
      "Epoch:38 Batch:320 Log-loss0.04147724434733391\n",
      "Epoch:38 Batch:360 Log-loss0.031413111835718155\n",
      "Epoch:38 Batch:400 Log-loss0.03198281675577164\n",
      "Epoch:38 Batch:440 Log-loss0.028671547770500183\n",
      "Epoch:38 Batch:480 Log-loss0.02374567650258541\n",
      "Epoch:38 Batch:520 Log-loss0.047395963221788406\n",
      "Epoch:38 Batch:560 Log-loss0.028944676741957664\n",
      "Epoch average log loss:0.03676470772562815\n",
      "In Epoch38, val_loss:0.03833293636328325, best_val_loss:0.038327332179147876, best_auc:0.9907846981884747\n",
      "Epoch:39 Batch:0 Log-loss0.052718598395586014\n",
      "Epoch:39 Batch:40 Log-loss0.02582903765141964\n",
      "Epoch:39 Batch:80 Log-loss0.03480033203959465\n",
      "Epoch:39 Batch:120 Log-loss0.041168708354234695\n",
      "Epoch:39 Batch:160 Log-loss0.03809254243969917\n",
      "Epoch:39 Batch:200 Log-loss0.028193876147270203\n",
      "Epoch:39 Batch:240 Log-loss0.03669868782162666\n",
      "Epoch:39 Batch:280 Log-loss0.03743292763829231\n",
      "Epoch:39 Batch:320 Log-loss0.03942251205444336\n",
      "Epoch:39 Batch:360 Log-loss0.03371391445398331\n",
      "Epoch:39 Batch:400 Log-loss0.029498165473341942\n",
      "Epoch:39 Batch:440 Log-loss0.04345904290676117\n",
      "Epoch:39 Batch:480 Log-loss0.026235148310661316\n",
      "Epoch:39 Batch:520 Log-loss0.04128080606460571\n",
      "Epoch:39 Batch:560 Log-loss0.032650768756866455\n",
      "Epoch average log loss:0.03671045389990987\n",
      "In Epoch39, val_loss:0.03842335751080899, best_val_loss:0.038327332179147876, best_auc:0.9907846981884747\n",
      "Epoch:40 Batch:0 Log-loss0.032605838030576706\n",
      "Epoch:40 Batch:40 Log-loss0.03644023835659027\n",
      "Epoch:40 Batch:80 Log-loss0.03453151509165764\n",
      "Epoch:40 Batch:120 Log-loss0.04729697108268738\n",
      "Epoch:40 Batch:160 Log-loss0.03937115520238876\n",
      "Epoch:40 Batch:200 Log-loss0.02595549263060093\n",
      "Epoch:40 Batch:240 Log-loss0.05045941844582558\n",
      "Epoch:40 Batch:280 Log-loss0.033892568200826645\n",
      "Epoch:40 Batch:320 Log-loss0.026689579710364342\n",
      "Epoch:40 Batch:360 Log-loss0.0273128654807806\n",
      "Epoch:40 Batch:400 Log-loss0.034342486411333084\n",
      "Epoch:40 Batch:440 Log-loss0.02225331962108612\n",
      "Epoch:40 Batch:480 Log-loss0.040037546306848526\n",
      "Epoch:40 Batch:520 Log-loss0.039445336908102036\n",
      "Epoch:40 Batch:560 Log-loss0.04383651167154312\n",
      "Epoch average log loss:0.03654735396682684\n",
      "In Epoch40, val_loss:0.03839356245074562, best_val_loss:0.038327332179147876, best_auc:0.9907846981884747\n",
      "Epoch:41 Batch:0 Log-loss0.04139534756541252\n",
      "Epoch:41 Batch:40 Log-loss0.03547089919447899\n",
      "Epoch:41 Batch:80 Log-loss0.052432671189308167\n",
      "Epoch:41 Batch:120 Log-loss0.03816361725330353\n",
      "Epoch:41 Batch:160 Log-loss0.04670645296573639\n",
      "Epoch:41 Batch:200 Log-loss0.02934950403869152\n",
      "Epoch:41 Batch:240 Log-loss0.0482211709022522\n",
      "Epoch:41 Batch:280 Log-loss0.0365435853600502\n",
      "Epoch:41 Batch:320 Log-loss0.032821886241436005\n",
      "Epoch:41 Batch:360 Log-loss0.03405686095356941\n",
      "Epoch:41 Batch:400 Log-loss0.042547330260276794\n",
      "Epoch:41 Batch:440 Log-loss0.03161971643567085\n",
      "Epoch:41 Batch:480 Log-loss0.03724794462323189\n",
      "Epoch:41 Batch:520 Log-loss0.030012456700205803\n",
      "Epoch:41 Batch:560 Log-loss0.02660941518843174\n",
      "Epoch average log loss:0.036596343347004484\n",
      "In Epoch41, val_loss:0.038428551899730955, best_val_loss:0.038327332179147876, best_auc:0.9907846981884747\n",
      "Epoch:42 Batch:0 Log-loss0.035900622606277466\n",
      "Epoch:42 Batch:40 Log-loss0.030352100729942322\n",
      "Epoch:42 Batch:80 Log-loss0.0342310331761837\n",
      "Epoch:42 Batch:120 Log-loss0.03726394847035408\n",
      "Epoch:42 Batch:160 Log-loss0.03620383143424988\n",
      "Epoch:42 Batch:200 Log-loss0.04217027127742767\n",
      "Epoch:42 Batch:240 Log-loss0.027581697329878807\n",
      "Epoch:42 Batch:280 Log-loss0.03772306814789772\n",
      "Epoch:42 Batch:320 Log-loss0.03798210993409157\n",
      "Epoch:42 Batch:360 Log-loss0.04986685886979103\n",
      "Epoch:42 Batch:400 Log-loss0.03039456345140934\n",
      "Epoch:42 Batch:440 Log-loss0.04386422038078308\n",
      "Epoch:42 Batch:480 Log-loss0.028385959565639496\n",
      "Epoch:42 Batch:520 Log-loss0.04903498291969299\n",
      "Epoch:42 Batch:560 Log-loss0.03319163993000984\n",
      "Epoch average log loss:0.03664262499327638\n",
      "In Epoch42, val_loss:0.03840637369150251, best_val_loss:0.038327332179147876, best_auc:0.9907846981884747\n",
      "Epoch:43 Batch:0 Log-loss0.032297369092702866\n",
      "Epoch:43 Batch:40 Log-loss0.03740040585398674\n",
      "Epoch:43 Batch:80 Log-loss0.03923671692609787\n",
      "Epoch:43 Batch:120 Log-loss0.04119645804166794\n",
      "Epoch:43 Batch:160 Log-loss0.023576175794005394\n",
      "Epoch:43 Batch:200 Log-loss0.05089433491230011\n",
      "Epoch:43 Batch:240 Log-loss0.03735814243555069\n",
      "Epoch:43 Batch:280 Log-loss0.032177507877349854\n",
      "Epoch:43 Batch:320 Log-loss0.04307720065116882\n",
      "Epoch:43 Batch:360 Log-loss0.033533718436956406\n",
      "Epoch:43 Batch:400 Log-loss0.03149111196398735\n",
      "Epoch:43 Batch:440 Log-loss0.04219291731715202\n",
      "Epoch:43 Batch:480 Log-loss0.029526211321353912\n",
      "Epoch:43 Batch:520 Log-loss0.03365549072623253\n",
      "Epoch:43 Batch:560 Log-loss0.023580245673656464\n",
      "Epoch average log loss:0.03642294460109302\n",
      "In Epoch43, val_loss:0.03847599553875076, best_val_loss:0.038327332179147876, best_auc:0.9907846981884747\n",
      "Epoch:44 Batch:0 Log-loss0.03837737441062927\n",
      "Epoch:44 Batch:40 Log-loss0.03502965345978737\n",
      "Epoch:44 Batch:80 Log-loss0.030342981219291687\n",
      "Epoch:44 Batch:120 Log-loss0.052422281354665756\n",
      "Epoch:44 Batch:160 Log-loss0.04670095443725586\n",
      "Epoch:44 Batch:200 Log-loss0.032355595380067825\n",
      "Epoch:44 Batch:240 Log-loss0.025137213990092278\n",
      "Epoch:44 Batch:280 Log-loss0.0377265140414238\n",
      "Epoch:44 Batch:320 Log-loss0.04410749673843384\n",
      "Epoch:44 Batch:360 Log-loss0.02441100962460041\n",
      "Epoch:44 Batch:400 Log-loss0.04394104704260826\n",
      "Epoch:44 Batch:440 Log-loss0.03415122628211975\n",
      "Epoch:44 Batch:480 Log-loss0.035502079874277115\n",
      "Epoch:44 Batch:520 Log-loss0.03698817268013954\n",
      "Epoch:44 Batch:560 Log-loss0.03453919664025307\n",
      "Epoch average log loss:0.03651261478329876\n",
      "In Epoch44, val_loss:0.03838693427506219, best_val_loss:0.038327332179147876, best_auc:0.9907846981884747\n",
      "Epoch:45 Batch:0 Log-loss0.04066948965191841\n",
      "Epoch:45 Batch:40 Log-loss0.032666321843862534\n",
      "Epoch:45 Batch:80 Log-loss0.040593598037958145\n",
      "Epoch:45 Batch:120 Log-loss0.027234675362706184\n",
      "Epoch:45 Batch:160 Log-loss0.05528222396969795\n",
      "Epoch:45 Batch:200 Log-loss0.03643824905157089\n",
      "Epoch:45 Batch:240 Log-loss0.026389434933662415\n",
      "Epoch:45 Batch:280 Log-loss0.04598049819469452\n",
      "Epoch:45 Batch:320 Log-loss0.04243434593081474\n",
      "Epoch:45 Batch:360 Log-loss0.024411702528595924\n",
      "Epoch:45 Batch:400 Log-loss0.049717117100954056\n",
      "Epoch:45 Batch:440 Log-loss0.028743097558617592\n",
      "Epoch:45 Batch:480 Log-loss0.024769539013504982\n",
      "Epoch:45 Batch:520 Log-loss0.04441727697849274\n",
      "Epoch:45 Batch:560 Log-loss0.03554776310920715\n",
      "Epoch average log loss:0.03654156979506037\n",
      "In Epoch45, val_loss:0.03840888433832565, best_val_loss:0.038327332179147876, best_auc:0.9907846981884747\n",
      "Epoch:46 Batch:0 Log-loss0.04822288826107979\n",
      "Epoch:46 Batch:40 Log-loss0.03701406344771385\n",
      "Epoch:46 Batch:80 Log-loss0.025873607024550438\n",
      "Epoch:46 Batch:120 Log-loss0.03291633352637291\n",
      "Epoch:46 Batch:160 Log-loss0.05966527760028839\n",
      "Epoch:46 Batch:200 Log-loss0.024091577157378197\n",
      "Epoch:46 Batch:240 Log-loss0.03283175453543663\n",
      "Epoch:46 Batch:280 Log-loss0.033899594098329544\n",
      "Epoch:46 Batch:320 Log-loss0.029717514291405678\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:46 Batch:360 Log-loss0.028098128736019135\n",
      "Epoch:46 Batch:400 Log-loss0.02508808858692646\n",
      "Epoch:46 Batch:440 Log-loss0.04062319174408913\n",
      "Epoch:46 Batch:480 Log-loss0.05267810449004173\n",
      "Epoch:46 Batch:520 Log-loss0.03798578307032585\n",
      "Epoch:46 Batch:560 Log-loss0.04819203540682793\n",
      "Epoch average log loss:0.03658836573595181\n",
      "In Epoch46, val_loss:0.03837748371079889, best_val_loss:0.038327332179147876, best_auc:0.9907846981884747\n",
      "Epoch:47 Batch:0 Log-loss0.04228948429226875\n",
      "Epoch:47 Batch:40 Log-loss0.053644049912691116\n",
      "Epoch:47 Batch:80 Log-loss0.0299371387809515\n",
      "Epoch:47 Batch:120 Log-loss0.03826596960425377\n",
      "Epoch:47 Batch:160 Log-loss0.04309915378689766\n",
      "Epoch:47 Batch:200 Log-loss0.03421245515346527\n",
      "Epoch:47 Batch:240 Log-loss0.05905935540795326\n",
      "Epoch:47 Batch:280 Log-loss0.04731204733252525\n",
      "Epoch:47 Batch:320 Log-loss0.04027416184544563\n",
      "Epoch:47 Batch:360 Log-loss0.0433635376393795\n",
      "Epoch:47 Batch:400 Log-loss0.048432428389787674\n",
      "Epoch:47 Batch:440 Log-loss0.032580431550741196\n",
      "Epoch:47 Batch:480 Log-loss0.040981441736221313\n",
      "Epoch:47 Batch:520 Log-loss0.025363504886627197\n",
      "Epoch:47 Batch:560 Log-loss0.04108262062072754\n",
      "Epoch average log loss:0.036533123509226634\n",
      "A pre-trained model at temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt has been loaded.\n",
      "\n",
      "Model has been saved as checkpoints/FASTTXT_BGRU_64_641.pt.\n",
      "\n",
      "Choose the torch base model.\n",
      "Training on fold 2\n",
      "Epoch:1 Batch:0 Log-loss0.6928703188896179\n",
      "Epoch:1 Batch:40 Log-loss0.13741891086101532\n",
      "Epoch:1 Batch:80 Log-loss0.08087212592363358\n",
      "Epoch:1 Batch:120 Log-loss0.06089150905609131\n",
      "Epoch:1 Batch:160 Log-loss0.06759323924779892\n",
      "Epoch:1 Batch:200 Log-loss0.060083236545324326\n",
      "Epoch:1 Batch:240 Log-loss0.05709236487746239\n",
      "Epoch:1 Batch:280 Log-loss0.0585591085255146\n",
      "Epoch:1 Batch:320 Log-loss0.048386529088020325\n",
      "Epoch:1 Batch:360 Log-loss0.05101696029305458\n",
      "Epoch:1 Batch:400 Log-loss0.0377175509929657\n",
      "Epoch:1 Batch:440 Log-loss0.06078179180622101\n",
      "Epoch:1 Batch:480 Log-loss0.061230335384607315\n",
      "Epoch:1 Batch:520 Log-loss0.05453072115778923\n",
      "Epoch:1 Batch:560 Log-loss0.05769366770982742\n",
      "Epoch average log loss:0.07792876437306404\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch1, val_loss:0.046625844337742174, best_val_loss:0.046625844337742174, best_auc:0.9779375516895453\n",
      "Epoch:2 Batch:0 Log-loss0.06141050532460213\n",
      "Epoch:2 Batch:40 Log-loss0.06197091564536095\n",
      "Epoch:2 Batch:80 Log-loss0.05124599114060402\n",
      "Epoch:2 Batch:120 Log-loss0.04217907786369324\n",
      "Epoch:2 Batch:160 Log-loss0.041186507791280746\n",
      "Epoch:2 Batch:200 Log-loss0.040652502328157425\n",
      "Epoch:2 Batch:240 Log-loss0.04890696704387665\n",
      "Epoch:2 Batch:280 Log-loss0.04221348464488983\n",
      "Epoch:2 Batch:320 Log-loss0.07009810209274292\n",
      "Epoch:2 Batch:360 Log-loss0.03717415779829025\n",
      "Epoch:2 Batch:400 Log-loss0.04542696475982666\n",
      "Epoch:2 Batch:440 Log-loss0.05036729574203491\n",
      "Epoch:2 Batch:480 Log-loss0.0456293523311615\n",
      "Epoch:2 Batch:520 Log-loss0.057056959718465805\n",
      "Epoch:2 Batch:560 Log-loss0.05003466084599495\n",
      "Epoch average log loss:0.04962383274521146\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch2, val_loss:0.04514235156313163, best_val_loss:0.04514235156313163, best_auc:0.9822080881748119\n",
      "Epoch:3 Batch:0 Log-loss0.03335085138678551\n",
      "Epoch:3 Batch:40 Log-loss0.06011822819709778\n",
      "Epoch:3 Batch:80 Log-loss0.028101513162255287\n",
      "Epoch:3 Batch:120 Log-loss0.05916588380932808\n",
      "Epoch:3 Batch:160 Log-loss0.03896266967058182\n",
      "Epoch:3 Batch:200 Log-loss0.051912251859903336\n",
      "Epoch:3 Batch:240 Log-loss0.038363803178071976\n",
      "Epoch:3 Batch:280 Log-loss0.06774355471134186\n",
      "Epoch:3 Batch:320 Log-loss0.04007725045084953\n",
      "Epoch:3 Batch:360 Log-loss0.05438389256596565\n",
      "Epoch:3 Batch:400 Log-loss0.04791267588734627\n",
      "Epoch:3 Batch:440 Log-loss0.056407202035188675\n",
      "Epoch:3 Batch:480 Log-loss0.05312284454703331\n",
      "Epoch:3 Batch:520 Log-loss0.06000146642327309\n",
      "Epoch:3 Batch:560 Log-loss0.05476024001836777\n",
      "Epoch average log loss:0.04748112331144512\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch3, val_loss:0.04327715993823401, best_val_loss:0.04327715993823401, best_auc:0.9848200546351588\n",
      "Epoch:4 Batch:0 Log-loss0.04429585859179497\n",
      "Epoch:4 Batch:40 Log-loss0.05746098980307579\n",
      "Epoch:4 Batch:80 Log-loss0.043072592467069626\n",
      "Epoch:4 Batch:120 Log-loss0.0523291677236557\n",
      "Epoch:4 Batch:160 Log-loss0.066464364528656\n",
      "Epoch:4 Batch:200 Log-loss0.057495031505823135\n",
      "Epoch:4 Batch:240 Log-loss0.045616596937179565\n",
      "Epoch:4 Batch:280 Log-loss0.05921653285622597\n",
      "Epoch:4 Batch:320 Log-loss0.044818129390478134\n",
      "Epoch:4 Batch:360 Log-loss0.0335150845348835\n",
      "Epoch:4 Batch:400 Log-loss0.04476849362254143\n",
      "Epoch:4 Batch:440 Log-loss0.04324735328555107\n",
      "Epoch:4 Batch:480 Log-loss0.04573209583759308\n",
      "Epoch:4 Batch:520 Log-loss0.04689997807145119\n",
      "Epoch:4 Batch:560 Log-loss0.05522151291370392\n",
      "Epoch average log loss:0.04563971364405006\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch4, val_loss:0.04265378853598, best_val_loss:0.04265378853598, best_auc:0.9869044600476345\n",
      "Epoch:5 Batch:0 Log-loss0.04909389093518257\n",
      "Epoch:5 Batch:40 Log-loss0.04433010518550873\n",
      "Epoch:5 Batch:80 Log-loss0.06495063751935959\n",
      "Epoch:5 Batch:120 Log-loss0.030279912054538727\n",
      "Epoch:5 Batch:160 Log-loss0.04490276798605919\n",
      "Epoch:5 Batch:200 Log-loss0.029577812179923058\n",
      "Epoch:5 Batch:240 Log-loss0.03991992771625519\n",
      "Epoch:5 Batch:280 Log-loss0.04596048966050148\n",
      "Epoch:5 Batch:320 Log-loss0.042429447174072266\n",
      "Epoch:5 Batch:360 Log-loss0.01913551054894924\n",
      "Epoch:5 Batch:400 Log-loss0.030120356008410454\n",
      "Epoch:5 Batch:440 Log-loss0.044404949992895126\n",
      "Epoch:5 Batch:480 Log-loss0.031529877334833145\n",
      "Epoch:5 Batch:520 Log-loss0.037567660212516785\n",
      "Epoch:5 Batch:560 Log-loss0.04079128056764603\n",
      "Epoch average log loss:0.04457886347414127\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch5, val_loss:0.04158672323961483, best_val_loss:0.04158672323961483, best_auc:0.9879017506522526\n",
      "Epoch:6 Batch:0 Log-loss0.03328610584139824\n",
      "Epoch:6 Batch:40 Log-loss0.04574345052242279\n",
      "Epoch:6 Batch:80 Log-loss0.03834918141365051\n",
      "Epoch:6 Batch:120 Log-loss0.04635148122906685\n",
      "Epoch:6 Batch:160 Log-loss0.044530078768730164\n",
      "Epoch:6 Batch:200 Log-loss0.04357682541012764\n",
      "Epoch:6 Batch:240 Log-loss0.034290872514247894\n",
      "Epoch:6 Batch:280 Log-loss0.041356440633535385\n",
      "Epoch:6 Batch:320 Log-loss0.0407344214618206\n",
      "Epoch:6 Batch:360 Log-loss0.03984351083636284\n",
      "Epoch:6 Batch:400 Log-loss0.04697701334953308\n",
      "Epoch:6 Batch:440 Log-loss0.0483773797750473\n",
      "Epoch:6 Batch:480 Log-loss0.04298090934753418\n",
      "Epoch:6 Batch:520 Log-loss0.05624474957585335\n",
      "Epoch:6 Batch:560 Log-loss0.03654629737138748\n",
      "Epoch average log loss:0.04318194537357028\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch6, val_loss:0.04091054883855156, best_val_loss:0.04091054883855156, best_auc:0.9888278430001544\n",
      "Epoch:7 Batch:0 Log-loss0.025823427364230156\n",
      "Epoch:7 Batch:40 Log-loss0.054370831698179245\n",
      "Epoch:7 Batch:80 Log-loss0.05008721724152565\n",
      "Epoch:7 Batch:120 Log-loss0.03641465678811073\n",
      "Epoch:7 Batch:160 Log-loss0.04898731783032417\n",
      "Epoch:7 Batch:200 Log-loss0.030017472803592682\n",
      "Epoch:7 Batch:240 Log-loss0.044669438153505325\n",
      "Epoch:7 Batch:280 Log-loss0.04323868826031685\n",
      "Epoch:7 Batch:320 Log-loss0.030304796993732452\n",
      "Epoch:7 Batch:360 Log-loss0.05659133568406105\n",
      "Epoch:7 Batch:400 Log-loss0.048093076795339584\n",
      "Epoch:7 Batch:440 Log-loss0.04988396540284157\n",
      "Epoch:7 Batch:480 Log-loss0.0414276085793972\n",
      "Epoch:7 Batch:520 Log-loss0.03262503445148468\n",
      "Epoch:7 Batch:560 Log-loss0.05871284753084183\n",
      "Epoch average log loss:0.04256915923740182\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch7, val_loss:0.03961960659440927, best_val_loss:0.03961960659440927, best_auc:0.9892107998170271\n",
      "Epoch:8 Batch:0 Log-loss0.035870280116796494\n",
      "Epoch:8 Batch:40 Log-loss0.05470597371459007\n",
      "Epoch:8 Batch:80 Log-loss0.06276452541351318\n",
      "Epoch:8 Batch:120 Log-loss0.05183902010321617\n",
      "Epoch:8 Batch:160 Log-loss0.04995383322238922\n",
      "Epoch:8 Batch:200 Log-loss0.035522688180208206\n",
      "Epoch:8 Batch:240 Log-loss0.046604037284851074\n",
      "Epoch:8 Batch:280 Log-loss0.04952894151210785\n",
      "Epoch:8 Batch:320 Log-loss0.051489707082509995\n",
      "Epoch:8 Batch:360 Log-loss0.039719872176647186\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:8 Batch:400 Log-loss0.03662538155913353\n",
      "Epoch:8 Batch:440 Log-loss0.03899862989783287\n",
      "Epoch:8 Batch:480 Log-loss0.04006917029619217\n",
      "Epoch:8 Batch:520 Log-loss0.0517415814101696\n",
      "Epoch:8 Batch:560 Log-loss0.035872798413038254\n",
      "Epoch average log loss:0.04204534596231367\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch8, val_loss:0.03950107477793544, best_val_loss:0.03950107477793544, best_auc:0.989833465663108\n",
      "Epoch:9 Batch:0 Log-loss0.0630677193403244\n",
      "Epoch:9 Batch:40 Log-loss0.032331325113773346\n",
      "Epoch:9 Batch:80 Log-loss0.047606054693460464\n",
      "Epoch:9 Batch:120 Log-loss0.054586682468652725\n",
      "Epoch:9 Batch:160 Log-loss0.047248248010873795\n",
      "Epoch:9 Batch:200 Log-loss0.02666633389890194\n",
      "Epoch:9 Batch:240 Log-loss0.04046201333403587\n",
      "Epoch:9 Batch:280 Log-loss0.032455701380968094\n",
      "Epoch:9 Batch:320 Log-loss0.03451235219836235\n",
      "Epoch:9 Batch:360 Log-loss0.03911256045103073\n",
      "Epoch:9 Batch:400 Log-loss0.030111514031887054\n",
      "Epoch:9 Batch:440 Log-loss0.05952444672584534\n",
      "Epoch:9 Batch:480 Log-loss0.03191159665584564\n",
      "Epoch:9 Batch:520 Log-loss0.0485529750585556\n",
      "Epoch:9 Batch:560 Log-loss0.06810222566127777\n",
      "Epoch average log loss:0.04151580774757479\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch9, val_loss:0.039299820091780506, best_val_loss:0.039299820091780506, best_auc:0.9899954952891495\n",
      "Epoch:10 Batch:0 Log-loss0.024337440729141235\n",
      "Epoch:10 Batch:40 Log-loss0.054267723113298416\n",
      "Epoch:10 Batch:80 Log-loss0.024722551926970482\n",
      "Epoch:10 Batch:120 Log-loss0.03751711547374725\n",
      "Epoch:10 Batch:160 Log-loss0.03169744089245796\n",
      "Epoch:10 Batch:200 Log-loss0.0469374917447567\n",
      "Epoch:10 Batch:240 Log-loss0.036286611109972\n",
      "Epoch:10 Batch:280 Log-loss0.05758748948574066\n",
      "Epoch:10 Batch:320 Log-loss0.044614460319280624\n",
      "Epoch:10 Batch:360 Log-loss0.02717798389494419\n",
      "Epoch:10 Batch:400 Log-loss0.04301437363028526\n",
      "Epoch:10 Batch:440 Log-loss0.04589441418647766\n",
      "Epoch:10 Batch:480 Log-loss0.052281588315963745\n",
      "Epoch:10 Batch:520 Log-loss0.03986307233572006\n",
      "Epoch:10 Batch:560 Log-loss0.04001610726118088\n",
      "Epoch average log loss:0.040849499469290354\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch10, val_loss:0.038782851197542156, best_val_loss:0.038782851197542156, best_auc:0.9901860012797642\n",
      "Epoch:11 Batch:0 Log-loss0.04028342291712761\n",
      "Epoch:11 Batch:40 Log-loss0.042978957295417786\n",
      "Epoch:11 Batch:80 Log-loss0.041277918964624405\n",
      "Epoch:11 Batch:120 Log-loss0.03345264121890068\n",
      "Epoch:11 Batch:160 Log-loss0.03200631961226463\n",
      "Epoch:11 Batch:200 Log-loss0.030004212632775307\n",
      "Epoch:11 Batch:240 Log-loss0.03735075518488884\n",
      "Epoch:11 Batch:280 Log-loss0.039856430143117905\n",
      "Epoch:11 Batch:320 Log-loss0.04281102120876312\n",
      "Epoch:11 Batch:360 Log-loss0.0493350476026535\n",
      "Epoch:11 Batch:400 Log-loss0.035422321408987045\n",
      "Epoch:11 Batch:440 Log-loss0.03792629763484001\n",
      "Epoch:11 Batch:480 Log-loss0.032001446932554245\n",
      "Epoch:11 Batch:520 Log-loss0.0467241071164608\n",
      "Epoch:11 Batch:560 Log-loss0.03683655709028244\n",
      "Epoch average log loss:0.04049182879805033\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch11, val_loss:0.038310279637248, best_val_loss:0.038310279637248, best_auc:0.9903923975064167\n",
      "Epoch:12 Batch:0 Log-loss0.04148334264755249\n",
      "Epoch:12 Batch:40 Log-loss0.04804639518260956\n",
      "Epoch:12 Batch:80 Log-loss0.030943913385272026\n",
      "Epoch:12 Batch:120 Log-loss0.03439882770180702\n",
      "Epoch:12 Batch:160 Log-loss0.03299998864531517\n",
      "Epoch:12 Batch:200 Log-loss0.02846428006887436\n",
      "Epoch:12 Batch:240 Log-loss0.04001440852880478\n",
      "Epoch:12 Batch:280 Log-loss0.037067804485559464\n",
      "Epoch:12 Batch:320 Log-loss0.03687363862991333\n",
      "Epoch:12 Batch:360 Log-loss0.036079246550798416\n",
      "Epoch:12 Batch:400 Log-loss0.04696764424443245\n",
      "Epoch:12 Batch:440 Log-loss0.029513360932469368\n",
      "Epoch:12 Batch:480 Log-loss0.05089746043086052\n",
      "Epoch:12 Batch:520 Log-loss0.03382499888539314\n",
      "Epoch:12 Batch:560 Log-loss0.04745011776685715\n",
      "Epoch average log loss:0.039856560611432154\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch12, val_loss:0.03795102165399885, best_val_loss:0.03795102165399885, best_auc:0.9904461102925978\n",
      "Epoch:13 Batch:0 Log-loss0.03586495295166969\n",
      "Epoch:13 Batch:40 Log-loss0.029652126133441925\n",
      "Epoch:13 Batch:80 Log-loss0.027871474623680115\n",
      "Epoch:13 Batch:120 Log-loss0.0412600077688694\n",
      "Epoch:13 Batch:160 Log-loss0.02791735529899597\n",
      "Epoch:13 Batch:200 Log-loss0.03972065821290016\n",
      "Epoch:13 Batch:240 Log-loss0.03724301978945732\n",
      "Epoch:13 Batch:280 Log-loss0.02923210710287094\n",
      "Epoch:13 Batch:320 Log-loss0.03338851034641266\n",
      "Epoch:13 Batch:360 Log-loss0.041641294956207275\n",
      "Epoch:13 Batch:400 Log-loss0.036343980580568314\n",
      "Epoch:13 Batch:440 Log-loss0.039476413279771805\n",
      "Epoch:13 Batch:480 Log-loss0.02864190936088562\n",
      "Epoch:13 Batch:520 Log-loss0.0560058169066906\n",
      "Epoch:13 Batch:560 Log-loss0.0406799279153347\n",
      "Epoch average log loss:0.03982397252361157\n",
      "In Epoch13, val_loss:0.03834595008305127, best_val_loss:0.03795102165399885, best_auc:0.9904461102925978\n",
      "Epoch:14 Batch:0 Log-loss0.02360472083091736\n",
      "Epoch:14 Batch:40 Log-loss0.043772995471954346\n",
      "Epoch:14 Batch:80 Log-loss0.044469982385635376\n",
      "Epoch:14 Batch:120 Log-loss0.04090387374162674\n",
      "Epoch:14 Batch:160 Log-loss0.02998700924217701\n",
      "Epoch:14 Batch:200 Log-loss0.029073841869831085\n",
      "Epoch:14 Batch:240 Log-loss0.03474513813853264\n",
      "Epoch:14 Batch:280 Log-loss0.04185404255986214\n",
      "Epoch:14 Batch:320 Log-loss0.030516639351844788\n",
      "Epoch:14 Batch:360 Log-loss0.031710509210824966\n",
      "Epoch:14 Batch:400 Log-loss0.050654277205467224\n",
      "Epoch:14 Batch:440 Log-loss0.043831780552864075\n",
      "Epoch:14 Batch:480 Log-loss0.03936177119612694\n",
      "Epoch:14 Batch:520 Log-loss0.02852933667600155\n",
      "Epoch:14 Batch:560 Log-loss0.03848551586270332\n",
      "Epoch average log loss:0.039428533068192856\n",
      "In Epoch14, val_loss:0.03835661695449988, best_val_loss:0.03795102165399885, best_auc:0.9904461102925978\n",
      "Epoch:15 Batch:0 Log-loss0.028331583365797997\n",
      "Epoch:15 Batch:40 Log-loss0.03219308704137802\n",
      "Epoch:15 Batch:80 Log-loss0.034813184291124344\n",
      "Epoch:15 Batch:120 Log-loss0.028493473306298256\n",
      "Epoch:15 Batch:160 Log-loss0.030672525987029076\n",
      "Epoch:15 Batch:200 Log-loss0.05371078476309776\n",
      "Epoch:15 Batch:240 Log-loss0.04828621447086334\n",
      "Epoch:15 Batch:280 Log-loss0.043323416262865067\n",
      "Epoch:15 Batch:320 Log-loss0.032927822321653366\n",
      "Epoch:15 Batch:360 Log-loss0.05479094013571739\n",
      "Epoch:15 Batch:400 Log-loss0.03814859315752983\n",
      "Epoch:15 Batch:440 Log-loss0.03734656050801277\n",
      "Epoch:15 Batch:480 Log-loss0.045749273151159286\n",
      "Epoch:15 Batch:520 Log-loss0.046068090945482254\n",
      "Epoch:15 Batch:560 Log-loss0.03779882937669754\n",
      "Epoch average log loss:0.0392482830710443\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch15, val_loss:0.0379426083377579, best_val_loss:0.0379426083377579, best_auc:0.9905999223071892\n",
      "Epoch:16 Batch:0 Log-loss0.0465434305369854\n",
      "Epoch:16 Batch:40 Log-loss0.03553055599331856\n",
      "Epoch:16 Batch:80 Log-loss0.03777426481246948\n",
      "Epoch:16 Batch:120 Log-loss0.04047202691435814\n",
      "Epoch:16 Batch:160 Log-loss0.05390375852584839\n",
      "Epoch:16 Batch:200 Log-loss0.03600732982158661\n",
      "Epoch:16 Batch:240 Log-loss0.03528046980500221\n",
      "Epoch:16 Batch:280 Log-loss0.04521110653877258\n",
      "Epoch:16 Batch:320 Log-loss0.02067331038415432\n",
      "Epoch:16 Batch:360 Log-loss0.03184175118803978\n",
      "Epoch:16 Batch:400 Log-loss0.03834965080022812\n",
      "Epoch:16 Batch:440 Log-loss0.046109866350889206\n",
      "Epoch:16 Batch:480 Log-loss0.034112598747015\n",
      "Epoch:16 Batch:520 Log-loss0.03983728960156441\n",
      "Epoch:16 Batch:560 Log-loss0.04578253626823425\n",
      "Epoch average log loss:0.03895667488354125\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch16, val_loss:0.0377604549079151, best_val_loss:0.0377604549079151, best_auc:0.990652319559914\n",
      "Epoch:17 Batch:0 Log-loss0.029431330040097237\n",
      "Epoch:17 Batch:40 Log-loss0.05665382370352745\n",
      "Epoch:17 Batch:80 Log-loss0.04645794257521629\n",
      "Epoch:17 Batch:120 Log-loss0.0407097227871418\n",
      "Epoch:17 Batch:160 Log-loss0.03570893034338951\n",
      "Epoch:17 Batch:200 Log-loss0.0460493303835392\n",
      "Epoch:17 Batch:240 Log-loss0.03755299746990204\n",
      "Epoch:17 Batch:280 Log-loss0.03398856893181801\n",
      "Epoch:17 Batch:320 Log-loss0.03770415112376213\n",
      "Epoch:17 Batch:360 Log-loss0.038698624819517136\n",
      "Epoch:17 Batch:400 Log-loss0.03448640555143356\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:17 Batch:440 Log-loss0.050244856625795364\n",
      "Epoch:17 Batch:480 Log-loss0.03357003629207611\n",
      "Epoch:17 Batch:520 Log-loss0.045512083917856216\n",
      "Epoch:17 Batch:560 Log-loss0.037274282425642014\n",
      "Epoch average log loss:0.038705564383417365\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch17, val_loss:0.03753429972140312, best_val_loss:0.03753429972140312, best_auc:0.9907392010568472\n",
      "Epoch:18 Batch:0 Log-loss0.03486906737089157\n",
      "Epoch:18 Batch:40 Log-loss0.040532927960157394\n",
      "Epoch:18 Batch:80 Log-loss0.037907350808382034\n",
      "Epoch:18 Batch:120 Log-loss0.05321185290813446\n",
      "Epoch:18 Batch:160 Log-loss0.04002360999584198\n",
      "Epoch:18 Batch:200 Log-loss0.048030223697423935\n",
      "Epoch:18 Batch:240 Log-loss0.03564560040831566\n",
      "Epoch:18 Batch:280 Log-loss0.05232837423682213\n",
      "Epoch:18 Batch:320 Log-loss0.033754441887140274\n",
      "Epoch:18 Batch:360 Log-loss0.024255527183413506\n",
      "Epoch:18 Batch:400 Log-loss0.03634358569979668\n",
      "Epoch:18 Batch:440 Log-loss0.03060762770473957\n",
      "Epoch:18 Batch:480 Log-loss0.030194764956831932\n",
      "Epoch:18 Batch:520 Log-loss0.04299859330058098\n",
      "Epoch:18 Batch:560 Log-loss0.0384242907166481\n",
      "Epoch average log loss:0.0386364275057401\n",
      "In Epoch18, val_loss:0.03790959434035578, best_val_loss:0.03753429972140312, best_auc:0.9907392010568472\n",
      "Epoch:19 Batch:0 Log-loss0.03413819894194603\n",
      "Epoch:19 Batch:40 Log-loss0.0303022563457489\n",
      "Epoch:19 Batch:80 Log-loss0.039421334862709045\n",
      "Epoch:19 Batch:120 Log-loss0.051185932010412216\n",
      "Epoch:19 Batch:160 Log-loss0.04105416312813759\n",
      "Epoch:19 Batch:200 Log-loss0.030440853908658028\n",
      "Epoch:19 Batch:240 Log-loss0.03916696086525917\n",
      "Epoch:19 Batch:280 Log-loss0.040150176733732224\n",
      "Epoch:19 Batch:320 Log-loss0.03423193469643593\n",
      "Epoch:19 Batch:360 Log-loss0.060524359345436096\n",
      "Epoch:19 Batch:400 Log-loss0.038640622049570084\n",
      "Epoch:19 Batch:440 Log-loss0.02986324578523636\n",
      "Epoch:19 Batch:480 Log-loss0.04476431384682655\n",
      "Epoch:19 Batch:520 Log-loss0.03342575952410698\n",
      "Epoch:19 Batch:560 Log-loss0.024102911353111267\n",
      "Epoch average log loss:0.03840362134256533\n",
      "In Epoch19, val_loss:0.03765058803836623, best_val_loss:0.03753429972140312, best_auc:0.9907392010568472\n",
      "Epoch:20 Batch:0 Log-loss0.025528589263558388\n",
      "Epoch:20 Batch:40 Log-loss0.036758244037628174\n",
      "Epoch:20 Batch:80 Log-loss0.027567483484745026\n",
      "Epoch:20 Batch:120 Log-loss0.03922842815518379\n",
      "Epoch:20 Batch:160 Log-loss0.02417660504579544\n",
      "Epoch:20 Batch:200 Log-loss0.029249390587210655\n",
      "Epoch:20 Batch:240 Log-loss0.035298846662044525\n",
      "Epoch:20 Batch:280 Log-loss0.041169438511133194\n",
      "Epoch:20 Batch:320 Log-loss0.04059450328350067\n",
      "Epoch:20 Batch:360 Log-loss0.02514701336622238\n",
      "Epoch:20 Batch:400 Log-loss0.05654831603169441\n",
      "Epoch:20 Batch:440 Log-loss0.03687482699751854\n",
      "Epoch:20 Batch:480 Log-loss0.03417056053876877\n",
      "Epoch:20 Batch:520 Log-loss0.041902199387550354\n",
      "Epoch:20 Batch:560 Log-loss0.037453439086675644\n",
      "Epoch average log loss:0.03836283105026398\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch20, val_loss:0.037408166070452796, best_val_loss:0.037408166070452796, best_auc:0.990843676832069\n",
      "Epoch:21 Batch:0 Log-loss0.04353359341621399\n",
      "Epoch:21 Batch:40 Log-loss0.03881264850497246\n",
      "Epoch:21 Batch:80 Log-loss0.046610042452812195\n",
      "Epoch:21 Batch:120 Log-loss0.039024535566568375\n",
      "Epoch:21 Batch:160 Log-loss0.028751440346240997\n",
      "Epoch:21 Batch:200 Log-loss0.04233742132782936\n",
      "Epoch:21 Batch:240 Log-loss0.032723501324653625\n",
      "Epoch:21 Batch:280 Log-loss0.020987382158637047\n",
      "Epoch:21 Batch:320 Log-loss0.029667483642697334\n",
      "Epoch:21 Batch:360 Log-loss0.04363049194216728\n",
      "Epoch:21 Batch:400 Log-loss0.0355210043489933\n",
      "Epoch:21 Batch:440 Log-loss0.04169869050383568\n",
      "Epoch:21 Batch:480 Log-loss0.03355199471116066\n",
      "Epoch:21 Batch:520 Log-loss0.05163900926709175\n",
      "Epoch:21 Batch:560 Log-loss0.03254113346338272\n",
      "Epoch average log loss:0.038161646524843364\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch21, val_loss:0.03733554633664588, best_val_loss:0.03733554633664588, best_auc:0.990945842787065\n",
      "Epoch:22 Batch:0 Log-loss0.03271177411079407\n",
      "Epoch:22 Batch:40 Log-loss0.025723015889525414\n",
      "Epoch:22 Batch:80 Log-loss0.04055812582373619\n",
      "Epoch:22 Batch:120 Log-loss0.05725717917084694\n",
      "Epoch:22 Batch:160 Log-loss0.04136880487203598\n",
      "Epoch:22 Batch:200 Log-loss0.042168501764535904\n",
      "Epoch:22 Batch:240 Log-loss0.04615047946572304\n",
      "Epoch:22 Batch:280 Log-loss0.03872884809970856\n",
      "Epoch:22 Batch:320 Log-loss0.03559502959251404\n",
      "Epoch:22 Batch:360 Log-loss0.047807175666093826\n",
      "Epoch:22 Batch:400 Log-loss0.021501071751117706\n",
      "Epoch:22 Batch:440 Log-loss0.02286287397146225\n",
      "Epoch:22 Batch:480 Log-loss0.05280636250972748\n",
      "Epoch:22 Batch:520 Log-loss0.03324781358242035\n",
      "Epoch:22 Batch:560 Log-loss0.02994603104889393\n",
      "Epoch average log loss:0.03807821830601565\n",
      "In Epoch22, val_loss:0.037636434181642385, best_val_loss:0.03733554633664588, best_auc:0.990945842787065\n",
      "Epoch:23 Batch:0 Log-loss0.02577272802591324\n",
      "Epoch:23 Batch:40 Log-loss0.034274905920028687\n",
      "Epoch:23 Batch:80 Log-loss0.04005409777164459\n",
      "Epoch:23 Batch:120 Log-loss0.04096686840057373\n",
      "Epoch:23 Batch:160 Log-loss0.03913796320557594\n",
      "Epoch:23 Batch:200 Log-loss0.03929689899086952\n",
      "Epoch:23 Batch:240 Log-loss0.025365419685840607\n",
      "Epoch:23 Batch:280 Log-loss0.04799972474575043\n",
      "Epoch:23 Batch:320 Log-loss0.026053866371512413\n",
      "Epoch:23 Batch:360 Log-loss0.04018155485391617\n",
      "Epoch:23 Batch:400 Log-loss0.02252846397459507\n",
      "Epoch:23 Batch:440 Log-loss0.0352703258395195\n",
      "Epoch:23 Batch:480 Log-loss0.039031755179166794\n",
      "Epoch:23 Batch:520 Log-loss0.03579721599817276\n",
      "Epoch:23 Batch:560 Log-loss0.02142004668712616\n",
      "Epoch average log loss:0.037686815086219995\n",
      "In Epoch23, val_loss:0.03748472242344011, best_val_loss:0.03733554633664588, best_auc:0.990945842787065\n",
      "Epoch:24 Batch:0 Log-loss0.025095144286751747\n",
      "Epoch:24 Batch:40 Log-loss0.04032072052359581\n",
      "Epoch:24 Batch:80 Log-loss0.027819799259305\n",
      "Epoch:24 Batch:120 Log-loss0.040683142840862274\n",
      "Epoch:24 Batch:160 Log-loss0.04492028057575226\n",
      "Epoch:24 Batch:200 Log-loss0.04306287690997124\n",
      "Epoch:24 Batch:240 Log-loss0.037106119096279144\n",
      "Epoch:24 Batch:280 Log-loss0.027376897633075714\n",
      "Epoch:24 Batch:320 Log-loss0.031175727024674416\n",
      "Epoch:24 Batch:360 Log-loss0.03036065213382244\n",
      "Epoch:24 Batch:400 Log-loss0.04644479230046272\n",
      "Epoch:24 Batch:440 Log-loss0.042844340205192566\n",
      "Epoch:24 Batch:480 Log-loss0.03550352901220322\n",
      "Epoch:24 Batch:520 Log-loss0.019797593355178833\n",
      "Epoch:24 Batch:560 Log-loss0.036437634378671646\n",
      "Epoch average log loss:0.03764686605427414\n",
      "In Epoch24, val_loss:0.03740309855701005, best_val_loss:0.03733554633664588, best_auc:0.990945842787065\n",
      "Epoch:25 Batch:0 Log-loss0.025205975398421288\n",
      "Epoch:25 Batch:40 Log-loss0.041505638509988785\n",
      "Epoch:25 Batch:80 Log-loss0.02499012090265751\n",
      "Epoch:25 Batch:120 Log-loss0.03731862083077431\n",
      "Epoch:25 Batch:160 Log-loss0.037969157099723816\n",
      "Epoch:25 Batch:200 Log-loss0.03886739909648895\n",
      "Epoch:25 Batch:240 Log-loss0.044254180043935776\n",
      "Epoch:25 Batch:280 Log-loss0.04026693478226662\n",
      "Epoch:25 Batch:320 Log-loss0.03317466005682945\n",
      "Epoch:25 Batch:360 Log-loss0.028581460937857628\n",
      "Epoch:25 Batch:400 Log-loss0.034844402223825455\n",
      "Epoch:25 Batch:440 Log-loss0.027417434379458427\n",
      "Epoch:25 Batch:480 Log-loss0.039165183901786804\n",
      "Epoch:25 Batch:520 Log-loss0.03599189594388008\n",
      "Epoch:25 Batch:560 Log-loss0.030318595468997955\n",
      "Epoch average log loss:0.037638174885484786\n",
      "In Epoch25, val_loss:0.03745724523994442, best_val_loss:0.03733554633664588, best_auc:0.990945842787065\n",
      "Epoch:26 Batch:0 Log-loss0.036973897367715836\n",
      "Epoch:26 Batch:40 Log-loss0.055210769176483154\n",
      "Epoch:26 Batch:80 Log-loss0.035854365676641464\n",
      "Epoch:26 Batch:120 Log-loss0.05026581883430481\n",
      "Epoch:26 Batch:160 Log-loss0.04413658380508423\n",
      "Epoch:26 Batch:200 Log-loss0.027872847393155098\n",
      "Epoch:26 Batch:240 Log-loss0.0304498802870512\n",
      "Epoch:26 Batch:280 Log-loss0.031428758054971695\n",
      "Epoch:26 Batch:320 Log-loss0.026178887113928795\n",
      "Epoch:26 Batch:360 Log-loss0.029984548687934875\n",
      "Epoch:26 Batch:400 Log-loss0.023615824058651924\n",
      "Epoch:26 Batch:440 Log-loss0.05363977327942848\n",
      "Epoch:26 Batch:480 Log-loss0.03534262254834175\n",
      "Epoch:26 Batch:520 Log-loss0.03606637567281723\n",
      "Epoch:26 Batch:560 Log-loss0.028045549988746643\n",
      "Epoch average log loss:0.03744764811292823\n",
      "In Epoch26, val_loss:0.03744152673484704, best_val_loss:0.03733554633664588, best_auc:0.990945842787065\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:27 Batch:0 Log-loss0.04765066131949425\n",
      "Epoch:27 Batch:40 Log-loss0.030533581972122192\n",
      "Epoch:27 Batch:80 Log-loss0.03516639769077301\n",
      "Epoch:27 Batch:120 Log-loss0.04590623080730438\n",
      "Epoch:27 Batch:160 Log-loss0.028501277789473534\n",
      "Epoch:27 Batch:200 Log-loss0.02199750579893589\n",
      "Epoch:27 Batch:240 Log-loss0.03329619765281677\n",
      "Epoch:27 Batch:280 Log-loss0.02977646328508854\n",
      "Epoch:27 Batch:320 Log-loss0.046193405985832214\n",
      "Epoch:27 Batch:360 Log-loss0.0385616160929203\n",
      "Epoch:27 Batch:400 Log-loss0.04397210851311684\n",
      "Epoch:27 Batch:440 Log-loss0.03921296074986458\n",
      "Epoch:27 Batch:480 Log-loss0.034743670374155045\n",
      "Epoch:27 Batch:520 Log-loss0.03560959920287132\n",
      "Epoch:27 Batch:560 Log-loss0.03977116942405701\n",
      "Epoch average log loss:0.03747429199782865\n",
      "In Epoch27, val_loss:0.03735585581873504, best_val_loss:0.03733554633664588, best_auc:0.990945842787065\n",
      "Epoch:28 Batch:0 Log-loss0.041897520422935486\n",
      "Epoch:28 Batch:40 Log-loss0.029437685385346413\n",
      "Epoch:28 Batch:80 Log-loss0.04550371691584587\n",
      "Epoch:28 Batch:120 Log-loss0.031984709203243256\n",
      "Epoch:28 Batch:160 Log-loss0.04224340245127678\n",
      "Epoch:28 Batch:200 Log-loss0.03830043599009514\n",
      "Epoch:28 Batch:240 Log-loss0.03222541883587837\n",
      "Epoch:28 Batch:280 Log-loss0.029182644560933113\n",
      "Epoch:28 Batch:320 Log-loss0.01971314288675785\n",
      "Epoch:28 Batch:360 Log-loss0.0470844991505146\n",
      "Epoch:28 Batch:400 Log-loss0.03811819478869438\n",
      "Epoch:28 Batch:440 Log-loss0.028816692531108856\n",
      "Epoch:28 Batch:480 Log-loss0.03254137933254242\n",
      "Epoch:28 Batch:520 Log-loss0.049903661012649536\n",
      "Epoch:28 Batch:560 Log-loss0.03229675069451332\n",
      "Epoch average log loss:0.037307998847349415\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch28, val_loss:0.037304429734430704, best_val_loss:0.037304429734430704, best_auc:0.9910044825886049\n",
      "Epoch:29 Batch:0 Log-loss0.0334039069712162\n",
      "Epoch:29 Batch:40 Log-loss0.03596082702279091\n",
      "Epoch:29 Batch:80 Log-loss0.04963244870305061\n",
      "Epoch:29 Batch:120 Log-loss0.0356021486222744\n",
      "Epoch:29 Batch:160 Log-loss0.040690675377845764\n",
      "Epoch:29 Batch:200 Log-loss0.033041466027498245\n",
      "Epoch:29 Batch:240 Log-loss0.03235315531492233\n",
      "Epoch:29 Batch:280 Log-loss0.035773199051618576\n",
      "Epoch:29 Batch:320 Log-loss0.038118261843919754\n",
      "Epoch:29 Batch:360 Log-loss0.04498462378978729\n",
      "Epoch:29 Batch:400 Log-loss0.023096343502402306\n",
      "Epoch:29 Batch:440 Log-loss0.03383990749716759\n",
      "Epoch:29 Batch:480 Log-loss0.028813928365707397\n",
      "Epoch:29 Batch:520 Log-loss0.03282356634736061\n",
      "Epoch:29 Batch:560 Log-loss0.03133593499660492\n",
      "Epoch average log loss:0.03733751744896706\n",
      "In Epoch29, val_loss:0.037420041735503444, best_val_loss:0.037304429734430704, best_auc:0.9910044825886049\n",
      "Epoch:30 Batch:0 Log-loss0.03911552205681801\n",
      "Epoch:30 Batch:40 Log-loss0.03467226028442383\n",
      "Epoch:30 Batch:80 Log-loss0.03250691294670105\n",
      "Epoch:30 Batch:120 Log-loss0.03251340612769127\n",
      "Epoch:30 Batch:160 Log-loss0.023781048133969307\n",
      "Epoch:30 Batch:200 Log-loss0.04295306280255318\n",
      "Epoch:30 Batch:240 Log-loss0.02543996460735798\n",
      "Epoch:30 Batch:280 Log-loss0.03443111851811409\n",
      "Epoch:30 Batch:320 Log-loss0.02882387489080429\n",
      "Epoch:30 Batch:360 Log-loss0.0342765748500824\n",
      "Epoch:30 Batch:400 Log-loss0.039308033883571625\n",
      "Epoch:30 Batch:440 Log-loss0.04348209500312805\n",
      "Epoch:30 Batch:480 Log-loss0.03645095229148865\n",
      "Epoch:30 Batch:520 Log-loss0.03152773529291153\n",
      "Epoch:30 Batch:560 Log-loss0.046589504927396774\n",
      "Epoch average log loss:0.037214027421682006\n",
      "In Epoch30, val_loss:0.03740864849559552, best_val_loss:0.037304429734430704, best_auc:0.9910044825886049\n",
      "Epoch:31 Batch:0 Log-loss0.03621390089392662\n",
      "Epoch:31 Batch:40 Log-loss0.0366111658513546\n",
      "Epoch:31 Batch:80 Log-loss0.037726882845163345\n",
      "Epoch:31 Batch:120 Log-loss0.035577792674303055\n",
      "Epoch:31 Batch:160 Log-loss0.04476091265678406\n",
      "Epoch:31 Batch:200 Log-loss0.04078008607029915\n",
      "Epoch:31 Batch:240 Log-loss0.028133900836110115\n",
      "Epoch:31 Batch:280 Log-loss0.04739617183804512\n",
      "Epoch:31 Batch:320 Log-loss0.028336739167571068\n",
      "Epoch:31 Batch:360 Log-loss0.03200688585639\n",
      "Epoch:31 Batch:400 Log-loss0.05288603901863098\n",
      "Epoch:31 Batch:440 Log-loss0.04009615257382393\n",
      "Epoch:31 Batch:480 Log-loss0.02746540866792202\n",
      "Epoch:31 Batch:520 Log-loss0.045566096901893616\n",
      "Epoch:31 Batch:560 Log-loss0.0390903577208519\n",
      "Epoch average log loss:0.03710601618513465\n",
      "In Epoch31, val_loss:0.03751212920434779, best_val_loss:0.037304429734430704, best_auc:0.9910044825886049\n",
      "Epoch:32 Batch:0 Log-loss0.03235878050327301\n",
      "Epoch:32 Batch:40 Log-loss0.03951994702219963\n",
      "Epoch:32 Batch:80 Log-loss0.03793124482035637\n",
      "Epoch:32 Batch:120 Log-loss0.03609497845172882\n",
      "Epoch:32 Batch:160 Log-loss0.029579564929008484\n",
      "Epoch:32 Batch:200 Log-loss0.03583548963069916\n",
      "Epoch:32 Batch:240 Log-loss0.028522349894046783\n",
      "Epoch:32 Batch:280 Log-loss0.02778896689414978\n",
      "Epoch:32 Batch:320 Log-loss0.041072964668273926\n",
      "Epoch:32 Batch:360 Log-loss0.04971577599644661\n",
      "Epoch:32 Batch:400 Log-loss0.035357046872377396\n",
      "Epoch:32 Batch:440 Log-loss0.029520394280552864\n",
      "Epoch:32 Batch:480 Log-loss0.04104345291852951\n",
      "Epoch:32 Batch:520 Log-loss0.048072654753923416\n",
      "Epoch:32 Batch:560 Log-loss0.04245525225996971\n",
      "Epoch average log loss:0.03726025967126978\n",
      "In Epoch32, val_loss:0.037323445443438, best_val_loss:0.037304429734430704, best_auc:0.9910044825886049\n",
      "Epoch:33 Batch:0 Log-loss0.039888303726911545\n",
      "Epoch:33 Batch:40 Log-loss0.044500455260276794\n",
      "Epoch:33 Batch:80 Log-loss0.03703587129712105\n",
      "Epoch:33 Batch:120 Log-loss0.027686061337590218\n",
      "Epoch:33 Batch:160 Log-loss0.025301454588770866\n",
      "Epoch:33 Batch:200 Log-loss0.039506182074546814\n",
      "Epoch:33 Batch:240 Log-loss0.03543456271290779\n",
      "Epoch:33 Batch:280 Log-loss0.052560921758413315\n",
      "Epoch:33 Batch:320 Log-loss0.03167984262108803\n",
      "Epoch:33 Batch:360 Log-loss0.06750019639730453\n",
      "Epoch:33 Batch:400 Log-loss0.04931351915001869\n",
      "Epoch:33 Batch:440 Log-loss0.02251945249736309\n",
      "Epoch:33 Batch:480 Log-loss0.048371847718954086\n",
      "Epoch:33 Batch:520 Log-loss0.05554429814219475\n",
      "Epoch:33 Batch:560 Log-loss0.042006947100162506\n",
      "Epoch average log loss:0.03698379938058289\n",
      "In Epoch33, val_loss:0.03736172140330222, best_val_loss:0.037304429734430704, best_auc:0.9910044825886049\n",
      "Epoch:34 Batch:0 Log-loss0.04706115648150444\n",
      "Epoch:34 Batch:40 Log-loss0.03505145385861397\n",
      "Epoch:34 Batch:80 Log-loss0.02830786071717739\n",
      "Epoch:34 Batch:120 Log-loss0.03441610559821129\n",
      "Epoch:34 Batch:160 Log-loss0.035055410116910934\n",
      "Epoch:34 Batch:200 Log-loss0.03024425357580185\n",
      "Epoch:34 Batch:240 Log-loss0.022813303396105766\n",
      "Epoch:34 Batch:280 Log-loss0.02995261549949646\n",
      "Epoch:34 Batch:320 Log-loss0.03702149912714958\n",
      "Epoch:34 Batch:360 Log-loss0.037547703832387924\n",
      "Epoch:34 Batch:400 Log-loss0.04885684326291084\n",
      "Epoch:34 Batch:440 Log-loss0.04003254696726799\n",
      "Epoch:34 Batch:480 Log-loss0.04238361492753029\n",
      "Epoch:34 Batch:520 Log-loss0.03836144506931305\n",
      "Epoch:34 Batch:560 Log-loss0.04289707541465759\n",
      "Epoch average log loss:0.03697435142738478\n",
      "In Epoch34, val_loss:0.03737018997549799, best_val_loss:0.037304429734430704, best_auc:0.9910044825886049\n",
      "Epoch:35 Batch:0 Log-loss0.04004276543855667\n",
      "Epoch:35 Batch:40 Log-loss0.04587090015411377\n",
      "Epoch:35 Batch:80 Log-loss0.03218475729227066\n",
      "Epoch:35 Batch:120 Log-loss0.027905261144042015\n",
      "Epoch:35 Batch:160 Log-loss0.03228725865483284\n",
      "Epoch:35 Batch:200 Log-loss0.03752206265926361\n",
      "Epoch:35 Batch:240 Log-loss0.04767546430230141\n",
      "Epoch:35 Batch:280 Log-loss0.039307091385126114\n",
      "Epoch:35 Batch:320 Log-loss0.0179136972874403\n",
      "Epoch:35 Batch:360 Log-loss0.04061436280608177\n",
      "Epoch:35 Batch:400 Log-loss0.03762539103627205\n",
      "Epoch:35 Batch:440 Log-loss0.045557986944913864\n",
      "Epoch:35 Batch:480 Log-loss0.04173189774155617\n",
      "Epoch:35 Batch:520 Log-loss0.04216820374131203\n",
      "Epoch:35 Batch:560 Log-loss0.030253440141677856\n",
      "Epoch average log loss:0.036922520528813556\n",
      "In Epoch35, val_loss:0.03731684651871435, best_val_loss:0.037304429734430704, best_auc:0.9910044825886049\n",
      "Epoch:36 Batch:0 Log-loss0.035305093973875046\n",
      "Epoch:36 Batch:40 Log-loss0.04011394456028938\n",
      "Epoch:36 Batch:80 Log-loss0.04056064411997795\n",
      "Epoch:36 Batch:120 Log-loss0.04096108302474022\n",
      "Epoch:36 Batch:160 Log-loss0.03373657539486885\n",
      "Epoch:36 Batch:200 Log-loss0.02063695900142193\n",
      "Epoch:36 Batch:240 Log-loss0.03962268307805061\n",
      "Epoch:36 Batch:280 Log-loss0.03508223220705986\n",
      "Epoch:36 Batch:320 Log-loss0.033051423728466034\n",
      "Epoch:36 Batch:360 Log-loss0.038693804293870926\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:36 Batch:400 Log-loss0.025833694264292717\n",
      "Epoch:36 Batch:440 Log-loss0.034409526735544205\n",
      "Epoch:36 Batch:480 Log-loss0.025018826127052307\n",
      "Epoch:36 Batch:520 Log-loss0.031979165971279144\n",
      "Epoch:36 Batch:560 Log-loss0.03731510415673256\n",
      "Epoch average log loss:0.03684867032071842\n",
      "In Epoch36, val_loss:0.03731820794399781, best_val_loss:0.037304429734430704, best_auc:0.9910044825886049\n",
      "Epoch:37 Batch:0 Log-loss0.041120968759059906\n",
      "Epoch:37 Batch:40 Log-loss0.04017426818609238\n",
      "Epoch:37 Batch:80 Log-loss0.052230436354875565\n",
      "Epoch:37 Batch:120 Log-loss0.03885991871356964\n",
      "Epoch:37 Batch:160 Log-loss0.02336188219487667\n",
      "Epoch:37 Batch:200 Log-loss0.037928953766822815\n",
      "Epoch:37 Batch:240 Log-loss0.0404653437435627\n",
      "Epoch:37 Batch:280 Log-loss0.01976540870964527\n",
      "Epoch:37 Batch:320 Log-loss0.0346110463142395\n",
      "Epoch:37 Batch:360 Log-loss0.04622014984488487\n",
      "Epoch:37 Batch:400 Log-loss0.03186018019914627\n",
      "Epoch:37 Batch:440 Log-loss0.04804309830069542\n",
      "Epoch:37 Batch:480 Log-loss0.03703493997454643\n",
      "Epoch:37 Batch:520 Log-loss0.028458960354328156\n",
      "Epoch:37 Batch:560 Log-loss0.03594613075256348\n",
      "Epoch average log loss:0.036987041351052796\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch37, val_loss:0.03721452095062645, best_val_loss:0.03721452095062645, best_auc:0.9910563496841317\n",
      "Epoch:38 Batch:0 Log-loss0.03987783566117287\n",
      "Epoch:38 Batch:40 Log-loss0.027346575632691383\n",
      "Epoch:38 Batch:80 Log-loss0.0414263978600502\n",
      "Epoch:38 Batch:120 Log-loss0.028205784037709236\n",
      "Epoch:38 Batch:160 Log-loss0.035298142582178116\n",
      "Epoch:38 Batch:200 Log-loss0.039726171642541885\n",
      "Epoch:38 Batch:240 Log-loss0.03260711207985878\n",
      "Epoch:38 Batch:280 Log-loss0.038873862475156784\n",
      "Epoch:38 Batch:320 Log-loss0.03600887209177017\n",
      "Epoch:38 Batch:360 Log-loss0.04509643092751503\n",
      "Epoch:38 Batch:400 Log-loss0.03643272817134857\n",
      "Epoch:38 Batch:440 Log-loss0.03178714960813522\n",
      "Epoch:38 Batch:480 Log-loss0.04762449860572815\n",
      "Epoch:38 Batch:520 Log-loss0.036666061729192734\n",
      "Epoch:38 Batch:560 Log-loss0.03847813606262207\n",
      "Epoch average log loss:0.03686469574791512\n",
      "In Epoch38, val_loss:0.03723224801060842, best_val_loss:0.03721452095062645, best_auc:0.9910563496841317\n",
      "Epoch:39 Batch:0 Log-loss0.028133779764175415\n",
      "Epoch:39 Batch:40 Log-loss0.029791109263896942\n",
      "Epoch:39 Batch:80 Log-loss0.030254440382122993\n",
      "Epoch:39 Batch:120 Log-loss0.0467497743666172\n",
      "Epoch:39 Batch:160 Log-loss0.04260368272662163\n",
      "Epoch:39 Batch:200 Log-loss0.03271186351776123\n",
      "Epoch:39 Batch:240 Log-loss0.044240813702344894\n",
      "Epoch:39 Batch:280 Log-loss0.031158223748207092\n",
      "Epoch:39 Batch:320 Log-loss0.034835439175367355\n",
      "Epoch:39 Batch:360 Log-loss0.04626837372779846\n",
      "Epoch:39 Batch:400 Log-loss0.039524663239717484\n",
      "Epoch:39 Batch:440 Log-loss0.04671291634440422\n",
      "Epoch:39 Batch:480 Log-loss0.033346205949783325\n",
      "Epoch:39 Batch:520 Log-loss0.03543519601225853\n",
      "Epoch:39 Batch:560 Log-loss0.027995886281132698\n",
      "Epoch average log loss:0.036896304806162204\n",
      "In Epoch39, val_loss:0.03723334782922209, best_val_loss:0.03721452095062645, best_auc:0.9910563496841317\n",
      "Epoch:40 Batch:0 Log-loss0.043709058314561844\n",
      "Epoch:40 Batch:40 Log-loss0.043579112738370895\n",
      "Epoch:40 Batch:80 Log-loss0.03198954090476036\n",
      "Epoch:40 Batch:120 Log-loss0.037705205380916595\n",
      "Epoch:40 Batch:160 Log-loss0.04166216775774956\n",
      "Epoch:40 Batch:200 Log-loss0.02933821827173233\n",
      "Epoch:40 Batch:240 Log-loss0.03673989698290825\n",
      "Epoch:40 Batch:280 Log-loss0.01995866931974888\n",
      "Epoch:40 Batch:320 Log-loss0.03707017004489899\n",
      "Epoch:40 Batch:360 Log-loss0.041313398629426956\n",
      "Epoch:40 Batch:400 Log-loss0.027599966153502464\n",
      "Epoch:40 Batch:440 Log-loss0.03295275568962097\n",
      "Epoch:40 Batch:480 Log-loss0.03807887062430382\n",
      "Epoch:40 Batch:520 Log-loss0.0326271615922451\n",
      "Epoch:40 Batch:560 Log-loss0.040498729795217514\n",
      "Epoch average log loss:0.03663945028757942\n",
      "In Epoch40, val_loss:0.03726435313428463, best_val_loss:0.03721452095062645, best_auc:0.9910563496841317\n",
      "Epoch:41 Batch:0 Log-loss0.03612907975912094\n",
      "Epoch:41 Batch:40 Log-loss0.03994673863053322\n",
      "Epoch:41 Batch:80 Log-loss0.03579460456967354\n",
      "Epoch:41 Batch:120 Log-loss0.030826019123196602\n",
      "Epoch:41 Batch:160 Log-loss0.03325476497411728\n",
      "Epoch:41 Batch:200 Log-loss0.03340085223317146\n",
      "Epoch:41 Batch:240 Log-loss0.02527732215821743\n",
      "Epoch:41 Batch:280 Log-loss0.028964413329958916\n",
      "Epoch:41 Batch:320 Log-loss0.05372707173228264\n",
      "Epoch:41 Batch:360 Log-loss0.03269604220986366\n",
      "Epoch:41 Batch:400 Log-loss0.03749432414770126\n",
      "Epoch:41 Batch:440 Log-loss0.038371626287698746\n",
      "Epoch:41 Batch:480 Log-loss0.03621261939406395\n",
      "Epoch:41 Batch:520 Log-loss0.04131593927741051\n",
      "Epoch:41 Batch:560 Log-loss0.033118054270744324\n",
      "Epoch average log loss:0.03666430843794452\n",
      "In Epoch41, val_loss:0.0373618681303052, best_val_loss:0.03721452095062645, best_auc:0.9910563496841317\n",
      "Epoch:42 Batch:0 Log-loss0.02290797233581543\n",
      "Epoch:42 Batch:40 Log-loss0.045806076377630234\n",
      "Epoch:42 Batch:80 Log-loss0.053405825048685074\n",
      "Epoch:42 Batch:120 Log-loss0.040981970727443695\n",
      "Epoch:42 Batch:160 Log-loss0.025832049548625946\n",
      "Epoch:42 Batch:200 Log-loss0.04555807635188103\n",
      "Epoch:42 Batch:240 Log-loss0.04294823482632637\n",
      "Epoch:42 Batch:280 Log-loss0.03391856327652931\n",
      "Epoch:42 Batch:320 Log-loss0.0366503931581974\n",
      "Epoch:42 Batch:360 Log-loss0.031503837555646896\n",
      "Epoch:42 Batch:400 Log-loss0.024079924449324608\n",
      "Epoch:42 Batch:440 Log-loss0.0375630259513855\n",
      "Epoch:42 Batch:480 Log-loss0.03705831989645958\n",
      "Epoch:42 Batch:520 Log-loss0.03819258511066437\n",
      "Epoch:42 Batch:560 Log-loss0.03642009571194649\n",
      "Epoch average log loss:0.036643676947070554\n",
      "In Epoch42, val_loss:0.03722733843375871, best_val_loss:0.03721452095062645, best_auc:0.9910563496841317\n",
      "Epoch:43 Batch:0 Log-loss0.039488647133111954\n",
      "Epoch:43 Batch:40 Log-loss0.0329151451587677\n",
      "Epoch:43 Batch:80 Log-loss0.03364669159054756\n",
      "Epoch:43 Batch:120 Log-loss0.04449990764260292\n",
      "Epoch:43 Batch:160 Log-loss0.026626383885741234\n",
      "Epoch:43 Batch:200 Log-loss0.04286964610219002\n",
      "Epoch:43 Batch:240 Log-loss0.03729188069701195\n",
      "Epoch:43 Batch:280 Log-loss0.02560483105480671\n",
      "Epoch:43 Batch:320 Log-loss0.03476215526461601\n",
      "Epoch:43 Batch:360 Log-loss0.043883729726076126\n",
      "Epoch:43 Batch:400 Log-loss0.03585006296634674\n",
      "Epoch:43 Batch:440 Log-loss0.038850173354148865\n",
      "Epoch:43 Batch:480 Log-loss0.028934350237250328\n",
      "Epoch:43 Batch:520 Log-loss0.03674283251166344\n",
      "Epoch:43 Batch:560 Log-loss0.03481120243668556\n",
      "Epoch average log loss:0.03669755923162613\n",
      "In Epoch43, val_loss:0.03727318923817192, best_val_loss:0.03721452095062645, best_auc:0.9910563496841317\n",
      "Epoch:44 Batch:0 Log-loss0.037337884306907654\n",
      "Epoch:44 Batch:40 Log-loss0.030317308381199837\n",
      "Epoch:44 Batch:80 Log-loss0.024984577670693398\n",
      "Epoch:44 Batch:120 Log-loss0.021501801908016205\n",
      "Epoch:44 Batch:160 Log-loss0.028642093762755394\n",
      "Epoch:44 Batch:200 Log-loss0.04227946698665619\n",
      "Epoch:44 Batch:240 Log-loss0.03831757977604866\n",
      "Epoch:44 Batch:280 Log-loss0.02147788368165493\n",
      "Epoch:44 Batch:320 Log-loss0.043839242309331894\n",
      "Epoch:44 Batch:360 Log-loss0.06582730263471603\n",
      "Epoch:44 Batch:400 Log-loss0.04409192129969597\n",
      "Epoch:44 Batch:440 Log-loss0.03337136283516884\n",
      "Epoch:44 Batch:480 Log-loss0.04505902901291847\n",
      "Epoch:44 Batch:520 Log-loss0.057101499289274216\n",
      "Epoch:44 Batch:560 Log-loss0.03644115850329399\n",
      "Epoch average log loss:0.036401422029095035\n",
      "In Epoch44, val_loss:0.03725870813084214, best_val_loss:0.03721452095062645, best_auc:0.9910563496841317\n",
      "Epoch:45 Batch:0 Log-loss0.04566578567028046\n",
      "Epoch:45 Batch:40 Log-loss0.02607955038547516\n",
      "Epoch:45 Batch:80 Log-loss0.031336404383182526\n",
      "Epoch:45 Batch:120 Log-loss0.04661528766155243\n",
      "Epoch:45 Batch:160 Log-loss0.040248800069093704\n",
      "Epoch:45 Batch:200 Log-loss0.04499531164765358\n",
      "Epoch:45 Batch:240 Log-loss0.024962330237030983\n",
      "Epoch:45 Batch:280 Log-loss0.04568517208099365\n",
      "Epoch:45 Batch:320 Log-loss0.026689253747463226\n",
      "Epoch:45 Batch:360 Log-loss0.04200820252299309\n",
      "Epoch:45 Batch:400 Log-loss0.04147621616721153\n",
      "Epoch:45 Batch:440 Log-loss0.029302703216671944\n",
      "Epoch:45 Batch:480 Log-loss0.04504382982850075\n",
      "Epoch:45 Batch:520 Log-loss0.047286082059144974\n",
      "Epoch:45 Batch:560 Log-loss0.030183210968971252\n",
      "Epoch average log loss:0.03665167796425521\n",
      "In Epoch45, val_loss:0.03723726259943901, best_val_loss:0.03721452095062645, best_auc:0.9910563496841317\n",
      "Epoch:46 Batch:0 Log-loss0.03107951022684574\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:46 Batch:40 Log-loss0.04640065133571625\n",
      "Epoch:46 Batch:80 Log-loss0.03536185994744301\n",
      "Epoch:46 Batch:120 Log-loss0.024372020736336708\n",
      "Epoch:46 Batch:160 Log-loss0.04049365594983101\n",
      "Epoch:46 Batch:200 Log-loss0.03226381912827492\n",
      "Epoch:46 Batch:240 Log-loss0.029100291430950165\n",
      "Epoch:46 Batch:280 Log-loss0.048882756382226944\n",
      "Epoch:46 Batch:320 Log-loss0.033108893781900406\n",
      "Epoch:46 Batch:360 Log-loss0.03262598440051079\n",
      "Epoch:46 Batch:400 Log-loss0.032616760581731796\n",
      "Epoch:46 Batch:440 Log-loss0.030659863725304604\n",
      "Epoch:46 Batch:480 Log-loss0.03488423302769661\n",
      "Epoch:46 Batch:520 Log-loss0.02750486694276333\n",
      "Epoch:46 Batch:560 Log-loss0.0325506292283535\n",
      "Epoch average log loss:0.03659242326061108\n",
      "In Epoch46, val_loss:0.03727975592343704, best_val_loss:0.03721452095062645, best_auc:0.9910563496841317\n",
      "Epoch:47 Batch:0 Log-loss0.028656980022788048\n",
      "Epoch:47 Batch:40 Log-loss0.03158046677708626\n",
      "Epoch:47 Batch:80 Log-loss0.04087788984179497\n",
      "Epoch:47 Batch:120 Log-loss0.0454706996679306\n",
      "Epoch:47 Batch:160 Log-loss0.046894192695617676\n",
      "Epoch:47 Batch:200 Log-loss0.029633397236466408\n",
      "Epoch:47 Batch:240 Log-loss0.0429418683052063\n",
      "Epoch:47 Batch:280 Log-loss0.04532318189740181\n",
      "Epoch:47 Batch:320 Log-loss0.030384710058569908\n",
      "Epoch:47 Batch:360 Log-loss0.05217564478516579\n",
      "Epoch:47 Batch:400 Log-loss0.035148490220308304\n",
      "Epoch:47 Batch:440 Log-loss0.04682619497179985\n",
      "Epoch:47 Batch:480 Log-loss0.023982038721442223\n",
      "Epoch:47 Batch:520 Log-loss0.03694840893149376\n",
      "Epoch:47 Batch:560 Log-loss0.0369701124727726\n",
      "Epoch average log loss:0.03666075868199446\n",
      "A pre-trained model at temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt has been loaded.\n",
      "\n",
      "Model has been saved as checkpoints/FASTTXT_BGRU_64_642.pt.\n",
      "\n",
      "Choose the torch base model.\n",
      "Training on fold 3\n",
      "Epoch:1 Batch:0 Log-loss0.6911925673484802\n",
      "Epoch:1 Batch:40 Log-loss0.16118021309375763\n",
      "Epoch:1 Batch:80 Log-loss0.08144784718751907\n",
      "Epoch:1 Batch:120 Log-loss0.0568089485168457\n",
      "Epoch:1 Batch:160 Log-loss0.05854390189051628\n",
      "Epoch:1 Batch:200 Log-loss0.046212974935770035\n",
      "Epoch:1 Batch:240 Log-loss0.056100621819496155\n",
      "Epoch:1 Batch:280 Log-loss0.04901039972901344\n",
      "Epoch:1 Batch:320 Log-loss0.057980041950941086\n",
      "Epoch:1 Batch:360 Log-loss0.052766647189855576\n",
      "Epoch:1 Batch:400 Log-loss0.0496533028781414\n",
      "Epoch:1 Batch:440 Log-loss0.05953006446361542\n",
      "Epoch:1 Batch:480 Log-loss0.05294634401798248\n",
      "Epoch:1 Batch:520 Log-loss0.04784952476620674\n",
      "Epoch:1 Batch:560 Log-loss0.0740078017115593\n",
      "Epoch average log loss:0.08224491564157818\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch1, val_loss:0.04864362156977225, best_val_loss:0.04864362156977225, best_auc:0.9700824643159547\n",
      "Epoch:2 Batch:0 Log-loss0.04907487332820892\n",
      "Epoch:2 Batch:40 Log-loss0.03561002388596535\n",
      "Epoch:2 Batch:80 Log-loss0.07106103748083115\n",
      "Epoch:2 Batch:120 Log-loss0.06324150413274765\n",
      "Epoch:2 Batch:160 Log-loss0.06156940758228302\n",
      "Epoch:2 Batch:200 Log-loss0.060762908309698105\n",
      "Epoch:2 Batch:240 Log-loss0.049155037850141525\n",
      "Epoch:2 Batch:280 Log-loss0.055987510830163956\n",
      "Epoch:2 Batch:320 Log-loss0.05644068494439125\n",
      "Epoch:2 Batch:360 Log-loss0.04107281193137169\n",
      "Epoch:2 Batch:400 Log-loss0.05409202352166176\n",
      "Epoch:2 Batch:440 Log-loss0.05741746351122856\n",
      "Epoch:2 Batch:480 Log-loss0.05258642137050629\n",
      "Epoch:2 Batch:520 Log-loss0.04661299288272858\n",
      "Epoch:2 Batch:560 Log-loss0.06830805540084839\n",
      "Epoch average log loss:0.049705012149310536\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch2, val_loss:0.0462583321082248, best_val_loss:0.0462583321082248, best_auc:0.9770363948102466\n",
      "Epoch:3 Batch:0 Log-loss0.059345513582229614\n",
      "Epoch:3 Batch:40 Log-loss0.042615752667188644\n",
      "Epoch:3 Batch:80 Log-loss0.04435621201992035\n",
      "Epoch:3 Batch:120 Log-loss0.04132319986820221\n",
      "Epoch:3 Batch:160 Log-loss0.05350503697991371\n",
      "Epoch:3 Batch:200 Log-loss0.03188984841108322\n",
      "Epoch:3 Batch:240 Log-loss0.05391533672809601\n",
      "Epoch:3 Batch:280 Log-loss0.05395090952515602\n",
      "Epoch:3 Batch:320 Log-loss0.04282085597515106\n",
      "Epoch:3 Batch:360 Log-loss0.05661889538168907\n",
      "Epoch:3 Batch:400 Log-loss0.04024967551231384\n",
      "Epoch:3 Batch:440 Log-loss0.04257488250732422\n",
      "Epoch:3 Batch:480 Log-loss0.0431118905544281\n",
      "Epoch:3 Batch:520 Log-loss0.0488051176071167\n",
      "Epoch:3 Batch:560 Log-loss0.034407444298267365\n",
      "Epoch average log loss:0.04720795983282317\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch3, val_loss:0.04449434554228072, best_val_loss:0.04449434554228072, best_auc:0.9801683127883707\n",
      "Epoch:4 Batch:0 Log-loss0.04236316680908203\n",
      "Epoch:4 Batch:40 Log-loss0.054248277097940445\n",
      "Epoch:4 Batch:80 Log-loss0.04171714559197426\n",
      "Epoch:4 Batch:120 Log-loss0.05088618025183678\n",
      "Epoch:4 Batch:160 Log-loss0.04507080838084221\n",
      "Epoch:4 Batch:200 Log-loss0.04883298650383949\n",
      "Epoch:4 Batch:240 Log-loss0.05659480392932892\n",
      "Epoch:4 Batch:280 Log-loss0.05886213853955269\n",
      "Epoch:4 Batch:320 Log-loss0.04107537865638733\n",
      "Epoch:4 Batch:360 Log-loss0.03261169418692589\n",
      "Epoch:4 Batch:400 Log-loss0.05327650532126427\n",
      "Epoch:4 Batch:440 Log-loss0.04158240929245949\n",
      "Epoch:4 Batch:480 Log-loss0.0571051649749279\n",
      "Epoch:4 Batch:520 Log-loss0.05032431706786156\n",
      "Epoch:4 Batch:560 Log-loss0.034923262894153595\n",
      "Epoch average log loss:0.04583480191483561\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch4, val_loss:0.04385006792729496, best_val_loss:0.04385006792729496, best_auc:0.9816364463171\n",
      "Epoch:5 Batch:0 Log-loss0.04507191479206085\n",
      "Epoch:5 Batch:40 Log-loss0.03533023223280907\n",
      "Epoch:5 Batch:80 Log-loss0.03872460126876831\n",
      "Epoch:5 Batch:120 Log-loss0.051133379340171814\n",
      "Epoch:5 Batch:160 Log-loss0.03845546767115593\n",
      "Epoch:5 Batch:200 Log-loss0.0510370247066021\n",
      "Epoch:5 Batch:240 Log-loss0.034767698496580124\n",
      "Epoch:5 Batch:280 Log-loss0.04441557452082634\n",
      "Epoch:5 Batch:320 Log-loss0.04622005298733711\n",
      "Epoch:5 Batch:360 Log-loss0.04317426308989525\n",
      "Epoch:5 Batch:400 Log-loss0.036369115114212036\n",
      "Epoch:5 Batch:440 Log-loss0.04984060302376747\n",
      "Epoch:5 Batch:480 Log-loss0.056038931012153625\n",
      "Epoch:5 Batch:520 Log-loss0.046743664890527725\n",
      "Epoch:5 Batch:560 Log-loss0.043532054871320724\n",
      "Epoch average log loss:0.04456721129494586\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch5, val_loss:0.04304901746311496, best_val_loss:0.04304901746311496, best_auc:0.9828399880441578\n",
      "Epoch:6 Batch:0 Log-loss0.029115991666913033\n",
      "Epoch:6 Batch:40 Log-loss0.05255502089858055\n",
      "Epoch:6 Batch:80 Log-loss0.03053230606019497\n",
      "Epoch:6 Batch:120 Log-loss0.029944242909550667\n",
      "Epoch:6 Batch:160 Log-loss0.03332075849175453\n",
      "Epoch:6 Batch:200 Log-loss0.03336980566382408\n",
      "Epoch:6 Batch:240 Log-loss0.04507264494895935\n",
      "Epoch:6 Batch:280 Log-loss0.035126496106386185\n",
      "Epoch:6 Batch:320 Log-loss0.049618203192949295\n",
      "Epoch:6 Batch:360 Log-loss0.0379767082631588\n",
      "Epoch:6 Batch:400 Log-loss0.03681567311286926\n",
      "Epoch:6 Batch:440 Log-loss0.04152696207165718\n",
      "Epoch:6 Batch:480 Log-loss0.04730619117617607\n",
      "Epoch:6 Batch:520 Log-loss0.07678546011447906\n",
      "Epoch:6 Batch:560 Log-loss0.03040321171283722\n",
      "Epoch average log loss:0.043719632923603056\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch6, val_loss:0.04232006919364501, best_val_loss:0.04232006919364501, best_auc:0.9838490034800205\n",
      "Epoch:7 Batch:0 Log-loss0.05257992818951607\n",
      "Epoch:7 Batch:40 Log-loss0.028911544010043144\n",
      "Epoch:7 Batch:80 Log-loss0.04554422199726105\n",
      "Epoch:7 Batch:120 Log-loss0.026432432234287262\n",
      "Epoch:7 Batch:160 Log-loss0.03258698433637619\n",
      "Epoch:7 Batch:200 Log-loss0.029201043769717216\n",
      "Epoch:7 Batch:240 Log-loss0.03640769422054291\n",
      "Epoch:7 Batch:280 Log-loss0.03481953591108322\n",
      "Epoch:7 Batch:320 Log-loss0.02683151327073574\n",
      "Epoch:7 Batch:360 Log-loss0.034897539764642715\n",
      "Epoch:7 Batch:400 Log-loss0.03464978560805321\n",
      "Epoch:7 Batch:440 Log-loss0.037285540252923965\n",
      "Epoch:7 Batch:480 Log-loss0.06519501656293869\n",
      "Epoch:7 Batch:520 Log-loss0.045262306928634644\n",
      "Epoch:7 Batch:560 Log-loss0.043975017964839935\n",
      "Epoch average log loss:0.042563923782602486\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch7, val_loss:0.041812803226995605, best_val_loss:0.041812803226995605, best_auc:0.9844266056327213\n",
      "Epoch:8 Batch:0 Log-loss0.035837363451719284\n",
      "Epoch:8 Batch:40 Log-loss0.03917704150080681\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:8 Batch:80 Log-loss0.026733243837952614\n",
      "Epoch:8 Batch:120 Log-loss0.024655088782310486\n",
      "Epoch:8 Batch:160 Log-loss0.0476449616253376\n",
      "Epoch:8 Batch:200 Log-loss0.04001067206263542\n",
      "Epoch:8 Batch:240 Log-loss0.04956318438053131\n",
      "Epoch:8 Batch:280 Log-loss0.03784343972802162\n",
      "Epoch:8 Batch:320 Log-loss0.05257384106516838\n",
      "Epoch:8 Batch:360 Log-loss0.029924876987934113\n",
      "Epoch:8 Batch:400 Log-loss0.04770452156662941\n",
      "Epoch:8 Batch:440 Log-loss0.04735509678721428\n",
      "Epoch:8 Batch:480 Log-loss0.045842722058296204\n",
      "Epoch:8 Batch:520 Log-loss0.03182806447148323\n",
      "Epoch:8 Batch:560 Log-loss0.044424816966056824\n",
      "Epoch average log loss:0.04223519116640091\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch8, val_loss:0.04113873117097322, best_val_loss:0.04113873117097322, best_auc:0.9854845190711595\n",
      "Epoch:9 Batch:0 Log-loss0.04345192387700081\n",
      "Epoch:9 Batch:40 Log-loss0.04898719862103462\n",
      "Epoch:9 Batch:80 Log-loss0.06000930070877075\n",
      "Epoch:9 Batch:120 Log-loss0.04804167151451111\n",
      "Epoch:9 Batch:160 Log-loss0.036462362855672836\n",
      "Epoch:9 Batch:200 Log-loss0.029616842046380043\n",
      "Epoch:9 Batch:240 Log-loss0.04544427618384361\n",
      "Epoch:9 Batch:280 Log-loss0.04075387120246887\n",
      "Epoch:9 Batch:320 Log-loss0.05407369136810303\n",
      "Epoch:9 Batch:360 Log-loss0.03629116341471672\n",
      "Epoch:9 Batch:400 Log-loss0.0333789587020874\n",
      "Epoch:9 Batch:440 Log-loss0.039542537182569504\n",
      "Epoch:9 Batch:480 Log-loss0.05656053498387337\n",
      "Epoch:9 Batch:520 Log-loss0.050507400184869766\n",
      "Epoch:9 Batch:560 Log-loss0.028922276571393013\n",
      "Epoch average log loss:0.04161241609669689\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch9, val_loss:0.04105491082541853, best_val_loss:0.04105491082541853, best_auc:0.985516595082177\n",
      "Epoch:10 Batch:0 Log-loss0.041305068880319595\n",
      "Epoch:10 Batch:40 Log-loss0.04514126107096672\n",
      "Epoch:10 Batch:80 Log-loss0.0537283681333065\n",
      "Epoch:10 Batch:120 Log-loss0.042842913419008255\n",
      "Epoch:10 Batch:160 Log-loss0.03558054938912392\n",
      "Epoch:10 Batch:200 Log-loss0.05462509021162987\n",
      "Epoch:10 Batch:240 Log-loss0.04473850131034851\n",
      "Epoch:10 Batch:280 Log-loss0.04107469320297241\n",
      "Epoch:10 Batch:320 Log-loss0.03052678517997265\n",
      "Epoch:10 Batch:360 Log-loss0.04492838308215141\n",
      "Epoch:10 Batch:400 Log-loss0.04406663775444031\n",
      "Epoch:10 Batch:440 Log-loss0.051702458411455154\n",
      "Epoch:10 Batch:480 Log-loss0.04290172457695007\n",
      "Epoch:10 Batch:520 Log-loss0.0539996437728405\n",
      "Epoch:10 Batch:560 Log-loss0.05281610041856766\n",
      "Epoch average log loss:0.04096412943742637\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch10, val_loss:0.0406388970143587, best_val_loss:0.0406388970143587, best_auc:0.986379104303129\n",
      "Epoch:11 Batch:0 Log-loss0.036584313958883286\n",
      "Epoch:11 Batch:40 Log-loss0.057383373379707336\n",
      "Epoch:11 Batch:80 Log-loss0.03537513688206673\n",
      "Epoch:11 Batch:120 Log-loss0.04983089491724968\n",
      "Epoch:11 Batch:160 Log-loss0.04112297669053078\n",
      "Epoch:11 Batch:200 Log-loss0.03242986276745796\n",
      "Epoch:11 Batch:240 Log-loss0.027837494388222694\n",
      "Epoch:11 Batch:280 Log-loss0.03949887678027153\n",
      "Epoch:11 Batch:320 Log-loss0.040615715086460114\n",
      "Epoch:11 Batch:360 Log-loss0.03859047591686249\n",
      "Epoch:11 Batch:400 Log-loss0.036360714584589005\n",
      "Epoch:11 Batch:440 Log-loss0.04400143027305603\n",
      "Epoch:11 Batch:480 Log-loss0.03524499014019966\n",
      "Epoch:11 Batch:520 Log-loss0.031263116747140884\n",
      "Epoch:11 Batch:560 Log-loss0.039828527718782425\n",
      "Epoch average log loss:0.04052995417732745\n",
      "In Epoch11, val_loss:0.040684770126903816, best_val_loss:0.0406388970143587, best_auc:0.986379104303129\n",
      "Epoch:12 Batch:0 Log-loss0.05004020035266876\n",
      "Epoch:12 Batch:40 Log-loss0.04116477444767952\n",
      "Epoch:12 Batch:80 Log-loss0.04769358038902283\n",
      "Epoch:12 Batch:120 Log-loss0.040055833756923676\n",
      "Epoch:12 Batch:160 Log-loss0.039949849247932434\n",
      "Epoch:12 Batch:200 Log-loss0.03153596445918083\n",
      "Epoch:12 Batch:240 Log-loss0.03463715314865112\n",
      "Epoch:12 Batch:280 Log-loss0.03186165541410446\n",
      "Epoch:12 Batch:320 Log-loss0.047728631645441055\n",
      "Epoch:12 Batch:360 Log-loss0.0476982407271862\n",
      "Epoch:12 Batch:400 Log-loss0.03540520742535591\n",
      "Epoch:12 Batch:440 Log-loss0.03966444358229637\n",
      "Epoch:12 Batch:480 Log-loss0.04689610004425049\n",
      "Epoch:12 Batch:520 Log-loss0.03877287730574608\n",
      "Epoch:12 Batch:560 Log-loss0.04238324239850044\n",
      "Epoch average log loss:0.04012675240103688\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch12, val_loss:0.04038717792688718, best_val_loss:0.04038717792688718, best_auc:0.9867575504772522\n",
      "Epoch:13 Batch:0 Log-loss0.04362007975578308\n",
      "Epoch:13 Batch:40 Log-loss0.04331636056303978\n",
      "Epoch:13 Batch:80 Log-loss0.03608546778559685\n",
      "Epoch:13 Batch:120 Log-loss0.035634033381938934\n",
      "Epoch:13 Batch:160 Log-loss0.04794049263000488\n",
      "Epoch:13 Batch:200 Log-loss0.03444548323750496\n",
      "Epoch:13 Batch:240 Log-loss0.0394890122115612\n",
      "Epoch:13 Batch:280 Log-loss0.04419383406639099\n",
      "Epoch:13 Batch:320 Log-loss0.0421270877122879\n",
      "Epoch:13 Batch:360 Log-loss0.036817897111177444\n",
      "Epoch:13 Batch:400 Log-loss0.04676416888833046\n",
      "Epoch:13 Batch:440 Log-loss0.04593288525938988\n",
      "Epoch:13 Batch:480 Log-loss0.0346481055021286\n",
      "Epoch:13 Batch:520 Log-loss0.03604202717542648\n",
      "Epoch:13 Batch:560 Log-loss0.04342382401227951\n",
      "Epoch average log loss:0.039665462475802216\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch13, val_loss:0.04010869097425087, best_val_loss:0.04010869097425087, best_auc:0.9870293146887302\n",
      "Epoch:14 Batch:0 Log-loss0.04258454218506813\n",
      "Epoch:14 Batch:40 Log-loss0.03879258781671524\n",
      "Epoch:14 Batch:80 Log-loss0.03102661669254303\n",
      "Epoch:14 Batch:120 Log-loss0.03939244523644447\n",
      "Epoch:14 Batch:160 Log-loss0.0742049589753151\n",
      "Epoch:14 Batch:200 Log-loss0.032459158450365067\n",
      "Epoch:14 Batch:240 Log-loss0.03623804822564125\n",
      "Epoch:14 Batch:280 Log-loss0.037813398987054825\n",
      "Epoch:14 Batch:320 Log-loss0.025630073621869087\n",
      "Epoch:14 Batch:360 Log-loss0.04758834466338158\n",
      "Epoch:14 Batch:400 Log-loss0.03379642218351364\n",
      "Epoch:14 Batch:440 Log-loss0.03715865686535835\n",
      "Epoch:14 Batch:480 Log-loss0.040465522557497025\n",
      "Epoch:14 Batch:520 Log-loss0.0416395328938961\n",
      "Epoch:14 Batch:560 Log-loss0.04538166522979736\n",
      "Epoch average log loss:0.039337050299426274\n",
      "In Epoch14, val_loss:0.04027508008516128, best_val_loss:0.04010869097425087, best_auc:0.9870293146887302\n",
      "Epoch:15 Batch:0 Log-loss0.041112322360277176\n",
      "Epoch:15 Batch:40 Log-loss0.04615861549973488\n",
      "Epoch:15 Batch:80 Log-loss0.03900519013404846\n",
      "Epoch:15 Batch:120 Log-loss0.04227766767144203\n",
      "Epoch:15 Batch:160 Log-loss0.029696213081479073\n",
      "Epoch:15 Batch:200 Log-loss0.05254432559013367\n",
      "Epoch:15 Batch:240 Log-loss0.0528242327272892\n",
      "Epoch:15 Batch:280 Log-loss0.053050149232149124\n",
      "Epoch:15 Batch:320 Log-loss0.041954487562179565\n",
      "Epoch:15 Batch:360 Log-loss0.04965021833777428\n",
      "Epoch:15 Batch:400 Log-loss0.03692000359296799\n",
      "Epoch:15 Batch:440 Log-loss0.041625265032052994\n",
      "Epoch:15 Batch:480 Log-loss0.043988510966300964\n",
      "Epoch:15 Batch:520 Log-loss0.03403458371758461\n",
      "Epoch:15 Batch:560 Log-loss0.04291795194149017\n",
      "Epoch average log loss:0.03900007550005934\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch15, val_loss:0.040001944089166214, best_val_loss:0.040001944089166214, best_auc:0.987347869350642\n",
      "Epoch:16 Batch:0 Log-loss0.04705185815691948\n",
      "Epoch:16 Batch:40 Log-loss0.04203365370631218\n",
      "Epoch:16 Batch:80 Log-loss0.04584623873233795\n",
      "Epoch:16 Batch:120 Log-loss0.04263225197792053\n",
      "Epoch:16 Batch:160 Log-loss0.044441837817430496\n",
      "Epoch:16 Batch:200 Log-loss0.028613626956939697\n",
      "Epoch:16 Batch:240 Log-loss0.04861168563365936\n",
      "Epoch:16 Batch:280 Log-loss0.04297227784991264\n",
      "Epoch:16 Batch:320 Log-loss0.050521086901426315\n",
      "Epoch:16 Batch:360 Log-loss0.03778441250324249\n",
      "Epoch:16 Batch:400 Log-loss0.033074986189603806\n",
      "Epoch:16 Batch:440 Log-loss0.022824300453066826\n",
      "Epoch:16 Batch:480 Log-loss0.03694048896431923\n",
      "Epoch:16 Batch:520 Log-loss0.05390022695064545\n",
      "Epoch:16 Batch:560 Log-loss0.0433659590780735\n",
      "Epoch average log loss:0.038948424150502044\n",
      "In Epoch16, val_loss:0.0403569265414679, best_val_loss:0.040001944089166214, best_auc:0.987347869350642\n",
      "Epoch:17 Batch:0 Log-loss0.037440355867147446\n",
      "Epoch:17 Batch:40 Log-loss0.03931276872754097\n",
      "Epoch:17 Batch:80 Log-loss0.043159183114767075\n",
      "Epoch:17 Batch:120 Log-loss0.03450457379221916\n",
      "Epoch:17 Batch:160 Log-loss0.047377318143844604\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:17 Batch:200 Log-loss0.03365105763077736\n",
      "Epoch:17 Batch:240 Log-loss0.03986181318759918\n",
      "Epoch:17 Batch:280 Log-loss0.03372947871685028\n",
      "Epoch:17 Batch:320 Log-loss0.04586018994450569\n",
      "Epoch:17 Batch:360 Log-loss0.03453269973397255\n",
      "Epoch:17 Batch:400 Log-loss0.040486890822649\n",
      "Epoch:17 Batch:440 Log-loss0.037645578384399414\n",
      "Epoch:17 Batch:480 Log-loss0.030009711161255836\n",
      "Epoch:17 Batch:520 Log-loss0.040392402559518814\n",
      "Epoch:17 Batch:560 Log-loss0.03294892609119415\n",
      "Epoch average log loss:0.03877530910540372\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch17, val_loss:0.03991126317927755, best_val_loss:0.03991126317927755, best_auc:0.9880556862096544\n",
      "Epoch:18 Batch:0 Log-loss0.040074821561574936\n",
      "Epoch:18 Batch:40 Log-loss0.040846195071935654\n",
      "Epoch:18 Batch:80 Log-loss0.037810105830430984\n",
      "Epoch:18 Batch:120 Log-loss0.0367041677236557\n",
      "Epoch:18 Batch:160 Log-loss0.040505167096853256\n",
      "Epoch:18 Batch:200 Log-loss0.034350231289863586\n",
      "Epoch:18 Batch:240 Log-loss0.037482354789972305\n",
      "Epoch:18 Batch:280 Log-loss0.04299391806125641\n",
      "Epoch:18 Batch:320 Log-loss0.036482322961091995\n",
      "Epoch:18 Batch:360 Log-loss0.038322657346725464\n",
      "Epoch:18 Batch:400 Log-loss0.0393458753824234\n",
      "Epoch:18 Batch:440 Log-loss0.031200610101222992\n",
      "Epoch:18 Batch:480 Log-loss0.06079860031604767\n",
      "Epoch:18 Batch:520 Log-loss0.04315680265426636\n",
      "Epoch:18 Batch:560 Log-loss0.03612951561808586\n",
      "Epoch average log loss:0.03858039028543447\n",
      "In Epoch18, val_loss:0.03994859513341296, best_val_loss:0.03991126317927755, best_auc:0.9880556862096544\n",
      "Epoch:19 Batch:0 Log-loss0.04450860992074013\n",
      "Epoch:19 Batch:40 Log-loss0.061168599873781204\n",
      "Epoch:19 Batch:80 Log-loss0.03649032115936279\n",
      "Epoch:19 Batch:120 Log-loss0.032987773418426514\n",
      "Epoch:19 Batch:160 Log-loss0.023938216269016266\n",
      "Epoch:19 Batch:200 Log-loss0.044326577335596085\n",
      "Epoch:19 Batch:240 Log-loss0.03488237038254738\n",
      "Epoch:19 Batch:280 Log-loss0.0402800515294075\n",
      "Epoch:19 Batch:320 Log-loss0.030560052022337914\n",
      "Epoch:19 Batch:360 Log-loss0.04171783849596977\n",
      "Epoch:19 Batch:400 Log-loss0.0471147745847702\n",
      "Epoch:19 Batch:440 Log-loss0.029946185648441315\n",
      "Epoch:19 Batch:480 Log-loss0.03612474724650383\n",
      "Epoch:19 Batch:520 Log-loss0.034044310450553894\n",
      "Epoch:19 Batch:560 Log-loss0.03823896124958992\n",
      "Epoch average log loss:0.03831573251102652\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch19, val_loss:0.0398429300155575, best_val_loss:0.0398429300155575, best_auc:0.9879269447095385\n",
      "Epoch:20 Batch:0 Log-loss0.02983924187719822\n",
      "Epoch:20 Batch:40 Log-loss0.03863152489066124\n",
      "Epoch:20 Batch:80 Log-loss0.027213560417294502\n",
      "Epoch:20 Batch:120 Log-loss0.025858068838715553\n",
      "Epoch:20 Batch:160 Log-loss0.06701960414648056\n",
      "Epoch:20 Batch:200 Log-loss0.05101548507809639\n",
      "Epoch:20 Batch:240 Log-loss0.04590797424316406\n",
      "Epoch:20 Batch:280 Log-loss0.048649996519088745\n",
      "Epoch:20 Batch:320 Log-loss0.03812846168875694\n",
      "Epoch:20 Batch:360 Log-loss0.027038054540753365\n",
      "Epoch:20 Batch:400 Log-loss0.03464942425489426\n",
      "Epoch:20 Batch:440 Log-loss0.052201807498931885\n",
      "Epoch:20 Batch:480 Log-loss0.04024213179945946\n",
      "Epoch:20 Batch:520 Log-loss0.07803493738174438\n",
      "Epoch:20 Batch:560 Log-loss0.03805623948574066\n",
      "Epoch average log loss:0.03815802318443145\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch20, val_loss:0.039732313976872706, best_val_loss:0.039732313976872706, best_auc:0.9879905627652646\n",
      "Epoch:21 Batch:0 Log-loss0.02751695178449154\n",
      "Epoch:21 Batch:40 Log-loss0.036250997334718704\n",
      "Epoch:21 Batch:80 Log-loss0.042129721492528915\n",
      "Epoch:21 Batch:120 Log-loss0.024727890267968178\n",
      "Epoch:21 Batch:160 Log-loss0.03593147173523903\n",
      "Epoch:21 Batch:200 Log-loss0.04855059087276459\n",
      "Epoch:21 Batch:240 Log-loss0.03285956010222435\n",
      "Epoch:21 Batch:280 Log-loss0.03837204724550247\n",
      "Epoch:21 Batch:320 Log-loss0.037229787558317184\n",
      "Epoch:21 Batch:360 Log-loss0.04059554263949394\n",
      "Epoch:21 Batch:400 Log-loss0.050699394196271896\n",
      "Epoch:21 Batch:440 Log-loss0.04214555025100708\n",
      "Epoch:21 Batch:480 Log-loss0.03716372326016426\n",
      "Epoch:21 Batch:520 Log-loss0.037633124738931656\n",
      "Epoch:21 Batch:560 Log-loss0.027339262887835503\n",
      "Epoch average log loss:0.03803560889126467\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch21, val_loss:0.039680714568913304, best_val_loss:0.039680714568913304, best_auc:0.9880802553991476\n",
      "Epoch:22 Batch:0 Log-loss0.020729808136820793\n",
      "Epoch:22 Batch:40 Log-loss0.03156974911689758\n",
      "Epoch:22 Batch:80 Log-loss0.033150654286146164\n",
      "Epoch:22 Batch:120 Log-loss0.05692422017455101\n",
      "Epoch:22 Batch:160 Log-loss0.03398428484797478\n",
      "Epoch:22 Batch:200 Log-loss0.04364881291985512\n",
      "Epoch:22 Batch:240 Log-loss0.04557077959179878\n",
      "Epoch:22 Batch:280 Log-loss0.045920293778181076\n",
      "Epoch:22 Batch:320 Log-loss0.03262825682759285\n",
      "Epoch:22 Batch:360 Log-loss0.03155517578125\n",
      "Epoch:22 Batch:400 Log-loss0.03202701732516289\n",
      "Epoch:22 Batch:440 Log-loss0.031079737469553947\n",
      "Epoch:22 Batch:480 Log-loss0.038790229707956314\n",
      "Epoch:22 Batch:520 Log-loss0.0386899933218956\n",
      "Epoch:22 Batch:560 Log-loss0.03681391105055809\n",
      "Epoch average log loss:0.0379009210604376\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch22, val_loss:0.039645869015900496, best_val_loss:0.039645869015900496, best_auc:0.9881181991965767\n",
      "Epoch:23 Batch:0 Log-loss0.03253383934497833\n",
      "Epoch:23 Batch:40 Log-loss0.026243800297379494\n",
      "Epoch:23 Batch:80 Log-loss0.04276639223098755\n",
      "Epoch:23 Batch:120 Log-loss0.029574429616332054\n",
      "Epoch:23 Batch:160 Log-loss0.039755094796419144\n",
      "Epoch:23 Batch:200 Log-loss0.03959386795759201\n",
      "Epoch:23 Batch:240 Log-loss0.04192901775240898\n",
      "Epoch:23 Batch:280 Log-loss0.04037978872656822\n",
      "Epoch:23 Batch:320 Log-loss0.03438792750239372\n",
      "Epoch:23 Batch:360 Log-loss0.04452192410826683\n",
      "Epoch:23 Batch:400 Log-loss0.04271752014756203\n",
      "Epoch:23 Batch:440 Log-loss0.03423717990517616\n",
      "Epoch:23 Batch:480 Log-loss0.04727497696876526\n",
      "Epoch:23 Batch:520 Log-loss0.05367186293005943\n",
      "Epoch:23 Batch:560 Log-loss0.047507744282484055\n",
      "Epoch average log loss:0.03777715673703434\n",
      "In Epoch23, val_loss:0.03980387114057278, best_val_loss:0.039645869015900496, best_auc:0.9881181991965767\n",
      "Epoch:24 Batch:0 Log-loss0.039453763514757156\n",
      "Epoch:24 Batch:40 Log-loss0.040253184735774994\n",
      "Epoch:24 Batch:80 Log-loss0.02764750085771084\n",
      "Epoch:24 Batch:120 Log-loss0.030891284346580505\n",
      "Epoch:24 Batch:160 Log-loss0.045025527477264404\n",
      "Epoch:24 Batch:200 Log-loss0.031832728534936905\n",
      "Epoch:24 Batch:240 Log-loss0.04106670990586281\n",
      "Epoch:24 Batch:280 Log-loss0.04411778971552849\n",
      "Epoch:24 Batch:320 Log-loss0.025978468358516693\n",
      "Epoch:24 Batch:360 Log-loss0.04039729759097099\n",
      "Epoch:24 Batch:400 Log-loss0.05636836960911751\n",
      "Epoch:24 Batch:440 Log-loss0.03952712193131447\n",
      "Epoch:24 Batch:480 Log-loss0.03713332489132881\n",
      "Epoch:24 Batch:520 Log-loss0.031242722645401955\n",
      "Epoch:24 Batch:560 Log-loss0.02738570235669613\n",
      "Epoch average log loss:0.03772653946653008\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch24, val_loss:0.03960626397149949, best_val_loss:0.03960626397149949, best_auc:0.9882408352973778\n",
      "Epoch:25 Batch:0 Log-loss0.037219006568193436\n",
      "Epoch:25 Batch:40 Log-loss0.04851214587688446\n",
      "Epoch:25 Batch:80 Log-loss0.036449287086725235\n",
      "Epoch:25 Batch:120 Log-loss0.03228975459933281\n",
      "Epoch:25 Batch:160 Log-loss0.023570530116558075\n",
      "Epoch:25 Batch:200 Log-loss0.03904127702116966\n",
      "Epoch:25 Batch:240 Log-loss0.038426656275987625\n",
      "Epoch:25 Batch:280 Log-loss0.03408236429095268\n",
      "Epoch:25 Batch:320 Log-loss0.04715893790125847\n",
      "Epoch:25 Batch:360 Log-loss0.0494580902159214\n",
      "Epoch:25 Batch:400 Log-loss0.06909820437431335\n",
      "Epoch:25 Batch:440 Log-loss0.04380826652050018\n",
      "Epoch:25 Batch:480 Log-loss0.040931057184934616\n",
      "Epoch:25 Batch:520 Log-loss0.042900364845991135\n",
      "Epoch:25 Batch:560 Log-loss0.029865972697734833\n",
      "Epoch average log loss:0.03755537748802453\n",
      "In Epoch25, val_loss:0.03967393706770267, best_val_loss:0.03960626397149949, best_auc:0.9882408352973778\n",
      "Epoch:26 Batch:0 Log-loss0.022886356338858604\n",
      "Epoch:26 Batch:40 Log-loss0.027715453878045082\n",
      "Epoch:26 Batch:80 Log-loss0.05084042251110077\n",
      "Epoch:26 Batch:120 Log-loss0.0427374504506588\n",
      "Epoch:26 Batch:160 Log-loss0.02823389135301113\n",
      "Epoch:26 Batch:200 Log-loss0.059138860553503036\n",
      "Epoch:26 Batch:240 Log-loss0.045324523001909256\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:26 Batch:280 Log-loss0.0324312224984169\n",
      "Epoch:26 Batch:320 Log-loss0.043237343430519104\n",
      "Epoch:26 Batch:360 Log-loss0.048127006739377975\n",
      "Epoch:26 Batch:400 Log-loss0.02048318274319172\n",
      "Epoch:26 Batch:440 Log-loss0.03484311327338219\n",
      "Epoch:26 Batch:480 Log-loss0.04101413115859032\n",
      "Epoch:26 Batch:520 Log-loss0.03416096046566963\n",
      "Epoch:26 Batch:560 Log-loss0.032723695039749146\n",
      "Epoch average log loss:0.03758471753015848\n",
      "In Epoch26, val_loss:0.03973707576121572, best_val_loss:0.03960626397149949, best_auc:0.9882408352973778\n",
      "Epoch:27 Batch:0 Log-loss0.05371787026524544\n",
      "Epoch:27 Batch:40 Log-loss0.04236404970288277\n",
      "Epoch:27 Batch:80 Log-loss0.05403284355998039\n",
      "Epoch:27 Batch:120 Log-loss0.024500088766217232\n",
      "Epoch:27 Batch:160 Log-loss0.042078617960214615\n",
      "Epoch:27 Batch:200 Log-loss0.033170174807310104\n",
      "Epoch:27 Batch:240 Log-loss0.0327983982861042\n",
      "Epoch:27 Batch:280 Log-loss0.04499615356326103\n",
      "Epoch:27 Batch:320 Log-loss0.04755761846899986\n",
      "Epoch:27 Batch:360 Log-loss0.04892247915267944\n",
      "Epoch:27 Batch:400 Log-loss0.02688419260084629\n",
      "Epoch:27 Batch:440 Log-loss0.026782726868987083\n",
      "Epoch:27 Batch:480 Log-loss0.03891800716519356\n",
      "Epoch:27 Batch:520 Log-loss0.028051868081092834\n",
      "Epoch:27 Batch:560 Log-loss0.04939776286482811\n",
      "Epoch average log loss:0.03740522820901658\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch27, val_loss:0.03960134621043647, best_val_loss:0.03960134621043647, best_auc:0.9883165898661361\n",
      "Epoch:28 Batch:0 Log-loss0.03645922616124153\n",
      "Epoch:28 Batch:40 Log-loss0.030812012031674385\n",
      "Epoch:28 Batch:80 Log-loss0.044883597642183304\n",
      "Epoch:28 Batch:120 Log-loss0.030533788725733757\n",
      "Epoch:28 Batch:160 Log-loss0.036540381610393524\n",
      "Epoch:28 Batch:200 Log-loss0.03925995156168938\n",
      "Epoch:28 Batch:240 Log-loss0.039563555270433426\n",
      "Epoch:28 Batch:280 Log-loss0.041989464312791824\n",
      "Epoch:28 Batch:320 Log-loss0.03624729812145233\n",
      "Epoch:28 Batch:360 Log-loss0.041719019412994385\n",
      "Epoch:28 Batch:400 Log-loss0.024957656860351562\n",
      "Epoch:28 Batch:440 Log-loss0.03012007661163807\n",
      "Epoch:28 Batch:480 Log-loss0.0379357673227787\n",
      "Epoch:28 Batch:520 Log-loss0.03738238289952278\n",
      "Epoch:28 Batch:560 Log-loss0.03550500422716141\n",
      "Epoch average log loss:0.03727846016907799\n",
      "In Epoch28, val_loss:0.03990594074043567, best_val_loss:0.03960134621043647, best_auc:0.9883165898661361\n",
      "Epoch:29 Batch:0 Log-loss0.029392793774604797\n",
      "Epoch:29 Batch:40 Log-loss0.032703202217817307\n",
      "Epoch:29 Batch:80 Log-loss0.04210419952869415\n",
      "Epoch:29 Batch:120 Log-loss0.05772899091243744\n",
      "Epoch:29 Batch:160 Log-loss0.0411553718149662\n",
      "Epoch:29 Batch:200 Log-loss0.045656099915504456\n",
      "Epoch:29 Batch:240 Log-loss0.032686855643987656\n",
      "Epoch:29 Batch:280 Log-loss0.043740060180425644\n",
      "Epoch:29 Batch:320 Log-loss0.03586067631840706\n",
      "Epoch:29 Batch:360 Log-loss0.01970842480659485\n",
      "Epoch:29 Batch:400 Log-loss0.027916567400097847\n",
      "Epoch:29 Batch:440 Log-loss0.03849014639854431\n",
      "Epoch:29 Batch:480 Log-loss0.04105723649263382\n",
      "Epoch:29 Batch:520 Log-loss0.03160668537020683\n",
      "Epoch:29 Batch:560 Log-loss0.03714841976761818\n",
      "Epoch average log loss:0.037113293034157585\n",
      "In Epoch29, val_loss:0.03978205163104612, best_val_loss:0.03960134621043647, best_auc:0.9883165898661361\n",
      "Epoch:30 Batch:0 Log-loss0.028491586446762085\n",
      "Epoch:30 Batch:40 Log-loss0.037212613970041275\n",
      "Epoch:30 Batch:80 Log-loss0.04054703563451767\n",
      "Epoch:30 Batch:120 Log-loss0.05247289314866066\n",
      "Epoch:30 Batch:160 Log-loss0.0339026041328907\n",
      "Epoch:30 Batch:200 Log-loss0.04522855952382088\n",
      "Epoch:30 Batch:240 Log-loss0.0437026172876358\n",
      "Epoch:30 Batch:280 Log-loss0.03546847030520439\n",
      "Epoch:30 Batch:320 Log-loss0.03127969801425934\n",
      "Epoch:30 Batch:360 Log-loss0.03710707277059555\n",
      "Epoch:30 Batch:400 Log-loss0.04919816181063652\n",
      "Epoch:30 Batch:440 Log-loss0.03720974549651146\n",
      "Epoch:30 Batch:480 Log-loss0.027146779000759125\n",
      "Epoch:30 Batch:520 Log-loss0.05042945221066475\n",
      "Epoch:30 Batch:560 Log-loss0.037084635347127914\n",
      "Epoch average log loss:0.036993569202188936\n",
      "In Epoch30, val_loss:0.0396852748211211, best_val_loss:0.03960134621043647, best_auc:0.9883165898661361\n",
      "Epoch:31 Batch:0 Log-loss0.0352509580552578\n",
      "Epoch:31 Batch:40 Log-loss0.02780234068632126\n",
      "Epoch:31 Batch:80 Log-loss0.02819417417049408\n",
      "Epoch:31 Batch:120 Log-loss0.041867200285196304\n",
      "Epoch:31 Batch:160 Log-loss0.0319465771317482\n",
      "Epoch:31 Batch:200 Log-loss0.03480900451540947\n",
      "Epoch:31 Batch:240 Log-loss0.03145485371351242\n",
      "Epoch:31 Batch:280 Log-loss0.03588329628109932\n",
      "Epoch:31 Batch:320 Log-loss0.025617724284529686\n",
      "Epoch:31 Batch:360 Log-loss0.024479160085320473\n",
      "Epoch:31 Batch:400 Log-loss0.045395251363515854\n",
      "Epoch:31 Batch:440 Log-loss0.0269883181899786\n",
      "Epoch:31 Batch:480 Log-loss0.03747023269534111\n",
      "Epoch:31 Batch:520 Log-loss0.03710931912064552\n",
      "Epoch:31 Batch:560 Log-loss0.04152093082666397\n",
      "Epoch average log loss:0.03689841014732208\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch31, val_loss:0.039578173508909895, best_val_loss:0.039578173508909895, best_auc:0.9883901954473616\n",
      "Epoch:32 Batch:0 Log-loss0.03411639854311943\n",
      "Epoch:32 Batch:40 Log-loss0.03309960290789604\n",
      "Epoch:32 Batch:80 Log-loss0.0456276535987854\n",
      "Epoch:32 Batch:120 Log-loss0.04481903091073036\n",
      "Epoch:32 Batch:160 Log-loss0.025949561968445778\n",
      "Epoch:32 Batch:200 Log-loss0.034613918513059616\n",
      "Epoch:32 Batch:240 Log-loss0.050724562257528305\n",
      "Epoch:32 Batch:280 Log-loss0.04634535685181618\n",
      "Epoch:32 Batch:320 Log-loss0.03416780009865761\n",
      "Epoch:32 Batch:360 Log-loss0.03739107772707939\n",
      "Epoch:32 Batch:400 Log-loss0.03282976150512695\n",
      "Epoch:32 Batch:440 Log-loss0.044905710965394974\n",
      "Epoch:32 Batch:480 Log-loss0.04262161627411842\n",
      "Epoch:32 Batch:520 Log-loss0.029258335009217262\n",
      "Epoch:32 Batch:560 Log-loss0.03116542287170887\n",
      "Epoch average log loss:0.03708133922025029\n",
      "In Epoch32, val_loss:0.03963406806753134, best_val_loss:0.039578173508909895, best_auc:0.9883901954473616\n",
      "Epoch:33 Batch:0 Log-loss0.04494209215044975\n",
      "Epoch:33 Batch:40 Log-loss0.03806380182504654\n",
      "Epoch:33 Batch:80 Log-loss0.034147776663303375\n",
      "Epoch:33 Batch:120 Log-loss0.04344000294804573\n",
      "Epoch:33 Batch:160 Log-loss0.025720955803990364\n",
      "Epoch:33 Batch:200 Log-loss0.029528195038437843\n",
      "Epoch:33 Batch:240 Log-loss0.035433653742074966\n",
      "Epoch:33 Batch:280 Log-loss0.03545616194605827\n",
      "Epoch:33 Batch:320 Log-loss0.043631989508867264\n",
      "Epoch:33 Batch:360 Log-loss0.030433109030127525\n",
      "Epoch:33 Batch:400 Log-loss0.021857714280486107\n",
      "Epoch:33 Batch:440 Log-loss0.03202652558684349\n",
      "Epoch:33 Batch:480 Log-loss0.02405669540166855\n",
      "Epoch:33 Batch:520 Log-loss0.04182443395256996\n",
      "Epoch:33 Batch:560 Log-loss0.03634663298726082\n",
      "Epoch average log loss:0.03692989288829267\n",
      "In Epoch33, val_loss:0.03971989821728813, best_val_loss:0.039578173508909895, best_auc:0.9883901954473616\n",
      "Epoch:34 Batch:0 Log-loss0.0360104963183403\n",
      "Epoch:34 Batch:40 Log-loss0.03379857912659645\n",
      "Epoch:34 Batch:80 Log-loss0.023510945960879326\n",
      "Epoch:34 Batch:120 Log-loss0.047485750168561935\n",
      "Epoch:34 Batch:160 Log-loss0.037179216742515564\n",
      "Epoch:34 Batch:200 Log-loss0.03883546590805054\n",
      "Epoch:34 Batch:240 Log-loss0.019952071830630302\n",
      "Epoch:34 Batch:280 Log-loss0.03698747232556343\n",
      "Epoch:34 Batch:320 Log-loss0.026430970057845116\n",
      "Epoch:34 Batch:360 Log-loss0.03882357105612755\n",
      "Epoch:34 Batch:400 Log-loss0.028398355469107628\n",
      "Epoch:34 Batch:440 Log-loss0.03441137447953224\n",
      "Epoch:34 Batch:480 Log-loss0.04402099922299385\n",
      "Epoch:34 Batch:520 Log-loss0.025963373482227325\n",
      "Epoch:34 Batch:560 Log-loss0.041843634098768234\n",
      "Epoch average log loss:0.03684822796910469\n",
      "In Epoch34, val_loss:0.03958789865540842, best_val_loss:0.039578173508909895, best_auc:0.9883901954473616\n",
      "Epoch:35 Batch:0 Log-loss0.03882332146167755\n",
      "Epoch:35 Batch:40 Log-loss0.03920610621571541\n",
      "Epoch:35 Batch:80 Log-loss0.02557436190545559\n",
      "Epoch:35 Batch:120 Log-loss0.048151370137929916\n",
      "Epoch:35 Batch:160 Log-loss0.02730034478008747\n",
      "Epoch:35 Batch:200 Log-loss0.0547710657119751\n",
      "Epoch:35 Batch:240 Log-loss0.0366642028093338\n",
      "Epoch:35 Batch:280 Log-loss0.0384109690785408\n",
      "Epoch:35 Batch:320 Log-loss0.03890584036707878\n",
      "Epoch:35 Batch:360 Log-loss0.05087042972445488\n",
      "Epoch:35 Batch:400 Log-loss0.030623475089669228\n",
      "Epoch:35 Batch:440 Log-loss0.039654795080423355\n",
      "Epoch:35 Batch:480 Log-loss0.047837670892477036\n",
      "Epoch:35 Batch:520 Log-loss0.024006620049476624\n",
      "Epoch:35 Batch:560 Log-loss0.034968625754117966\n",
      "Epoch average log loss:0.036756075680854594\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In Epoch35, val_loss:0.039657340306191564, best_val_loss:0.039578173508909895, best_auc:0.9883901954473616\n",
      "Epoch:36 Batch:0 Log-loss0.028388721868395805\n",
      "Epoch:36 Batch:40 Log-loss0.027466779574751854\n",
      "Epoch:36 Batch:80 Log-loss0.02881482057273388\n",
      "Epoch:36 Batch:120 Log-loss0.037767112255096436\n",
      "Epoch:36 Batch:160 Log-loss0.028785988688468933\n",
      "Epoch:36 Batch:200 Log-loss0.033389244228601456\n",
      "Epoch:36 Batch:240 Log-loss0.022862069308757782\n",
      "Epoch:36 Batch:280 Log-loss0.02861057035624981\n",
      "Epoch:36 Batch:320 Log-loss0.03138485178351402\n",
      "Epoch:36 Batch:360 Log-loss0.03939642012119293\n",
      "Epoch:36 Batch:400 Log-loss0.026068901643157005\n",
      "Epoch:36 Batch:440 Log-loss0.0366431325674057\n",
      "Epoch:36 Batch:480 Log-loss0.02163195051252842\n",
      "Epoch:36 Batch:520 Log-loss0.02748493105173111\n",
      "Epoch:36 Batch:560 Log-loss0.04909767210483551\n",
      "Epoch average log loss:0.036708988179452715\n",
      "In Epoch36, val_loss:0.03967144006626593, best_val_loss:0.039578173508909895, best_auc:0.9883901954473616\n",
      "Epoch:37 Batch:0 Log-loss0.036945242434740067\n",
      "Epoch:37 Batch:40 Log-loss0.03829910606145859\n",
      "Epoch:37 Batch:80 Log-loss0.04132379963994026\n",
      "Epoch:37 Batch:120 Log-loss0.03258271887898445\n",
      "Epoch:37 Batch:160 Log-loss0.023665329441428185\n",
      "Epoch:37 Batch:200 Log-loss0.03260662779211998\n",
      "Epoch:37 Batch:240 Log-loss0.02864997275173664\n",
      "Epoch:37 Batch:280 Log-loss0.03468349203467369\n",
      "Epoch:37 Batch:320 Log-loss0.043295010924339294\n",
      "Epoch:37 Batch:360 Log-loss0.02686449885368347\n",
      "Epoch:37 Batch:400 Log-loss0.022971903905272484\n",
      "Epoch:37 Batch:440 Log-loss0.03620908409357071\n",
      "Epoch:37 Batch:480 Log-loss0.040585894137620926\n",
      "Epoch:37 Batch:520 Log-loss0.035035643726587296\n",
      "Epoch:37 Batch:560 Log-loss0.026100967079401016\n",
      "Epoch average log loss:0.03681612725236586\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch37, val_loss:0.039567854078721126, best_val_loss:0.039567854078721126, best_auc:0.9884251471091594\n",
      "Epoch:38 Batch:0 Log-loss0.03576769307255745\n",
      "Epoch:38 Batch:40 Log-loss0.03483723849058151\n",
      "Epoch:38 Batch:80 Log-loss0.0425143800675869\n",
      "Epoch:38 Batch:120 Log-loss0.031606730073690414\n",
      "Epoch:38 Batch:160 Log-loss0.03160088509321213\n",
      "Epoch:38 Batch:200 Log-loss0.04148406162858009\n",
      "Epoch:38 Batch:240 Log-loss0.04098685458302498\n",
      "Epoch:38 Batch:280 Log-loss0.055730536580085754\n",
      "Epoch:38 Batch:320 Log-loss0.03296936675906181\n",
      "Epoch:38 Batch:360 Log-loss0.03172796592116356\n",
      "Epoch:38 Batch:400 Log-loss0.03662976250052452\n",
      "Epoch:38 Batch:440 Log-loss0.04056026414036751\n",
      "Epoch:38 Batch:480 Log-loss0.036201078444719315\n",
      "Epoch:38 Batch:520 Log-loss0.02347760647535324\n",
      "Epoch:38 Batch:560 Log-loss0.03868301957845688\n",
      "Epoch average log loss:0.0367151590962229\n",
      "In Epoch38, val_loss:0.03969147227616679, best_val_loss:0.039567854078721126, best_auc:0.9884251471091594\n",
      "Epoch:39 Batch:0 Log-loss0.04495902732014656\n",
      "Epoch:39 Batch:40 Log-loss0.03015611506998539\n",
      "Epoch:39 Batch:80 Log-loss0.034349165856838226\n",
      "Epoch:39 Batch:120 Log-loss0.03797387704253197\n",
      "Epoch:39 Batch:160 Log-loss0.042839761823415756\n",
      "Epoch:39 Batch:200 Log-loss0.028172120451927185\n",
      "Epoch:39 Batch:240 Log-loss0.047989483922719955\n",
      "Epoch:39 Batch:280 Log-loss0.02811065874993801\n",
      "Epoch:39 Batch:320 Log-loss0.023562803864479065\n",
      "Epoch:39 Batch:360 Log-loss0.06007622182369232\n",
      "Epoch:39 Batch:400 Log-loss0.04337894916534424\n",
      "Epoch:39 Batch:440 Log-loss0.040629513561725616\n",
      "Epoch:39 Batch:480 Log-loss0.037891488522291183\n",
      "Epoch:39 Batch:520 Log-loss0.041670020669698715\n",
      "Epoch:39 Batch:560 Log-loss0.03358212485909462\n",
      "Epoch average log loss:0.03666330280414384\n",
      "In Epoch39, val_loss:0.03962571629755927, best_val_loss:0.039567854078721126, best_auc:0.9884251471091594\n",
      "Epoch:40 Batch:0 Log-loss0.040778450667858124\n",
      "Epoch:40 Batch:40 Log-loss0.04027805104851723\n",
      "Epoch:40 Batch:80 Log-loss0.030439818277955055\n",
      "Epoch:40 Batch:120 Log-loss0.03421783447265625\n",
      "Epoch:40 Batch:160 Log-loss0.031219908967614174\n",
      "Epoch:40 Batch:200 Log-loss0.028730368241667747\n",
      "Epoch:40 Batch:240 Log-loss0.026360981166362762\n",
      "Epoch:40 Batch:280 Log-loss0.036209385842084885\n",
      "Epoch:40 Batch:320 Log-loss0.026824943721294403\n",
      "Epoch:40 Batch:360 Log-loss0.025348633527755737\n",
      "Epoch:40 Batch:400 Log-loss0.029025793075561523\n",
      "Epoch:40 Batch:440 Log-loss0.04022575542330742\n",
      "Epoch:40 Batch:480 Log-loss0.0434846393764019\n",
      "Epoch:40 Batch:520 Log-loss0.03601204231381416\n",
      "Epoch:40 Batch:560 Log-loss0.03356168046593666\n",
      "Epoch average log loss:0.03647519479106579\n",
      "In Epoch40, val_loss:0.03978306565425841, best_val_loss:0.039567854078721126, best_auc:0.9884251471091594\n",
      "Epoch:41 Batch:0 Log-loss0.04386116564273834\n",
      "Epoch:41 Batch:40 Log-loss0.03570844233036041\n",
      "Epoch:41 Batch:80 Log-loss0.032790880650281906\n",
      "Epoch:41 Batch:120 Log-loss0.039493758231401443\n",
      "Epoch:41 Batch:160 Log-loss0.03291763737797737\n",
      "Epoch:41 Batch:200 Log-loss0.038496632128953934\n",
      "Epoch:41 Batch:240 Log-loss0.03715769574046135\n",
      "Epoch:41 Batch:280 Log-loss0.02878645621240139\n",
      "Epoch:41 Batch:320 Log-loss0.034670036286115646\n",
      "Epoch:41 Batch:360 Log-loss0.03277970850467682\n",
      "Epoch:41 Batch:400 Log-loss0.04071936383843422\n",
      "Epoch:41 Batch:440 Log-loss0.03852614015340805\n",
      "Epoch:41 Batch:480 Log-loss0.029861070215702057\n",
      "Epoch:41 Batch:520 Log-loss0.030032366514205933\n",
      "Epoch:41 Batch:560 Log-loss0.037909846752882004\n",
      "Epoch average log loss:0.036634447518736124\n",
      "In Epoch41, val_loss:0.03968970581094172, best_val_loss:0.039567854078721126, best_auc:0.9884251471091594\n",
      "Epoch:42 Batch:0 Log-loss0.041332703083753586\n",
      "Epoch:42 Batch:40 Log-loss0.03275677189230919\n",
      "Epoch:42 Batch:80 Log-loss0.03983212262392044\n",
      "Epoch:42 Batch:120 Log-loss0.04043218120932579\n",
      "Epoch:42 Batch:160 Log-loss0.03798113018274307\n",
      "Epoch:42 Batch:200 Log-loss0.029367772862315178\n",
      "Epoch:42 Batch:240 Log-loss0.04331676661968231\n",
      "Epoch:42 Batch:280 Log-loss0.024164512753486633\n",
      "Epoch:42 Batch:320 Log-loss0.046287160366773605\n",
      "Epoch:42 Batch:360 Log-loss0.039953771978616714\n",
      "Epoch:42 Batch:400 Log-loss0.027996575459837914\n",
      "Epoch:42 Batch:440 Log-loss0.030333230271935463\n",
      "Epoch:42 Batch:480 Log-loss0.03549046814441681\n",
      "Epoch:42 Batch:520 Log-loss0.03523653373122215\n",
      "Epoch:42 Batch:560 Log-loss0.037862442433834076\n",
      "Epoch average log loss:0.03651816999273641\n",
      "In Epoch42, val_loss:0.03967855995814321, best_val_loss:0.039567854078721126, best_auc:0.9884251471091594\n",
      "Epoch:43 Batch:0 Log-loss0.048378732055425644\n",
      "Epoch:43 Batch:40 Log-loss0.032483868300914764\n",
      "Epoch:43 Batch:80 Log-loss0.030908865854144096\n",
      "Epoch:43 Batch:120 Log-loss0.03687702864408493\n",
      "Epoch:43 Batch:160 Log-loss0.03289490565657616\n",
      "Epoch:43 Batch:200 Log-loss0.04485132917761803\n",
      "Epoch:43 Batch:240 Log-loss0.037242207676172256\n",
      "Epoch:43 Batch:280 Log-loss0.02870124578475952\n",
      "Epoch:43 Batch:320 Log-loss0.04074305295944214\n",
      "Epoch:43 Batch:360 Log-loss0.03637842461466789\n",
      "Epoch:43 Batch:400 Log-loss0.0326099656522274\n",
      "Epoch:43 Batch:440 Log-loss0.043273117393255234\n",
      "Epoch:43 Batch:480 Log-loss0.021025486290454865\n",
      "Epoch:43 Batch:520 Log-loss0.046953726559877396\n",
      "Epoch:43 Batch:560 Log-loss0.030827239155769348\n",
      "Epoch average log loss:0.03642876544035971\n",
      "In Epoch43, val_loss:0.03960306198645181, best_val_loss:0.039567854078721126, best_auc:0.9884251471091594\n",
      "Epoch:44 Batch:0 Log-loss0.03223938122391701\n",
      "Epoch:44 Batch:40 Log-loss0.040878232568502426\n",
      "Epoch:44 Batch:80 Log-loss0.03631209582090378\n",
      "Epoch:44 Batch:120 Log-loss0.02289118617773056\n",
      "Epoch:44 Batch:160 Log-loss0.024861963465809822\n",
      "Epoch:44 Batch:200 Log-loss0.031360283493995667\n",
      "Epoch:44 Batch:240 Log-loss0.041193995624780655\n",
      "Epoch:44 Batch:280 Log-loss0.039613161236047745\n",
      "Epoch:44 Batch:320 Log-loss0.042600709944963455\n",
      "Epoch:44 Batch:360 Log-loss0.034420520067214966\n",
      "Epoch:44 Batch:400 Log-loss0.04766956344246864\n",
      "Epoch:44 Batch:440 Log-loss0.045523833483457565\n",
      "Epoch:44 Batch:480 Log-loss0.03142665699124336\n",
      "Epoch:44 Batch:520 Log-loss0.027826523408293724\n",
      "Epoch:44 Batch:560 Log-loss0.04660638049244881\n",
      "Epoch average log loss:0.036441624307605835\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch44, val_loss:0.039563648721861154, best_val_loss:0.039563648721861154, best_auc:0.988465415625026\n",
      "Epoch:45 Batch:0 Log-loss0.029030373319983482\n",
      "Epoch:45 Batch:40 Log-loss0.04030758515000343\n",
      "Epoch:45 Batch:80 Log-loss0.03198248893022537\n",
      "Epoch:45 Batch:120 Log-loss0.036749500781297684\n",
      "Epoch:45 Batch:160 Log-loss0.030187224969267845\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:45 Batch:200 Log-loss0.037846554070711136\n",
      "Epoch:45 Batch:240 Log-loss0.04242054000496864\n",
      "Epoch:45 Batch:280 Log-loss0.05488096550107002\n",
      "Epoch:45 Batch:320 Log-loss0.026418335735797882\n",
      "Epoch:45 Batch:360 Log-loss0.03647127375006676\n",
      "Epoch:45 Batch:400 Log-loss0.0363917276263237\n",
      "Epoch:45 Batch:440 Log-loss0.03109787404537201\n",
      "Epoch:45 Batch:480 Log-loss0.036177489906549454\n",
      "Epoch:45 Batch:520 Log-loss0.033712368458509445\n",
      "Epoch:45 Batch:560 Log-loss0.030969778075814247\n",
      "Epoch average log loss:0.0364347140737144\n",
      "In Epoch45, val_loss:0.03958566948339255, best_val_loss:0.039563648721861154, best_auc:0.988465415625026\n",
      "Epoch:46 Batch:0 Log-loss0.03161366656422615\n",
      "Epoch:46 Batch:40 Log-loss0.04283595085144043\n",
      "Epoch:46 Batch:80 Log-loss0.041522856801748276\n",
      "Epoch:46 Batch:120 Log-loss0.043729234486818314\n",
      "Epoch:46 Batch:160 Log-loss0.02303955890238285\n",
      "Epoch:46 Batch:200 Log-loss0.03180466964840889\n",
      "Epoch:46 Batch:240 Log-loss0.040012892335653305\n",
      "Epoch:46 Batch:280 Log-loss0.05044377222657204\n",
      "Epoch:46 Batch:320 Log-loss0.036771390587091446\n",
      "Epoch:46 Batch:360 Log-loss0.029295019805431366\n",
      "Epoch:46 Batch:400 Log-loss0.046184640377759933\n",
      "Epoch:46 Batch:440 Log-loss0.04032221809029579\n",
      "Epoch:46 Batch:480 Log-loss0.0366017185151577\n",
      "Epoch:46 Batch:520 Log-loss0.03202950581908226\n",
      "Epoch:46 Batch:560 Log-loss0.03275755047798157\n",
      "Epoch average log loss:0.036467682014751646\n",
      "In Epoch46, val_loss:0.039624142861368804, best_val_loss:0.039563648721861154, best_auc:0.988465415625026\n",
      "Epoch:47 Batch:0 Log-loss0.02722298353910446\n",
      "Epoch:47 Batch:40 Log-loss0.03700931742787361\n",
      "Epoch:47 Batch:80 Log-loss0.03180548921227455\n",
      "Epoch:47 Batch:120 Log-loss0.01939813606441021\n",
      "Epoch:47 Batch:160 Log-loss0.031441524624824524\n",
      "Epoch:47 Batch:200 Log-loss0.026491845026612282\n",
      "Epoch:47 Batch:240 Log-loss0.03682893142104149\n",
      "Epoch:47 Batch:280 Log-loss0.0264647975564003\n",
      "Epoch:47 Batch:320 Log-loss0.043025802820920944\n",
      "Epoch:47 Batch:360 Log-loss0.028972236439585686\n",
      "Epoch:47 Batch:400 Log-loss0.03352978825569153\n",
      "Epoch:47 Batch:440 Log-loss0.03353860601782799\n",
      "Epoch:47 Batch:480 Log-loss0.03929548338055611\n",
      "Epoch:47 Batch:520 Log-loss0.0329405702650547\n",
      "Epoch:47 Batch:560 Log-loss0.027537818998098373\n",
      "Epoch average log loss:0.03646680074078696\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch47, val_loss:0.03956095000325142, best_val_loss:0.03956095000325142, best_auc:0.9884330467037216\n",
      "Epoch:48 Batch:0 Log-loss0.037353720515966415\n",
      "Epoch:48 Batch:40 Log-loss0.03465961292386055\n",
      "Epoch:48 Batch:80 Log-loss0.03440623730421066\n",
      "Epoch:48 Batch:120 Log-loss0.026437079533934593\n",
      "Epoch:48 Batch:160 Log-loss0.028249375522136688\n",
      "Epoch:48 Batch:200 Log-loss0.03738517686724663\n",
      "Epoch:48 Batch:240 Log-loss0.0314459465444088\n",
      "Epoch:48 Batch:280 Log-loss0.04266664385795593\n",
      "Epoch:48 Batch:320 Log-loss0.034039516001939774\n",
      "Epoch:48 Batch:360 Log-loss0.03789418563246727\n",
      "Epoch:48 Batch:400 Log-loss0.030169740319252014\n",
      "Epoch:48 Batch:440 Log-loss0.045850928872823715\n",
      "Epoch:48 Batch:480 Log-loss0.05497143790125847\n",
      "Epoch:48 Batch:520 Log-loss0.031286440789699554\n",
      "Epoch:48 Batch:560 Log-loss0.01694652996957302\n",
      "Epoch average log loss:0.03658032953473074\n",
      "In Epoch48, val_loss:0.039561594451648065, best_val_loss:0.03956095000325142, best_auc:0.9884330467037216\n",
      "Epoch:49 Batch:0 Log-loss0.05123318359255791\n",
      "Epoch:49 Batch:40 Log-loss0.040734220296144485\n",
      "Epoch:49 Batch:80 Log-loss0.0383751280605793\n",
      "Epoch:49 Batch:120 Log-loss0.0541226863861084\n",
      "Epoch:49 Batch:160 Log-loss0.03996769338846207\n",
      "Epoch:49 Batch:200 Log-loss0.03293861076235771\n",
      "Epoch:49 Batch:240 Log-loss0.03026244230568409\n",
      "Epoch:49 Batch:280 Log-loss0.04248392581939697\n",
      "Epoch:49 Batch:320 Log-loss0.03183959051966667\n",
      "Epoch:49 Batch:360 Log-loss0.04382992908358574\n",
      "Epoch:49 Batch:400 Log-loss0.044081758707761765\n",
      "Epoch:49 Batch:440 Log-loss0.0389932282269001\n",
      "Epoch:49 Batch:480 Log-loss0.03158693388104439\n",
      "Epoch:49 Batch:520 Log-loss0.04761933162808418\n",
      "Epoch:49 Batch:560 Log-loss0.02867039106786251\n",
      "Epoch average log loss:0.03645635412207671\n",
      "In Epoch49, val_loss:0.039644738674706394, best_val_loss:0.03956095000325142, best_auc:0.9884330467037216\n",
      "Epoch:50 Batch:0 Log-loss0.04143865779042244\n",
      "Epoch:50 Batch:40 Log-loss0.04156620055437088\n",
      "Epoch:50 Batch:80 Log-loss0.03555138036608696\n",
      "Epoch:50 Batch:120 Log-loss0.028686007484793663\n",
      "Epoch:50 Batch:160 Log-loss0.023911377415060997\n",
      "Epoch:50 Batch:200 Log-loss0.044528961181640625\n",
      "Epoch:50 Batch:240 Log-loss0.039073165506124496\n",
      "Epoch:50 Batch:280 Log-loss0.041509877890348434\n",
      "Epoch:50 Batch:320 Log-loss0.048391640186309814\n",
      "Epoch:50 Batch:360 Log-loss0.027509821578860283\n",
      "Epoch:50 Batch:400 Log-loss0.0327010415494442\n",
      "Epoch:50 Batch:440 Log-loss0.030252190306782722\n",
      "Epoch:50 Batch:480 Log-loss0.02912786602973938\n",
      "Epoch:50 Batch:520 Log-loss0.039222683757543564\n",
      "Epoch:50 Batch:560 Log-loss0.031026963144540787\n",
      "Epoch average log loss:0.03648092790972442\n",
      "In Epoch50, val_loss:0.039634098974457174, best_val_loss:0.03956095000325142, best_auc:0.9884330467037216\n",
      "Epoch:51 Batch:0 Log-loss0.033813200891017914\n",
      "Epoch:51 Batch:40 Log-loss0.03414024040102959\n",
      "Epoch:51 Batch:80 Log-loss0.03195153549313545\n",
      "Epoch:51 Batch:120 Log-loss0.038192860782146454\n",
      "Epoch:51 Batch:160 Log-loss0.038095783442258835\n",
      "Epoch:51 Batch:200 Log-loss0.027634769678115845\n",
      "Epoch:51 Batch:240 Log-loss0.032201673835515976\n",
      "Epoch:51 Batch:280 Log-loss0.03764203563332558\n",
      "Epoch:51 Batch:320 Log-loss0.039081163704395294\n",
      "Epoch:51 Batch:360 Log-loss0.029290853068232536\n",
      "Epoch:51 Batch:400 Log-loss0.02932385541498661\n",
      "Epoch:51 Batch:440 Log-loss0.029915517196059227\n",
      "Epoch:51 Batch:480 Log-loss0.03053434006869793\n",
      "Epoch:51 Batch:520 Log-loss0.03729807958006859\n",
      "Epoch:51 Batch:560 Log-loss0.042854756116867065\n",
      "Epoch average log loss:0.036391033968954746\n",
      "In Epoch51, val_loss:0.03956871569261685, best_val_loss:0.03956095000325142, best_auc:0.9884330467037216\n",
      "Epoch:52 Batch:0 Log-loss0.02880987338721752\n",
      "Epoch:52 Batch:40 Log-loss0.03038603998720646\n",
      "Epoch:52 Batch:80 Log-loss0.04193994402885437\n",
      "Epoch:52 Batch:120 Log-loss0.04210266098380089\n",
      "Epoch:52 Batch:160 Log-loss0.03865605220198631\n",
      "Epoch:52 Batch:200 Log-loss0.03784370794892311\n",
      "Epoch:52 Batch:240 Log-loss0.0456414557993412\n",
      "Epoch:52 Batch:280 Log-loss0.03961498290300369\n",
      "Epoch:52 Batch:320 Log-loss0.03422043099999428\n",
      "Epoch:52 Batch:360 Log-loss0.029734568670392036\n",
      "Epoch:52 Batch:400 Log-loss0.04641001299023628\n",
      "Epoch:52 Batch:440 Log-loss0.03880292922258377\n",
      "Epoch:52 Batch:480 Log-loss0.02789035625755787\n",
      "Epoch:52 Batch:520 Log-loss0.026914849877357483\n",
      "Epoch:52 Batch:560 Log-loss0.03900764137506485\n",
      "Epoch average log loss:0.03623008021885263\n",
      "In Epoch52, val_loss:0.03958178511623155, best_val_loss:0.03956095000325142, best_auc:0.9884330467037216\n",
      "Epoch:53 Batch:0 Log-loss0.029366254806518555\n",
      "Epoch:53 Batch:40 Log-loss0.03728725388646126\n",
      "Epoch:53 Batch:80 Log-loss0.03449774533510208\n",
      "Epoch:53 Batch:120 Log-loss0.03879328817129135\n",
      "Epoch:53 Batch:160 Log-loss0.023870842531323433\n",
      "Epoch:53 Batch:200 Log-loss0.02200845070183277\n",
      "Epoch:53 Batch:240 Log-loss0.0462050586938858\n",
      "Epoch:53 Batch:280 Log-loss0.041480910032987595\n",
      "Epoch:53 Batch:320 Log-loss0.03465253487229347\n",
      "Epoch:53 Batch:360 Log-loss0.024618564173579216\n",
      "Epoch:53 Batch:400 Log-loss0.025965595617890358\n",
      "Epoch:53 Batch:440 Log-loss0.03289501368999481\n",
      "Epoch:53 Batch:480 Log-loss0.03985114023089409\n",
      "Epoch:53 Batch:520 Log-loss0.035206686705350876\n",
      "Epoch:53 Batch:560 Log-loss0.045716237276792526\n",
      "Epoch average log loss:0.036425044575506556\n",
      "In Epoch53, val_loss:0.03962748899273338, best_val_loss:0.03956095000325142, best_auc:0.9884330467037216\n",
      "Epoch:54 Batch:0 Log-loss0.04876911640167236\n",
      "Epoch:54 Batch:40 Log-loss0.028766416013240814\n",
      "Epoch:54 Batch:80 Log-loss0.029212841764092445\n",
      "Epoch:54 Batch:120 Log-loss0.03744111582636833\n",
      "Epoch:54 Batch:160 Log-loss0.04126197472214699\n",
      "Epoch:54 Batch:200 Log-loss0.0315641351044178\n",
      "Epoch:54 Batch:240 Log-loss0.036402542144060135\n",
      "Epoch:54 Batch:280 Log-loss0.02957998216152191\n",
      "Epoch:54 Batch:320 Log-loss0.02792751044034958\n",
      "Epoch:54 Batch:360 Log-loss0.03832190856337547\n",
      "Epoch:54 Batch:400 Log-loss0.0305729191750288\n",
      "Epoch:54 Batch:440 Log-loss0.03349757567048073\n",
      "Epoch:54 Batch:480 Log-loss0.02767903544008732\n",
      "Epoch:54 Batch:520 Log-loss0.032382477074861526\n",
      "Epoch:54 Batch:560 Log-loss0.029742836952209473\n",
      "Epoch average log loss:0.036431567091494796\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In Epoch54, val_loss:0.039624828676775324, best_val_loss:0.03956095000325142, best_auc:0.9884330467037216\n",
      "Epoch:55 Batch:0 Log-loss0.034983109682798386\n",
      "Epoch:55 Batch:40 Log-loss0.043533142656087875\n",
      "Epoch:55 Batch:80 Log-loss0.035091809928417206\n",
      "Epoch:55 Batch:120 Log-loss0.05586490035057068\n",
      "Epoch:55 Batch:160 Log-loss0.034371551126241684\n",
      "Epoch:55 Batch:200 Log-loss0.020768629387021065\n",
      "Epoch:55 Batch:240 Log-loss0.044510677456855774\n",
      "Epoch:55 Batch:280 Log-loss0.046813011169433594\n",
      "Epoch:55 Batch:320 Log-loss0.03270531818270683\n",
      "Epoch:55 Batch:360 Log-loss0.04032721370458603\n",
      "Epoch:55 Batch:400 Log-loss0.04183114692568779\n",
      "Epoch:55 Batch:440 Log-loss0.028339119628071785\n",
      "Epoch:55 Batch:480 Log-loss0.040713679045438766\n",
      "Epoch:55 Batch:520 Log-loss0.03994131460785866\n",
      "Epoch:55 Batch:560 Log-loss0.03601524233818054\n",
      "Epoch average log loss:0.0363054969580844\n",
      "In Epoch55, val_loss:0.039604770953555024, best_val_loss:0.03956095000325142, best_auc:0.9884330467037216\n",
      "Epoch:56 Batch:0 Log-loss0.03473121300339699\n",
      "Epoch:56 Batch:40 Log-loss0.03990801423788071\n",
      "Epoch:56 Batch:80 Log-loss0.029676614329218864\n",
      "Epoch:56 Batch:120 Log-loss0.04154672846198082\n",
      "Epoch:56 Batch:160 Log-loss0.03258296474814415\n",
      "Epoch:56 Batch:200 Log-loss0.038446854799985886\n",
      "Epoch:56 Batch:240 Log-loss0.02859853394329548\n",
      "Epoch:56 Batch:280 Log-loss0.016661785542964935\n",
      "Epoch:56 Batch:320 Log-loss0.03533197566866875\n",
      "Epoch:56 Batch:360 Log-loss0.029840776696801186\n",
      "Epoch:56 Batch:400 Log-loss0.02665075846016407\n",
      "Epoch:56 Batch:440 Log-loss0.04492919147014618\n",
      "Epoch:56 Batch:480 Log-loss0.02536877989768982\n",
      "Epoch:56 Batch:520 Log-loss0.03511262685060501\n",
      "Epoch:56 Batch:560 Log-loss0.03967937082052231\n",
      "Epoch average log loss:0.03633974986816091\n",
      "In Epoch56, val_loss:0.03959533814612213, best_val_loss:0.03956095000325142, best_auc:0.9884330467037216\n",
      "Epoch:57 Batch:0 Log-loss0.04429575800895691\n",
      "Epoch:57 Batch:40 Log-loss0.021045679226517677\n",
      "Epoch:57 Batch:80 Log-loss0.039163410663604736\n",
      "Epoch:57 Batch:120 Log-loss0.037084709852933884\n",
      "Epoch:57 Batch:160 Log-loss0.04174540936946869\n",
      "Epoch:57 Batch:200 Log-loss0.03381502628326416\n",
      "Epoch:57 Batch:240 Log-loss0.05592810735106468\n",
      "Epoch:57 Batch:280 Log-loss0.032590314745903015\n",
      "Epoch:57 Batch:320 Log-loss0.04578182473778725\n",
      "Epoch:57 Batch:360 Log-loss0.05157661810517311\n",
      "Epoch:57 Batch:400 Log-loss0.04479190334677696\n",
      "Epoch:57 Batch:440 Log-loss0.05178583785891533\n",
      "Epoch:57 Batch:480 Log-loss0.041662368923425674\n",
      "Epoch:57 Batch:520 Log-loss0.033139750361442566\n",
      "Epoch:57 Batch:560 Log-loss0.034645430743694305\n",
      "Epoch average log loss:0.03632938542203712\n",
      "A pre-trained model at temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt has been loaded.\n",
      "\n",
      "Model has been saved as checkpoints/FASTTXT_BGRU_64_643.pt.\n",
      "\n",
      "Choose the torch base model.\n",
      "Training on fold 4\n",
      "Epoch:1 Batch:0 Log-loss0.659645140171051\n",
      "Epoch:1 Batch:40 Log-loss0.14446361362934113\n",
      "Epoch:1 Batch:80 Log-loss0.07998961210250854\n",
      "Epoch:1 Batch:120 Log-loss0.06288107484579086\n",
      "Epoch:1 Batch:160 Log-loss0.0726940706372261\n",
      "Epoch:1 Batch:200 Log-loss0.032135169953107834\n",
      "Epoch:1 Batch:240 Log-loss0.06576009839773178\n",
      "Epoch:1 Batch:280 Log-loss0.054645370692014694\n",
      "Epoch:1 Batch:320 Log-loss0.05584162846207619\n",
      "Epoch:1 Batch:360 Log-loss0.0548771508038044\n",
      "Epoch:1 Batch:400 Log-loss0.08226684480905533\n",
      "Epoch:1 Batch:440 Log-loss0.052752535790205\n",
      "Epoch:1 Batch:480 Log-loss0.0481075793504715\n",
      "Epoch:1 Batch:520 Log-loss0.04045679047703743\n",
      "Epoch:1 Batch:560 Log-loss0.059119533747434616\n",
      "Epoch average log loss:0.07760033625444131\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch1, val_loss:0.0468759781520754, best_val_loss:0.0468759781520754, best_auc:0.9767408597579266\n",
      "Epoch:2 Batch:0 Log-loss0.031133653596043587\n",
      "Epoch:2 Batch:40 Log-loss0.04699896648526192\n",
      "Epoch:2 Batch:80 Log-loss0.03741384670138359\n",
      "Epoch:2 Batch:120 Log-loss0.044730160385370255\n",
      "Epoch:2 Batch:160 Log-loss0.05451848730444908\n",
      "Epoch:2 Batch:200 Log-loss0.0463687889277935\n",
      "Epoch:2 Batch:240 Log-loss0.04603612795472145\n",
      "Epoch:2 Batch:280 Log-loss0.05511823296546936\n",
      "Epoch:2 Batch:320 Log-loss0.04783882200717926\n",
      "Epoch:2 Batch:360 Log-loss0.03137022256851196\n",
      "Epoch:2 Batch:400 Log-loss0.04321292042732239\n",
      "Epoch:2 Batch:440 Log-loss0.04614067077636719\n",
      "Epoch:2 Batch:480 Log-loss0.04445633664727211\n",
      "Epoch:2 Batch:520 Log-loss0.054650768637657166\n",
      "Epoch:2 Batch:560 Log-loss0.04607129469513893\n",
      "Epoch average log loss:0.049633930793165096\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch2, val_loss:0.043965016723217225, best_val_loss:0.043965016723217225, best_auc:0.9814274896902561\n",
      "Epoch:3 Batch:0 Log-loss0.045150671154260635\n",
      "Epoch:3 Batch:40 Log-loss0.05867539718747139\n",
      "Epoch:3 Batch:80 Log-loss0.08334663510322571\n",
      "Epoch:3 Batch:120 Log-loss0.03849736973643303\n",
      "Epoch:3 Batch:160 Log-loss0.04488544538617134\n",
      "Epoch:3 Batch:200 Log-loss0.04842612147331238\n",
      "Epoch:3 Batch:240 Log-loss0.0323815643787384\n",
      "Epoch:3 Batch:280 Log-loss0.053192779421806335\n",
      "Epoch:3 Batch:320 Log-loss0.0422791950404644\n",
      "Epoch:3 Batch:360 Log-loss0.04895016923546791\n",
      "Epoch:3 Batch:400 Log-loss0.042557816952466965\n",
      "Epoch:3 Batch:440 Log-loss0.04408751428127289\n",
      "Epoch:3 Batch:480 Log-loss0.051232580095529556\n",
      "Epoch:3 Batch:520 Log-loss0.03482208773493767\n",
      "Epoch:3 Batch:560 Log-loss0.04403195530176163\n",
      "Epoch average log loss:0.04718414786205228\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch3, val_loss:0.04258950750412499, best_val_loss:0.04258950750412499, best_auc:0.9846464160546698\n",
      "Epoch:4 Batch:0 Log-loss0.0603729747235775\n",
      "Epoch:4 Batch:40 Log-loss0.046958718448877335\n",
      "Epoch:4 Batch:80 Log-loss0.029989754781126976\n",
      "Epoch:4 Batch:120 Log-loss0.044527579098939896\n",
      "Epoch:4 Batch:160 Log-loss0.02247834950685501\n",
      "Epoch:4 Batch:200 Log-loss0.043032631278038025\n",
      "Epoch:4 Batch:240 Log-loss0.04395842179656029\n",
      "Epoch:4 Batch:280 Log-loss0.052530437707901\n",
      "Epoch:4 Batch:320 Log-loss0.03160036355257034\n",
      "Epoch:4 Batch:360 Log-loss0.044408127665519714\n",
      "Epoch:4 Batch:400 Log-loss0.04281621798872948\n",
      "Epoch:4 Batch:440 Log-loss0.039683904498815536\n",
      "Epoch:4 Batch:480 Log-loss0.04508295655250549\n",
      "Epoch:4 Batch:520 Log-loss0.04828350245952606\n",
      "Epoch:4 Batch:560 Log-loss0.03518211841583252\n",
      "Epoch average log loss:0.045573976496234535\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch4, val_loss:0.04135322183313946, best_val_loss:0.04135322183313946, best_auc:0.9864821556485506\n",
      "Epoch:5 Batch:0 Log-loss0.04911888763308525\n",
      "Epoch:5 Batch:40 Log-loss0.038342397660017014\n",
      "Epoch:5 Batch:80 Log-loss0.05336228385567665\n",
      "Epoch:5 Batch:120 Log-loss0.06449111551046371\n",
      "Epoch:5 Batch:160 Log-loss0.05421394109725952\n",
      "Epoch:5 Batch:200 Log-loss0.03508049622178078\n",
      "Epoch:5 Batch:240 Log-loss0.03470610827207565\n",
      "Epoch:5 Batch:280 Log-loss0.05332237109541893\n",
      "Epoch:5 Batch:320 Log-loss0.04227893427014351\n",
      "Epoch:5 Batch:360 Log-loss0.04008572921156883\n",
      "Epoch:5 Batch:400 Log-loss0.047896284610033035\n",
      "Epoch:5 Batch:440 Log-loss0.03972131758928299\n",
      "Epoch:5 Batch:480 Log-loss0.04787345230579376\n",
      "Epoch:5 Batch:520 Log-loss0.03496859595179558\n",
      "Epoch:5 Batch:560 Log-loss0.03808969259262085\n",
      "Epoch average log loss:0.04398595666778939\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch5, val_loss:0.04074859416773909, best_val_loss:0.04074859416773909, best_auc:0.9875287083550474\n",
      "Epoch:6 Batch:0 Log-loss0.032129768282175064\n",
      "Epoch:6 Batch:40 Log-loss0.043850626796483994\n",
      "Epoch:6 Batch:80 Log-loss0.0684269443154335\n",
      "Epoch:6 Batch:120 Log-loss0.055037859827280045\n",
      "Epoch:6 Batch:160 Log-loss0.05322861671447754\n",
      "Epoch:6 Batch:200 Log-loss0.03343113139271736\n",
      "Epoch:6 Batch:240 Log-loss0.047077521681785583\n",
      "Epoch:6 Batch:280 Log-loss0.04757243022322655\n",
      "Epoch:6 Batch:320 Log-loss0.03320103511214256\n",
      "Epoch:6 Batch:360 Log-loss0.03682112321257591\n",
      "Epoch:6 Batch:400 Log-loss0.046005379408597946\n",
      "Epoch:6 Batch:440 Log-loss0.041190650314092636\n",
      "Epoch:6 Batch:480 Log-loss0.0359928123652935\n",
      "Epoch:6 Batch:520 Log-loss0.041538115590810776\n",
      "Epoch:6 Batch:560 Log-loss0.028338247910141945\n",
      "Epoch average log loss:0.043055050537389304\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch6, val_loss:0.04040057876790801, best_val_loss:0.04040057876790801, best_auc:0.9878255893745806\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:7 Batch:0 Log-loss0.03548818454146385\n",
      "Epoch:7 Batch:40 Log-loss0.052665531635284424\n",
      "Epoch:7 Batch:80 Log-loss0.03596782684326172\n",
      "Epoch:7 Batch:120 Log-loss0.04696022346615791\n",
      "Epoch:7 Batch:160 Log-loss0.03771286457777023\n",
      "Epoch:7 Batch:200 Log-loss0.059957217425107956\n",
      "Epoch:7 Batch:240 Log-loss0.048102304339408875\n",
      "Epoch:7 Batch:280 Log-loss0.044488389045000076\n",
      "Epoch:7 Batch:320 Log-loss0.03202056512236595\n",
      "Epoch:7 Batch:360 Log-loss0.04730924963951111\n",
      "Epoch:7 Batch:400 Log-loss0.04715922102332115\n",
      "Epoch:7 Batch:440 Log-loss0.03295135498046875\n",
      "Epoch:7 Batch:480 Log-loss0.03756494075059891\n",
      "Epoch:7 Batch:520 Log-loss0.04242478311061859\n",
      "Epoch:7 Batch:560 Log-loss0.0506175272166729\n",
      "Epoch average log loss:0.04220668766980192\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch7, val_loss:0.03953565719962627, best_val_loss:0.03953565719962627, best_auc:0.9887815929665466\n",
      "Epoch:8 Batch:0 Log-loss0.035534877330064774\n",
      "Epoch:8 Batch:40 Log-loss0.02707836776971817\n",
      "Epoch:8 Batch:80 Log-loss0.034036245197057724\n",
      "Epoch:8 Batch:120 Log-loss0.05005462095141411\n",
      "Epoch:8 Batch:160 Log-loss0.05227324739098549\n",
      "Epoch:8 Batch:200 Log-loss0.027259858325123787\n",
      "Epoch:8 Batch:240 Log-loss0.04719015583395958\n",
      "Epoch:8 Batch:280 Log-loss0.039998870342969894\n",
      "Epoch:8 Batch:320 Log-loss0.042592599987983704\n",
      "Epoch:8 Batch:360 Log-loss0.050832349807024\n",
      "Epoch:8 Batch:400 Log-loss0.03647970035672188\n",
      "Epoch:8 Batch:440 Log-loss0.0475119948387146\n",
      "Epoch:8 Batch:480 Log-loss0.03961664065718651\n",
      "Epoch:8 Batch:520 Log-loss0.02980734594166279\n",
      "Epoch:8 Batch:560 Log-loss0.041342031210660934\n",
      "Epoch average log loss:0.04189465038278806\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch8, val_loss:0.03949839685808723, best_val_loss:0.03949839685808723, best_auc:0.9889933943864192\n",
      "Epoch:9 Batch:0 Log-loss0.04056664928793907\n",
      "Epoch:9 Batch:40 Log-loss0.029567359015345573\n",
      "Epoch:9 Batch:80 Log-loss0.03743496164679527\n",
      "Epoch:9 Batch:120 Log-loss0.04022417962551117\n",
      "Epoch:9 Batch:160 Log-loss0.05589474365115166\n",
      "Epoch:9 Batch:200 Log-loss0.03125044330954552\n",
      "Epoch:9 Batch:240 Log-loss0.036861345171928406\n",
      "Epoch:9 Batch:280 Log-loss0.04491109028458595\n",
      "Epoch:9 Batch:320 Log-loss0.04557633772492409\n",
      "Epoch:9 Batch:360 Log-loss0.04654925689101219\n",
      "Epoch:9 Batch:400 Log-loss0.023399317637085915\n",
      "Epoch:9 Batch:440 Log-loss0.04244418814778328\n",
      "Epoch:9 Batch:480 Log-loss0.030772646889090538\n",
      "Epoch:9 Batch:520 Log-loss0.036371465772390366\n",
      "Epoch:9 Batch:560 Log-loss0.027238866314291954\n",
      "Epoch average log loss:0.041149893100373446\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch9, val_loss:0.03911157026386084, best_val_loss:0.03911157026386084, best_auc:0.9893903966756845\n",
      "Epoch:10 Batch:0 Log-loss0.04738984629511833\n",
      "Epoch:10 Batch:40 Log-loss0.0389787033200264\n",
      "Epoch:10 Batch:80 Log-loss0.04890051111578941\n",
      "Epoch:10 Batch:120 Log-loss0.06107448413968086\n",
      "Epoch:10 Batch:160 Log-loss0.0382637120783329\n",
      "Epoch:10 Batch:200 Log-loss0.04146342724561691\n",
      "Epoch:10 Batch:240 Log-loss0.031920675188302994\n",
      "Epoch:10 Batch:280 Log-loss0.032925259321928024\n",
      "Epoch:10 Batch:320 Log-loss0.05290825292468071\n",
      "Epoch:10 Batch:360 Log-loss0.039462119340896606\n",
      "Epoch:10 Batch:400 Log-loss0.0761236771941185\n",
      "Epoch:10 Batch:440 Log-loss0.038305725902318954\n",
      "Epoch:10 Batch:480 Log-loss0.051441531628370285\n",
      "Epoch:10 Batch:520 Log-loss0.03457540646195412\n",
      "Epoch:10 Batch:560 Log-loss0.029388705268502235\n",
      "Epoch average log loss:0.04072283419248249\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch10, val_loss:0.03873746438430913, best_val_loss:0.03873746438430913, best_auc:0.989707351703454\n",
      "Epoch:11 Batch:0 Log-loss0.04420163854956627\n",
      "Epoch:11 Batch:40 Log-loss0.037799935787916183\n",
      "Epoch:11 Batch:80 Log-loss0.045092593878507614\n",
      "Epoch:11 Batch:120 Log-loss0.03926775977015495\n",
      "Epoch:11 Batch:160 Log-loss0.028663581237196922\n",
      "Epoch:11 Batch:200 Log-loss0.04207427799701691\n",
      "Epoch:11 Batch:240 Log-loss0.037258099764585495\n",
      "Epoch:11 Batch:280 Log-loss0.041280876845121384\n",
      "Epoch:11 Batch:320 Log-loss0.03894316032528877\n",
      "Epoch:11 Batch:360 Log-loss0.034106869250535965\n",
      "Epoch:11 Batch:400 Log-loss0.047021035104990005\n",
      "Epoch:11 Batch:440 Log-loss0.0391559898853302\n",
      "Epoch:11 Batch:480 Log-loss0.04476983845233917\n",
      "Epoch:11 Batch:520 Log-loss0.03950538858771324\n",
      "Epoch:11 Batch:560 Log-loss0.0360223762691021\n",
      "Epoch average log loss:0.04009424161631614\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch11, val_loss:0.038558241340337064, best_val_loss:0.038558241340337064, best_auc:0.9899859864746418\n",
      "Epoch:12 Batch:0 Log-loss0.03838730230927467\n",
      "Epoch:12 Batch:40 Log-loss0.04782569780945778\n",
      "Epoch:12 Batch:80 Log-loss0.06396006047725677\n",
      "Epoch:12 Batch:120 Log-loss0.04994070529937744\n",
      "Epoch:12 Batch:160 Log-loss0.03501191362738609\n",
      "Epoch:12 Batch:200 Log-loss0.03465748205780983\n",
      "Epoch:12 Batch:240 Log-loss0.025623196735978127\n",
      "Epoch:12 Batch:280 Log-loss0.0423363633453846\n",
      "Epoch:12 Batch:320 Log-loss0.04084746167063713\n",
      "Epoch:12 Batch:360 Log-loss0.05292125046253204\n",
      "Epoch:12 Batch:400 Log-loss0.02640448324382305\n",
      "Epoch:12 Batch:440 Log-loss0.045562390238046646\n",
      "Epoch:12 Batch:480 Log-loss0.04220292344689369\n",
      "Epoch:12 Batch:520 Log-loss0.036243926733732224\n",
      "Epoch:12 Batch:560 Log-loss0.04843365401029587\n",
      "Epoch average log loss:0.04005784836065556\n",
      "In Epoch12, val_loss:0.03874969759453919, best_val_loss:0.038558241340337064, best_auc:0.9899859864746418\n",
      "Epoch:13 Batch:0 Log-loss0.030302638188004494\n",
      "Epoch:13 Batch:40 Log-loss0.031055325642228127\n",
      "Epoch:13 Batch:80 Log-loss0.032627034932374954\n",
      "Epoch:13 Batch:120 Log-loss0.04031093046069145\n",
      "Epoch:13 Batch:160 Log-loss0.05082252249121666\n",
      "Epoch:13 Batch:200 Log-loss0.038438472896814346\n",
      "Epoch:13 Batch:240 Log-loss0.0396525077521801\n",
      "Epoch:13 Batch:280 Log-loss0.05352585390210152\n",
      "Epoch:13 Batch:320 Log-loss0.05182700231671333\n",
      "Epoch:13 Batch:360 Log-loss0.024199122563004494\n",
      "Epoch:13 Batch:400 Log-loss0.03168167173862457\n",
      "Epoch:13 Batch:440 Log-loss0.04687267541885376\n",
      "Epoch:13 Batch:480 Log-loss0.0392155759036541\n",
      "Epoch:13 Batch:520 Log-loss0.03644919767975807\n",
      "Epoch:13 Batch:560 Log-loss0.046558644622564316\n",
      "Epoch average log loss:0.03961135685177786\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch13, val_loss:0.03810529382589412, best_val_loss:0.03810529382589412, best_auc:0.9903077410517865\n",
      "Epoch:14 Batch:0 Log-loss0.039515405893325806\n",
      "Epoch:14 Batch:40 Log-loss0.037504643201828\n",
      "Epoch:14 Batch:80 Log-loss0.04196112975478172\n",
      "Epoch:14 Batch:120 Log-loss0.035789113491773605\n",
      "Epoch:14 Batch:160 Log-loss0.043014198541641235\n",
      "Epoch:14 Batch:200 Log-loss0.03341611847281456\n",
      "Epoch:14 Batch:240 Log-loss0.047238413244485855\n",
      "Epoch:14 Batch:280 Log-loss0.03526340797543526\n",
      "Epoch:14 Batch:320 Log-loss0.0469171516597271\n",
      "Epoch:14 Batch:360 Log-loss0.03995130956172943\n",
      "Epoch:14 Batch:400 Log-loss0.04543628171086311\n",
      "Epoch:14 Batch:440 Log-loss0.03962602838873863\n",
      "Epoch:14 Batch:480 Log-loss0.04091816395521164\n",
      "Epoch:14 Batch:520 Log-loss0.04274244233965874\n",
      "Epoch:14 Batch:560 Log-loss0.03490445762872696\n",
      "Epoch average log loss:0.039395811378822795\n",
      "In Epoch14, val_loss:0.038336738486321305, best_val_loss:0.03810529382589412, best_auc:0.9903077410517865\n",
      "Epoch:15 Batch:0 Log-loss0.028821831569075584\n",
      "Epoch:15 Batch:40 Log-loss0.03405437245965004\n",
      "Epoch:15 Batch:80 Log-loss0.03716019541025162\n",
      "Epoch:15 Batch:120 Log-loss0.04161692410707474\n",
      "Epoch:15 Batch:160 Log-loss0.03720926493406296\n",
      "Epoch:15 Batch:200 Log-loss0.055774811655282974\n",
      "Epoch:15 Batch:240 Log-loss0.048964034765958786\n",
      "Epoch:15 Batch:280 Log-loss0.032468684017658234\n",
      "Epoch:15 Batch:320 Log-loss0.028559477999806404\n",
      "Epoch:15 Batch:360 Log-loss0.05653304234147072\n",
      "Epoch:15 Batch:400 Log-loss0.039020586758852005\n",
      "Epoch:15 Batch:440 Log-loss0.03994608297944069\n",
      "Epoch:15 Batch:480 Log-loss0.04471792280673981\n",
      "Epoch:15 Batch:520 Log-loss0.03694408759474754\n",
      "Epoch:15 Batch:560 Log-loss0.03332788497209549\n",
      "Epoch average log loss:0.039138040808029476\n",
      "In Epoch15, val_loss:0.03827138365149401, best_val_loss:0.03810529382589412, best_auc:0.9903077410517865\n",
      "Epoch:16 Batch:0 Log-loss0.03604796901345253\n",
      "Epoch:16 Batch:40 Log-loss0.0517917275428772\n",
      "Epoch:16 Batch:80 Log-loss0.04573889076709747\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:16 Batch:120 Log-loss0.02847808413207531\n",
      "Epoch:16 Batch:160 Log-loss0.04073413088917732\n",
      "Epoch:16 Batch:200 Log-loss0.04970342293381691\n",
      "Epoch:16 Batch:240 Log-loss0.03921099379658699\n",
      "Epoch:16 Batch:280 Log-loss0.02875444106757641\n",
      "Epoch:16 Batch:320 Log-loss0.023150764405727386\n",
      "Epoch:16 Batch:360 Log-loss0.03523242846131325\n",
      "Epoch:16 Batch:400 Log-loss0.03908280283212662\n",
      "Epoch:16 Batch:440 Log-loss0.03343937173485756\n",
      "Epoch:16 Batch:480 Log-loss0.03022467903792858\n",
      "Epoch:16 Batch:520 Log-loss0.05012531206011772\n",
      "Epoch:16 Batch:560 Log-loss0.03380236029624939\n",
      "Epoch average log loss:0.03903021079687668\n",
      "In Epoch16, val_loss:0.03828643841673148, best_val_loss:0.03810529382589412, best_auc:0.9903077410517865\n",
      "Epoch:17 Batch:0 Log-loss0.035158198326826096\n",
      "Epoch:17 Batch:40 Log-loss0.03788575902581215\n",
      "Epoch:17 Batch:80 Log-loss0.041078221052885056\n",
      "Epoch:17 Batch:120 Log-loss0.03147970512509346\n",
      "Epoch:17 Batch:160 Log-loss0.03534971550107002\n",
      "Epoch:17 Batch:200 Log-loss0.05309704318642616\n",
      "Epoch:17 Batch:240 Log-loss0.02885316126048565\n",
      "Epoch:17 Batch:280 Log-loss0.03373369947075844\n",
      "Epoch:17 Batch:320 Log-loss0.025063464418053627\n",
      "Epoch:17 Batch:360 Log-loss0.03767812252044678\n",
      "Epoch:17 Batch:400 Log-loss0.04600685462355614\n",
      "Epoch:17 Batch:440 Log-loss0.030615108087658882\n",
      "Epoch:17 Batch:480 Log-loss0.02964244782924652\n",
      "Epoch:17 Batch:520 Log-loss0.035355519503355026\n",
      "Epoch:17 Batch:560 Log-loss0.058893535286188126\n",
      "Epoch average log loss:0.03862625375789191\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch17, val_loss:0.03807002535509998, best_val_loss:0.03807002535509998, best_auc:0.9904662659483471\n",
      "Epoch:18 Batch:0 Log-loss0.04259592294692993\n",
      "Epoch:18 Batch:40 Log-loss0.05103136971592903\n",
      "Epoch:18 Batch:80 Log-loss0.046156641095876694\n",
      "Epoch:18 Batch:120 Log-loss0.04897114261984825\n",
      "Epoch:18 Batch:160 Log-loss0.03729540482163429\n",
      "Epoch:18 Batch:200 Log-loss0.03336198627948761\n",
      "Epoch:18 Batch:240 Log-loss0.04987921193242073\n",
      "Epoch:18 Batch:280 Log-loss0.03168166056275368\n",
      "Epoch:18 Batch:320 Log-loss0.044567208737134933\n",
      "Epoch:18 Batch:360 Log-loss0.0351526252925396\n",
      "Epoch:18 Batch:400 Log-loss0.0418822318315506\n",
      "Epoch:18 Batch:440 Log-loss0.029944634065032005\n",
      "Epoch:18 Batch:480 Log-loss0.046458449214696884\n",
      "Epoch:18 Batch:520 Log-loss0.028823278844356537\n",
      "Epoch:18 Batch:560 Log-loss0.03401486575603485\n",
      "Epoch average log loss:0.03857414972236646\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch18, val_loss:0.03775931840143904, best_val_loss:0.03775931840143904, best_auc:0.9905469236527327\n",
      "Epoch:19 Batch:0 Log-loss0.04768053814768791\n",
      "Epoch:19 Batch:40 Log-loss0.033250030130147934\n",
      "Epoch:19 Batch:80 Log-loss0.03673900291323662\n",
      "Epoch:19 Batch:120 Log-loss0.03926431015133858\n",
      "Epoch:19 Batch:160 Log-loss0.031879693269729614\n",
      "Epoch:19 Batch:200 Log-loss0.04057524353265762\n",
      "Epoch:19 Batch:240 Log-loss0.032838255167007446\n",
      "Epoch:19 Batch:280 Log-loss0.046915214508771896\n",
      "Epoch:19 Batch:320 Log-loss0.03473503515124321\n",
      "Epoch:19 Batch:360 Log-loss0.04932260885834694\n",
      "Epoch:19 Batch:400 Log-loss0.04554557427763939\n",
      "Epoch:19 Batch:440 Log-loss0.043672043830156326\n",
      "Epoch:19 Batch:480 Log-loss0.034496355801820755\n",
      "Epoch:19 Batch:520 Log-loss0.032911088317632675\n",
      "Epoch:19 Batch:560 Log-loss0.031045900657773018\n",
      "Epoch average log loss:0.038338058298852826\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch19, val_loss:0.037688291174668744, best_val_loss:0.037688291174668744, best_auc:0.9905981971654102\n",
      "Epoch:20 Batch:0 Log-loss0.026723003014922142\n",
      "Epoch:20 Batch:40 Log-loss0.04062233865261078\n",
      "Epoch:20 Batch:80 Log-loss0.03262924775481224\n",
      "Epoch:20 Batch:120 Log-loss0.02988232485949993\n",
      "Epoch:20 Batch:160 Log-loss0.025508219376206398\n",
      "Epoch:20 Batch:200 Log-loss0.04034276679158211\n",
      "Epoch:20 Batch:240 Log-loss0.03889485448598862\n",
      "Epoch:20 Batch:280 Log-loss0.03249341621994972\n",
      "Epoch:20 Batch:320 Log-loss0.04927969351410866\n",
      "Epoch:20 Batch:360 Log-loss0.04126504436135292\n",
      "Epoch:20 Batch:400 Log-loss0.051531534641981125\n",
      "Epoch:20 Batch:440 Log-loss0.02705504186451435\n",
      "Epoch:20 Batch:480 Log-loss0.04234111309051514\n",
      "Epoch:20 Batch:520 Log-loss0.03704550862312317\n",
      "Epoch:20 Batch:560 Log-loss0.032170817255973816\n",
      "Epoch average log loss:0.03811852785625628\n",
      "In Epoch20, val_loss:0.03805095534893122, best_val_loss:0.037688291174668744, best_auc:0.9905981971654102\n",
      "Epoch:21 Batch:0 Log-loss0.03410138189792633\n",
      "Epoch:21 Batch:40 Log-loss0.03928488865494728\n",
      "Epoch:21 Batch:80 Log-loss0.025310000404715538\n",
      "Epoch:21 Batch:120 Log-loss0.03610420227050781\n",
      "Epoch:21 Batch:160 Log-loss0.042683541774749756\n",
      "Epoch:21 Batch:200 Log-loss0.03337988629937172\n",
      "Epoch:21 Batch:240 Log-loss0.054062891751527786\n",
      "Epoch:21 Batch:280 Log-loss0.053513508290052414\n",
      "Epoch:21 Batch:320 Log-loss0.04095509275794029\n",
      "Epoch:21 Batch:360 Log-loss0.016229979693889618\n",
      "Epoch:21 Batch:400 Log-loss0.02786250226199627\n",
      "Epoch:21 Batch:440 Log-loss0.045529965311288834\n",
      "Epoch:21 Batch:480 Log-loss0.03718707337975502\n",
      "Epoch:21 Batch:520 Log-loss0.030350184068083763\n",
      "Epoch:21 Batch:560 Log-loss0.027038784697651863\n",
      "Epoch average log loss:0.03793863884638995\n",
      "In Epoch21, val_loss:0.03776897584067273, best_val_loss:0.037688291174668744, best_auc:0.9905981971654102\n",
      "Epoch:22 Batch:0 Log-loss0.03955140337347984\n",
      "Epoch:22 Batch:40 Log-loss0.03269234672188759\n",
      "Epoch:22 Batch:80 Log-loss0.0358651839196682\n",
      "Epoch:22 Batch:120 Log-loss0.05404435098171234\n",
      "Epoch:22 Batch:160 Log-loss0.03879634290933609\n",
      "Epoch:22 Batch:200 Log-loss0.03567332401871681\n",
      "Epoch:22 Batch:240 Log-loss0.04160812869668007\n",
      "Epoch:22 Batch:280 Log-loss0.03486203774809837\n",
      "Epoch:22 Batch:320 Log-loss0.03597186133265495\n",
      "Epoch:22 Batch:360 Log-loss0.03940849006175995\n",
      "Epoch:22 Batch:400 Log-loss0.023325182497501373\n",
      "Epoch:22 Batch:440 Log-loss0.044091206043958664\n",
      "Epoch:22 Batch:480 Log-loss0.05692624673247337\n",
      "Epoch:22 Batch:520 Log-loss0.03980434685945511\n",
      "Epoch:22 Batch:560 Log-loss0.04084399715065956\n",
      "Epoch average log loss:0.03780152472733919\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch22, val_loss:0.037663528076649085, best_val_loss:0.037663528076649085, best_auc:0.99063638337479\n",
      "Epoch:23 Batch:0 Log-loss0.03871135413646698\n",
      "Epoch:23 Batch:40 Log-loss0.04198372736573219\n",
      "Epoch:23 Batch:80 Log-loss0.02776225097477436\n",
      "Epoch:23 Batch:120 Log-loss0.022875146940350533\n",
      "Epoch:23 Batch:160 Log-loss0.026630468666553497\n",
      "Epoch:23 Batch:200 Log-loss0.03720741719007492\n",
      "Epoch:23 Batch:240 Log-loss0.04589276388287544\n",
      "Epoch:23 Batch:280 Log-loss0.04206153377890587\n",
      "Epoch:23 Batch:320 Log-loss0.024369316175580025\n",
      "Epoch:23 Batch:360 Log-loss0.0436730682849884\n",
      "Epoch:23 Batch:400 Log-loss0.055392149835824966\n",
      "Epoch:23 Batch:440 Log-loss0.045453932136297226\n",
      "Epoch:23 Batch:480 Log-loss0.027243157848715782\n",
      "Epoch:23 Batch:520 Log-loss0.04611216112971306\n",
      "Epoch:23 Batch:560 Log-loss0.03431854769587517\n",
      "Epoch average log loss:0.03794066797542785\n",
      "In Epoch23, val_loss:0.03773853588102876, best_val_loss:0.037663528076649085, best_auc:0.99063638337479\n",
      "Epoch:24 Batch:0 Log-loss0.06660935282707214\n",
      "Epoch:24 Batch:40 Log-loss0.04765395447611809\n",
      "Epoch:24 Batch:80 Log-loss0.05076133832335472\n",
      "Epoch:24 Batch:120 Log-loss0.04127500578761101\n",
      "Epoch:24 Batch:160 Log-loss0.04229182004928589\n",
      "Epoch:24 Batch:200 Log-loss0.026332130655646324\n",
      "Epoch:24 Batch:240 Log-loss0.045406583696603775\n",
      "Epoch:24 Batch:280 Log-loss0.028605477884411812\n",
      "Epoch:24 Batch:320 Log-loss0.04145379364490509\n",
      "Epoch:24 Batch:360 Log-loss0.034648071974515915\n",
      "Epoch:24 Batch:400 Log-loss0.05184311792254448\n",
      "Epoch:24 Batch:440 Log-loss0.03705364838242531\n",
      "Epoch:24 Batch:480 Log-loss0.03137419372797012\n",
      "Epoch:24 Batch:520 Log-loss0.03103744052350521\n",
      "Epoch:24 Batch:560 Log-loss0.03853391855955124\n",
      "Epoch average log loss:0.03753017724957317\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch24, val_loss:0.03758535380011604, best_val_loss:0.03758535380011604, best_auc:0.9906366406253445\n",
      "Epoch:25 Batch:0 Log-loss0.02915312349796295\n",
      "Epoch:25 Batch:40 Log-loss0.03373638913035393\n",
      "Epoch:25 Batch:80 Log-loss0.03524315729737282\n",
      "Epoch:25 Batch:120 Log-loss0.05499933287501335\n",
      "Epoch:25 Batch:160 Log-loss0.030628932639956474\n",
      "Epoch:25 Batch:200 Log-loss0.03605921193957329\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:25 Batch:240 Log-loss0.042076244950294495\n",
      "Epoch:25 Batch:280 Log-loss0.04033327102661133\n",
      "Epoch:25 Batch:320 Log-loss0.042131200432777405\n",
      "Epoch:25 Batch:360 Log-loss0.0439172089099884\n",
      "Epoch:25 Batch:400 Log-loss0.04719462990760803\n",
      "Epoch:25 Batch:440 Log-loss0.035465557128190994\n",
      "Epoch:25 Batch:480 Log-loss0.040575526654720306\n",
      "Epoch:25 Batch:520 Log-loss0.03884885460138321\n",
      "Epoch:25 Batch:560 Log-loss0.035009708255529404\n",
      "Epoch average log loss:0.03756963840153601\n",
      "In Epoch25, val_loss:0.03762516986694113, best_val_loss:0.03758535380011604, best_auc:0.9906366406253445\n",
      "Epoch:26 Batch:0 Log-loss0.04991983249783516\n",
      "Epoch:26 Batch:40 Log-loss0.04616701975464821\n",
      "Epoch:26 Batch:80 Log-loss0.04436659440398216\n",
      "Epoch:26 Batch:120 Log-loss0.03783582150936127\n",
      "Epoch:26 Batch:160 Log-loss0.03593594953417778\n",
      "Epoch:26 Batch:200 Log-loss0.047565627843141556\n",
      "Epoch:26 Batch:240 Log-loss0.04818875715136528\n",
      "Epoch:26 Batch:280 Log-loss0.033925384283065796\n",
      "Epoch:26 Batch:320 Log-loss0.0288123469799757\n",
      "Epoch:26 Batch:360 Log-loss0.03444323316216469\n",
      "Epoch:26 Batch:400 Log-loss0.0245556328445673\n",
      "Epoch:26 Batch:440 Log-loss0.03124626912176609\n",
      "Epoch:26 Batch:480 Log-loss0.024615200236439705\n",
      "Epoch:26 Batch:520 Log-loss0.0361473374068737\n",
      "Epoch:26 Batch:560 Log-loss0.031498417258262634\n",
      "Epoch average log loss:0.03752526392927393\n",
      "In Epoch26, val_loss:0.03776690230173733, best_val_loss:0.03758535380011604, best_auc:0.9906366406253445\n",
      "Epoch:27 Batch:0 Log-loss0.037007950246334076\n",
      "Epoch:27 Batch:40 Log-loss0.062360573559999466\n",
      "Epoch:27 Batch:80 Log-loss0.0418049655854702\n",
      "Epoch:27 Batch:120 Log-loss0.05148252472281456\n",
      "Epoch:27 Batch:160 Log-loss0.04144638404250145\n",
      "Epoch:27 Batch:200 Log-loss0.027125688269734383\n",
      "Epoch:27 Batch:240 Log-loss0.027222424745559692\n",
      "Epoch:27 Batch:280 Log-loss0.04918043315410614\n",
      "Epoch:27 Batch:320 Log-loss0.041630182415246964\n",
      "Epoch:27 Batch:360 Log-loss0.03199784830212593\n",
      "Epoch:27 Batch:400 Log-loss0.02764221467077732\n",
      "Epoch:27 Batch:440 Log-loss0.02803480066359043\n",
      "Epoch:27 Batch:480 Log-loss0.033346254378557205\n",
      "Epoch:27 Batch:520 Log-loss0.03909960761666298\n",
      "Epoch:27 Batch:560 Log-loss0.027162369340658188\n",
      "Epoch average log loss:0.037536428615983046\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch27, val_loss:0.03756161837462838, best_val_loss:0.03756161837462838, best_auc:0.9906529421800466\n",
      "Epoch:28 Batch:0 Log-loss0.02574841119349003\n",
      "Epoch:28 Batch:40 Log-loss0.017620021477341652\n",
      "Epoch:28 Batch:80 Log-loss0.042722251266241074\n",
      "Epoch:28 Batch:120 Log-loss0.048033956438302994\n",
      "Epoch:28 Batch:160 Log-loss0.03560284525156021\n",
      "Epoch:28 Batch:200 Log-loss0.04026104137301445\n",
      "Epoch:28 Batch:240 Log-loss0.03942817077040672\n",
      "Epoch:28 Batch:280 Log-loss0.02927110157907009\n",
      "Epoch:28 Batch:320 Log-loss0.03243154287338257\n",
      "Epoch:28 Batch:360 Log-loss0.020532164722681046\n",
      "Epoch:28 Batch:400 Log-loss0.04533485695719719\n",
      "Epoch:28 Batch:440 Log-loss0.0295188557356596\n",
      "Epoch:28 Batch:480 Log-loss0.02864622324705124\n",
      "Epoch:28 Batch:520 Log-loss0.034142788499593735\n",
      "Epoch:28 Batch:560 Log-loss0.03154713660478592\n",
      "Epoch average log loss:0.037239501419078024\n",
      "In Epoch28, val_loss:0.03757130403679581, best_val_loss:0.03756161837462838, best_auc:0.9906529421800466\n",
      "Epoch:29 Batch:0 Log-loss0.043058473616838455\n",
      "Epoch:29 Batch:40 Log-loss0.03237989917397499\n",
      "Epoch:29 Batch:80 Log-loss0.037477727979421616\n",
      "Epoch:29 Batch:120 Log-loss0.02875979244709015\n",
      "Epoch:29 Batch:160 Log-loss0.03126174584031105\n",
      "Epoch:29 Batch:200 Log-loss0.03155761584639549\n",
      "Epoch:29 Batch:240 Log-loss0.03628905490040779\n",
      "Epoch:29 Batch:280 Log-loss0.020405204966664314\n",
      "Epoch:29 Batch:320 Log-loss0.028316760435700417\n",
      "Epoch:29 Batch:360 Log-loss0.026348663493990898\n",
      "Epoch:29 Batch:400 Log-loss0.040437206625938416\n",
      "Epoch:29 Batch:440 Log-loss0.03566420450806618\n",
      "Epoch:29 Batch:480 Log-loss0.030785061419010162\n",
      "Epoch:29 Batch:520 Log-loss0.03511311858892441\n",
      "Epoch:29 Batch:560 Log-loss0.040318168699741364\n",
      "Epoch average log loss:0.03720370587294123\n",
      "In Epoch29, val_loss:0.037717824749960875, best_val_loss:0.03756161837462838, best_auc:0.9906529421800466\n",
      "Epoch:30 Batch:0 Log-loss0.04036810249090195\n",
      "Epoch:30 Batch:40 Log-loss0.03403953090310097\n",
      "Epoch:30 Batch:80 Log-loss0.055116936564445496\n",
      "Epoch:30 Batch:120 Log-loss0.03979634493589401\n",
      "Epoch:30 Batch:160 Log-loss0.045175742357969284\n",
      "Epoch:30 Batch:200 Log-loss0.030736273154616356\n",
      "Epoch:30 Batch:240 Log-loss0.04602348804473877\n",
      "Epoch:30 Batch:280 Log-loss0.055865075439214706\n",
      "Epoch:30 Batch:320 Log-loss0.03399871662259102\n",
      "Epoch:30 Batch:360 Log-loss0.050478022545576096\n",
      "Epoch:30 Batch:400 Log-loss0.05040769279003143\n",
      "Epoch:30 Batch:440 Log-loss0.037648756057024\n",
      "Epoch:30 Batch:480 Log-loss0.04265691339969635\n",
      "Epoch:30 Batch:520 Log-loss0.03836357221007347\n",
      "Epoch:30 Batch:560 Log-loss0.03751519322395325\n",
      "Epoch average log loss:0.03715292576733711\n",
      "In Epoch30, val_loss:0.03757342089053286, best_val_loss:0.03756161837462838, best_auc:0.9906529421800466\n",
      "Epoch:31 Batch:0 Log-loss0.03629062697291374\n",
      "Epoch:31 Batch:40 Log-loss0.03210274875164032\n",
      "Epoch:31 Batch:80 Log-loss0.03547084704041481\n",
      "Epoch:31 Batch:120 Log-loss0.0532783679664135\n",
      "Epoch:31 Batch:160 Log-loss0.03435686230659485\n",
      "Epoch:31 Batch:200 Log-loss0.020743822678923607\n",
      "Epoch:31 Batch:240 Log-loss0.0353044755756855\n",
      "Epoch:31 Batch:280 Log-loss0.0397816076874733\n",
      "Epoch:31 Batch:320 Log-loss0.04601288214325905\n",
      "Epoch:31 Batch:360 Log-loss0.04645582661032677\n",
      "Epoch:31 Batch:400 Log-loss0.036808982491493225\n",
      "Epoch:31 Batch:440 Log-loss0.039341799914836884\n",
      "Epoch:31 Batch:480 Log-loss0.034331779927015305\n",
      "Epoch:31 Batch:520 Log-loss0.0297837033867836\n",
      "Epoch:31 Batch:560 Log-loss0.03369104862213135\n",
      "Epoch average log loss:0.03703375379887543\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch31, val_loss:0.03756022915740878, best_val_loss:0.03756022915740878, best_auc:0.990674681436794\n",
      "Epoch:32 Batch:0 Log-loss0.040270961821079254\n",
      "Epoch:32 Batch:40 Log-loss0.035581324249506\n",
      "Epoch:32 Batch:80 Log-loss0.030653303489089012\n",
      "Epoch:32 Batch:120 Log-loss0.034066565334796906\n",
      "Epoch:32 Batch:160 Log-loss0.05069318786263466\n",
      "Epoch:32 Batch:200 Log-loss0.021771462634205818\n",
      "Epoch:32 Batch:240 Log-loss0.03339231386780739\n",
      "Epoch:32 Batch:280 Log-loss0.03705660626292229\n",
      "Epoch:32 Batch:320 Log-loss0.03946482762694359\n",
      "Epoch:32 Batch:360 Log-loss0.033377379179000854\n",
      "Epoch:32 Batch:400 Log-loss0.05648709833621979\n",
      "Epoch:32 Batch:440 Log-loss0.030330510810017586\n",
      "Epoch:32 Batch:480 Log-loss0.047222625464200974\n",
      "Epoch:32 Batch:520 Log-loss0.029861032962799072\n",
      "Epoch:32 Batch:560 Log-loss0.03356908634305\n",
      "Epoch average log loss:0.03702559405937791\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch32, val_loss:0.03752064762243834, best_val_loss:0.03752064762243834, best_auc:0.9907191334804722\n",
      "Epoch:33 Batch:0 Log-loss0.02440115250647068\n",
      "Epoch:33 Batch:40 Log-loss0.036766212433576584\n",
      "Epoch:33 Batch:80 Log-loss0.02979958988726139\n",
      "Epoch:33 Batch:120 Log-loss0.045318812131881714\n",
      "Epoch:33 Batch:160 Log-loss0.03211308270692825\n",
      "Epoch:33 Batch:200 Log-loss0.03671051934361458\n",
      "Epoch:33 Batch:240 Log-loss0.03344113007187843\n",
      "Epoch:33 Batch:280 Log-loss0.023719625547528267\n",
      "Epoch:33 Batch:320 Log-loss0.03446203097701073\n",
      "Epoch:33 Batch:360 Log-loss0.05414395406842232\n",
      "Epoch:33 Batch:400 Log-loss0.033951230347156525\n",
      "Epoch:33 Batch:440 Log-loss0.028279118239879608\n",
      "Epoch:33 Batch:480 Log-loss0.03972211852669716\n",
      "Epoch:33 Batch:520 Log-loss0.04005526378750801\n",
      "Epoch:33 Batch:560 Log-loss0.03838134557008743\n",
      "Epoch average log loss:0.03691156999474125\n",
      "In Epoch33, val_loss:0.03758229029692807, best_val_loss:0.03752064762243834, best_auc:0.9907191334804722\n",
      "Epoch:34 Batch:0 Log-loss0.03595205396413803\n",
      "Epoch:34 Batch:40 Log-loss0.023007413372397423\n",
      "Epoch:34 Batch:80 Log-loss0.040718138217926025\n",
      "Epoch:34 Batch:120 Log-loss0.05209320783615112\n",
      "Epoch:34 Batch:160 Log-loss0.04028547927737236\n",
      "Epoch:34 Batch:200 Log-loss0.026034921407699585\n",
      "Epoch:34 Batch:240 Log-loss0.0373840257525444\n",
      "Epoch:34 Batch:280 Log-loss0.033594243228435516\n",
      "Epoch:34 Batch:320 Log-loss0.03630637004971504\n",
      "Epoch:34 Batch:360 Log-loss0.03658752888441086\n",
      "Epoch:34 Batch:400 Log-loss0.045118555426597595\n",
      "Epoch:34 Batch:440 Log-loss0.030279263854026794\n",
      "Epoch:34 Batch:480 Log-loss0.03201538696885109\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:34 Batch:520 Log-loss0.047328025102615356\n",
      "Epoch:34 Batch:560 Log-loss0.04677702486515045\n",
      "Epoch average log loss:0.036833271231236205\n",
      "In Epoch34, val_loss:0.037629429597550834, best_val_loss:0.03752064762243834, best_auc:0.9907191334804722\n",
      "Epoch:35 Batch:0 Log-loss0.03654155880212784\n",
      "Epoch:35 Batch:40 Log-loss0.03824657201766968\n",
      "Epoch:35 Batch:80 Log-loss0.042992692440748215\n",
      "Epoch:35 Batch:120 Log-loss0.03254116326570511\n",
      "Epoch:35 Batch:160 Log-loss0.0536394827067852\n",
      "Epoch:35 Batch:200 Log-loss0.027161823585629463\n",
      "Epoch:35 Batch:240 Log-loss0.04223228618502617\n",
      "Epoch:35 Batch:280 Log-loss0.035007674247026443\n",
      "Epoch:35 Batch:320 Log-loss0.050888512283563614\n",
      "Epoch:35 Batch:360 Log-loss0.028673483058810234\n",
      "Epoch:35 Batch:400 Log-loss0.03539141267538071\n",
      "Epoch:35 Batch:440 Log-loss0.053066834807395935\n",
      "Epoch:35 Batch:480 Log-loss0.028763746842741966\n",
      "Epoch:35 Batch:520 Log-loss0.02728216163814068\n",
      "Epoch:35 Batch:560 Log-loss0.02542814053595066\n",
      "Epoch average log loss:0.03684140107527908\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch35, val_loss:0.03748576615945485, best_val_loss:0.03748576615945485, best_auc:0.9907301496078559\n",
      "Epoch:36 Batch:0 Log-loss0.04039844498038292\n",
      "Epoch:36 Batch:40 Log-loss0.04057399928569794\n",
      "Epoch:36 Batch:80 Log-loss0.02983354963362217\n",
      "Epoch:36 Batch:120 Log-loss0.03525431081652641\n",
      "Epoch:36 Batch:160 Log-loss0.01778329536318779\n",
      "Epoch:36 Batch:200 Log-loss0.03155260160565376\n",
      "Epoch:36 Batch:240 Log-loss0.042107775807380676\n",
      "Epoch:36 Batch:280 Log-loss0.05935535952448845\n",
      "Epoch:36 Batch:320 Log-loss0.030718883499503136\n",
      "Epoch:36 Batch:360 Log-loss0.056798044592142105\n",
      "Epoch:36 Batch:400 Log-loss0.04443603754043579\n",
      "Epoch:36 Batch:440 Log-loss0.03816811367869377\n",
      "Epoch:36 Batch:480 Log-loss0.03726169839501381\n",
      "Epoch:36 Batch:520 Log-loss0.03686914220452309\n",
      "Epoch:36 Batch:560 Log-loss0.03669176250696182\n",
      "Epoch average log loss:0.03674982936520662\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch36, val_loss:0.037454443560324596, best_val_loss:0.037454443560324596, best_auc:0.9907401075675546\n",
      "Epoch:37 Batch:0 Log-loss0.03648534417152405\n",
      "Epoch:37 Batch:40 Log-loss0.04110075533390045\n",
      "Epoch:37 Batch:80 Log-loss0.03257828950881958\n",
      "Epoch:37 Batch:120 Log-loss0.0467073917388916\n",
      "Epoch:37 Batch:160 Log-loss0.026431242004036903\n",
      "Epoch:37 Batch:200 Log-loss0.036659758538007736\n",
      "Epoch:37 Batch:240 Log-loss0.03034829907119274\n",
      "Epoch:37 Batch:280 Log-loss0.03842698782682419\n",
      "Epoch:37 Batch:320 Log-loss0.051822248846292496\n",
      "Epoch:37 Batch:360 Log-loss0.04110092669725418\n",
      "Epoch:37 Batch:400 Log-loss0.032272063195705414\n",
      "Epoch:37 Batch:440 Log-loss0.03323899954557419\n",
      "Epoch:37 Batch:480 Log-loss0.031507182866334915\n",
      "Epoch:37 Batch:520 Log-loss0.05141263082623482\n",
      "Epoch:37 Batch:560 Log-loss0.03817077353596687\n",
      "Epoch average log loss:0.036733934738939364\n",
      "In Epoch37, val_loss:0.03751909081717049, best_val_loss:0.037454443560324596, best_auc:0.9907401075675546\n",
      "Epoch:38 Batch:0 Log-loss0.033710770308971405\n",
      "Epoch:38 Batch:40 Log-loss0.03577810153365135\n",
      "Epoch:38 Batch:80 Log-loss0.03390784561634064\n",
      "Epoch:38 Batch:120 Log-loss0.02772287279367447\n",
      "Epoch:38 Batch:160 Log-loss0.023910365998744965\n",
      "Epoch:38 Batch:200 Log-loss0.033671557903289795\n",
      "Epoch:38 Batch:240 Log-loss0.03722308948636055\n",
      "Epoch:38 Batch:280 Log-loss0.03662875294685364\n",
      "Epoch:38 Batch:320 Log-loss0.024596353992819786\n",
      "Epoch:38 Batch:360 Log-loss0.030148731544613838\n",
      "Epoch:38 Batch:400 Log-loss0.031560029834508896\n",
      "Epoch:38 Batch:440 Log-loss0.026050157845020294\n",
      "Epoch:38 Batch:480 Log-loss0.028126398101449013\n",
      "Epoch:38 Batch:520 Log-loss0.02855091542005539\n",
      "Epoch:38 Batch:560 Log-loss0.038513362407684326\n",
      "Epoch average log loss:0.0367778149433434\n",
      "In Epoch38, val_loss:0.0375262924853878, best_val_loss:0.037454443560324596, best_auc:0.9907401075675546\n",
      "Epoch:39 Batch:0 Log-loss0.03843310847878456\n",
      "Epoch:39 Batch:40 Log-loss0.03360533341765404\n",
      "Epoch:39 Batch:80 Log-loss0.03210759535431862\n",
      "Epoch:39 Batch:120 Log-loss0.028441129252314568\n",
      "Epoch:39 Batch:160 Log-loss0.0443623848259449\n",
      "Epoch:39 Batch:200 Log-loss0.04409251734614372\n",
      "Epoch:39 Batch:240 Log-loss0.02916918508708477\n",
      "Epoch:39 Batch:280 Log-loss0.03911900892853737\n",
      "Epoch:39 Batch:320 Log-loss0.044118236750364304\n",
      "Epoch:39 Batch:360 Log-loss0.02465968392789364\n",
      "Epoch:39 Batch:400 Log-loss0.02317124791443348\n",
      "Epoch:39 Batch:440 Log-loss0.034046005457639694\n",
      "Epoch:39 Batch:480 Log-loss0.023286568000912666\n",
      "Epoch:39 Batch:520 Log-loss0.033328231424093246\n",
      "Epoch:39 Batch:560 Log-loss0.03528204560279846\n",
      "Epoch average log loss:0.03676335844543895\n",
      "In Epoch39, val_loss:0.03750231846505076, best_val_loss:0.037454443560324596, best_auc:0.9907401075675546\n",
      "Epoch:40 Batch:0 Log-loss0.04334867373108864\n",
      "Epoch:40 Batch:40 Log-loss0.04096536710858345\n",
      "Epoch:40 Batch:80 Log-loss0.02870040386915207\n",
      "Epoch:40 Batch:120 Log-loss0.03209679201245308\n",
      "Epoch:40 Batch:160 Log-loss0.04332178458571434\n",
      "Epoch:40 Batch:200 Log-loss0.02938867174088955\n",
      "Epoch:40 Batch:240 Log-loss0.03752301633358002\n",
      "Epoch:40 Batch:280 Log-loss0.02979697473347187\n",
      "Epoch:40 Batch:320 Log-loss0.03126801550388336\n",
      "Epoch:40 Batch:360 Log-loss0.03844590112566948\n",
      "Epoch:40 Batch:400 Log-loss0.02605438232421875\n",
      "Epoch:40 Batch:440 Log-loss0.02812117151916027\n",
      "Epoch:40 Batch:480 Log-loss0.02416459657251835\n",
      "Epoch:40 Batch:520 Log-loss0.037754353135824203\n",
      "Epoch:40 Batch:560 Log-loss0.04585617035627365\n",
      "Epoch average log loss:0.03653361967299133\n",
      "In Epoch40, val_loss:0.03750989683283807, best_val_loss:0.037454443560324596, best_auc:0.9907401075675546\n",
      "Epoch:41 Batch:0 Log-loss0.03495131805539131\n",
      "Epoch:41 Batch:40 Log-loss0.04406316950917244\n",
      "Epoch:41 Batch:80 Log-loss0.041597288101911545\n",
      "Epoch:41 Batch:120 Log-loss0.02821780927479267\n",
      "Epoch:41 Batch:160 Log-loss0.0322282575070858\n",
      "Epoch:41 Batch:200 Log-loss0.036536265164613724\n",
      "Epoch:41 Batch:240 Log-loss0.032182447612285614\n",
      "Epoch:41 Batch:280 Log-loss0.03604556620121002\n",
      "Epoch:41 Batch:320 Log-loss0.02535991184413433\n",
      "Epoch:41 Batch:360 Log-loss0.033075373619794846\n",
      "Epoch:41 Batch:400 Log-loss0.02406797558069229\n",
      "Epoch:41 Batch:440 Log-loss0.03473605588078499\n",
      "Epoch:41 Batch:480 Log-loss0.03896999731659889\n",
      "Epoch:41 Batch:520 Log-loss0.04257621243596077\n",
      "Epoch:41 Batch:560 Log-loss0.027556737884879112\n",
      "Epoch average log loss:0.03672805463096925\n",
      "In Epoch41, val_loss:0.03749758653197784, best_val_loss:0.037454443560324596, best_auc:0.9907401075675546\n",
      "Epoch:42 Batch:0 Log-loss0.036560576409101486\n",
      "Epoch:42 Batch:40 Log-loss0.037621889263391495\n",
      "Epoch:42 Batch:80 Log-loss0.033891092985868454\n",
      "Epoch:42 Batch:120 Log-loss0.047813937067985535\n",
      "Epoch:42 Batch:160 Log-loss0.051469240337610245\n",
      "Epoch:42 Batch:200 Log-loss0.029677212238311768\n",
      "Epoch:42 Batch:240 Log-loss0.0362258180975914\n",
      "Epoch:42 Batch:280 Log-loss0.035153742879629135\n",
      "Epoch:42 Batch:320 Log-loss0.03792461380362511\n",
      "Epoch:42 Batch:360 Log-loss0.05313386023044586\n",
      "Epoch:42 Batch:400 Log-loss0.02747482806444168\n",
      "Epoch:42 Batch:440 Log-loss0.03630854934453964\n",
      "Epoch:42 Batch:480 Log-loss0.026947250589728355\n",
      "Epoch:42 Batch:520 Log-loss0.02765355072915554\n",
      "Epoch:42 Batch:560 Log-loss0.03604019060730934\n",
      "Epoch average log loss:0.03645817614825708\n",
      "In Epoch42, val_loss:0.037486091693382934, best_val_loss:0.037454443560324596, best_auc:0.9907401075675546\n",
      "Epoch:43 Batch:0 Log-loss0.037616800516843796\n",
      "Epoch:43 Batch:40 Log-loss0.03351197391748428\n",
      "Epoch:43 Batch:80 Log-loss0.055157918483018875\n",
      "Epoch:43 Batch:120 Log-loss0.03527913987636566\n",
      "Epoch:43 Batch:160 Log-loss0.05757206305861473\n",
      "Epoch:43 Batch:200 Log-loss0.03864036872982979\n",
      "Epoch:43 Batch:240 Log-loss0.028367841616272926\n",
      "Epoch:43 Batch:280 Log-loss0.04300248622894287\n",
      "Epoch:43 Batch:320 Log-loss0.04644845053553581\n",
      "Epoch:43 Batch:360 Log-loss0.02806922234594822\n",
      "Epoch:43 Batch:400 Log-loss0.032854363322257996\n",
      "Epoch:43 Batch:440 Log-loss0.04181409254670143\n",
      "Epoch:43 Batch:480 Log-loss0.03787432238459587\n",
      "Epoch:43 Batch:520 Log-loss0.027906538918614388\n",
      "Epoch:43 Batch:560 Log-loss0.031099284067749977\n",
      "Epoch average log loss:0.0367429534339213\n",
      "In Epoch43, val_loss:0.03754933584371548, best_val_loss:0.037454443560324596, best_auc:0.9907401075675546\n",
      "Epoch:44 Batch:0 Log-loss0.026778509840369225\n",
      "Epoch:44 Batch:40 Log-loss0.051695141941308975\n",
      "Epoch:44 Batch:80 Log-loss0.02625424601137638\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:44 Batch:120 Log-loss0.050662193447351456\n",
      "Epoch:44 Batch:160 Log-loss0.02383950911462307\n",
      "Epoch:44 Batch:200 Log-loss0.03868343308568001\n",
      "Epoch:44 Batch:240 Log-loss0.04093961417675018\n",
      "Epoch:44 Batch:280 Log-loss0.03775416314601898\n",
      "Epoch:44 Batch:320 Log-loss0.0525713674724102\n",
      "Epoch:44 Batch:360 Log-loss0.03795639052987099\n",
      "Epoch:44 Batch:400 Log-loss0.04328681901097298\n",
      "Epoch:44 Batch:440 Log-loss0.05027790740132332\n",
      "Epoch:44 Batch:480 Log-loss0.042690690606832504\n",
      "Epoch:44 Batch:520 Log-loss0.027824893593788147\n",
      "Epoch:44 Batch:560 Log-loss0.02921321429312229\n",
      "Epoch average log loss:0.0367065680180011\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch44, val_loss:0.037426348052701425, best_val_loss:0.037426348052701425, best_auc:0.9907144276621792\n",
      "Epoch:45 Batch:0 Log-loss0.029776304960250854\n",
      "Epoch:45 Batch:40 Log-loss0.03824174776673317\n",
      "Epoch:45 Batch:80 Log-loss0.025960206985473633\n",
      "Epoch:45 Batch:120 Log-loss0.03196068853139877\n",
      "Epoch:45 Batch:160 Log-loss0.03877422586083412\n",
      "Epoch:45 Batch:200 Log-loss0.03803889825940132\n",
      "Epoch:45 Batch:240 Log-loss0.034448351711034775\n",
      "Epoch:45 Batch:280 Log-loss0.035997312515974045\n",
      "Epoch:45 Batch:320 Log-loss0.0348794050514698\n",
      "Epoch:45 Batch:360 Log-loss0.030090751126408577\n",
      "Epoch:45 Batch:400 Log-loss0.040950484573841095\n",
      "Epoch:45 Batch:440 Log-loss0.04791778698563576\n",
      "Epoch:45 Batch:480 Log-loss0.031047871336340904\n",
      "Epoch:45 Batch:520 Log-loss0.048122942447662354\n",
      "Epoch:45 Batch:560 Log-loss0.038623787462711334\n",
      "Epoch average log loss:0.03666249436459371\n",
      "In Epoch45, val_loss:0.037511840853834244, best_val_loss:0.037426348052701425, best_auc:0.9907144276621792\n",
      "Epoch:46 Batch:0 Log-loss0.03893132135272026\n",
      "Epoch:46 Batch:40 Log-loss0.02618403173983097\n",
      "Epoch:46 Batch:80 Log-loss0.030804278329014778\n",
      "Epoch:46 Batch:120 Log-loss0.053766701370477676\n",
      "Epoch:46 Batch:160 Log-loss0.02505549043416977\n",
      "Epoch:46 Batch:200 Log-loss0.03404397889971733\n",
      "Epoch:46 Batch:240 Log-loss0.03572073206305504\n",
      "Epoch:46 Batch:280 Log-loss0.038214411586523056\n",
      "Epoch:46 Batch:320 Log-loss0.026595277711749077\n",
      "Epoch:46 Batch:360 Log-loss0.039390794932842255\n",
      "Epoch:46 Batch:400 Log-loss0.03701328858733177\n",
      "Epoch:46 Batch:440 Log-loss0.05529038608074188\n",
      "Epoch:46 Batch:480 Log-loss0.029541879892349243\n",
      "Epoch:46 Batch:520 Log-loss0.029570750892162323\n",
      "Epoch:46 Batch:560 Log-loss0.03246171772480011\n",
      "Epoch average log loss:0.036489212742474464\n",
      "In Epoch46, val_loss:0.03748223359314027, best_val_loss:0.037426348052701425, best_auc:0.9907144276621792\n",
      "Epoch:47 Batch:0 Log-loss0.05756397545337677\n",
      "Epoch:47 Batch:40 Log-loss0.03677517548203468\n",
      "Epoch:47 Batch:80 Log-loss0.03993769735097885\n",
      "Epoch:47 Batch:120 Log-loss0.03726543113589287\n",
      "Epoch:47 Batch:160 Log-loss0.04404352232813835\n",
      "Epoch:47 Batch:200 Log-loss0.029624538496136665\n",
      "Epoch:47 Batch:240 Log-loss0.03956606984138489\n",
      "Epoch:47 Batch:280 Log-loss0.037344932556152344\n",
      "Epoch:47 Batch:320 Log-loss0.03352384269237518\n",
      "Epoch:47 Batch:360 Log-loss0.04205654188990593\n",
      "Epoch:47 Batch:400 Log-loss0.03632781654596329\n",
      "Epoch:47 Batch:440 Log-loss0.031404003500938416\n",
      "Epoch:47 Batch:480 Log-loss0.04086321219801903\n",
      "Epoch:47 Batch:520 Log-loss0.024353845044970512\n",
      "Epoch:47 Batch:560 Log-loss0.03601770102977753\n",
      "Epoch average log loss:0.036459519359881856\n",
      "In Epoch47, val_loss:0.03747791491166586, best_val_loss:0.037426348052701425, best_auc:0.9907144276621792\n",
      "Epoch:48 Batch:0 Log-loss0.03993798419833183\n",
      "Epoch:48 Batch:40 Log-loss0.04136229306459427\n",
      "Epoch:48 Batch:80 Log-loss0.03095684014260769\n",
      "Epoch:48 Batch:120 Log-loss0.03163452073931694\n",
      "Epoch:48 Batch:160 Log-loss0.03070612996816635\n",
      "Epoch:48 Batch:200 Log-loss0.029772058129310608\n",
      "Epoch:48 Batch:240 Log-loss0.035828847438097\n",
      "Epoch:48 Batch:280 Log-loss0.03419138863682747\n",
      "Epoch:48 Batch:320 Log-loss0.04125576838850975\n",
      "Epoch:48 Batch:360 Log-loss0.052795376628637314\n",
      "Epoch:48 Batch:400 Log-loss0.02805689163506031\n",
      "Epoch:48 Batch:440 Log-loss0.040572188794612885\n",
      "Epoch:48 Batch:480 Log-loss0.02678816020488739\n",
      "Epoch:48 Batch:520 Log-loss0.043982986360788345\n",
      "Epoch:48 Batch:560 Log-loss0.01974210888147354\n",
      "Epoch average log loss:0.036467193569322785\n",
      "In Epoch48, val_loss:0.037480824776989156, best_val_loss:0.037426348052701425, best_auc:0.9907144276621792\n",
      "Epoch:49 Batch:0 Log-loss0.03979698568582535\n",
      "Epoch:49 Batch:40 Log-loss0.030554188415408134\n",
      "Epoch:49 Batch:80 Log-loss0.042933717370033264\n",
      "Epoch:49 Batch:120 Log-loss0.03585189953446388\n",
      "Epoch:49 Batch:160 Log-loss0.03061298280954361\n",
      "Epoch:49 Batch:200 Log-loss0.04586276039481163\n",
      "Epoch:49 Batch:240 Log-loss0.0269299503415823\n",
      "Epoch:49 Batch:280 Log-loss0.039108503609895706\n",
      "Epoch:49 Batch:320 Log-loss0.03237064555287361\n",
      "Epoch:49 Batch:360 Log-loss0.024235552176833153\n",
      "Epoch:49 Batch:400 Log-loss0.023169590160250664\n",
      "Epoch:49 Batch:440 Log-loss0.03132155165076256\n",
      "Epoch:49 Batch:480 Log-loss0.03644731268286705\n",
      "Epoch:49 Batch:520 Log-loss0.03643643483519554\n",
      "Epoch:49 Batch:560 Log-loss0.024633480235934258\n",
      "Epoch average log loss:0.036362853958936674\n",
      "In Epoch49, val_loss:0.03746752377865209, best_val_loss:0.037426348052701425, best_auc:0.9907144276621792\n",
      "Epoch:50 Batch:0 Log-loss0.026157086715102196\n",
      "Epoch:50 Batch:40 Log-loss0.035399604588747025\n",
      "Epoch:50 Batch:80 Log-loss0.02751690149307251\n",
      "Epoch:50 Batch:120 Log-loss0.027655385434627533\n",
      "Epoch:50 Batch:160 Log-loss0.04211702570319176\n",
      "Epoch:50 Batch:200 Log-loss0.04288533329963684\n",
      "Epoch:50 Batch:240 Log-loss0.048481911420822144\n",
      "Epoch:50 Batch:280 Log-loss0.0324837900698185\n",
      "Epoch:50 Batch:320 Log-loss0.02990337647497654\n",
      "Epoch:50 Batch:360 Log-loss0.03233316168189049\n",
      "Epoch:50 Batch:400 Log-loss0.03411749377846718\n",
      "Epoch:50 Batch:440 Log-loss0.043309565633535385\n",
      "Epoch:50 Batch:480 Log-loss0.02988816611468792\n",
      "Epoch:50 Batch:520 Log-loss0.02856338582932949\n",
      "Epoch:50 Batch:560 Log-loss0.033494435250759125\n",
      "Epoch average log loss:0.03656118582834356\n",
      "In Epoch50, val_loss:0.03748242194896308, best_val_loss:0.037426348052701425, best_auc:0.9907144276621792\n",
      "Epoch:51 Batch:0 Log-loss0.03854780271649361\n",
      "Epoch:51 Batch:40 Log-loss0.03070794977247715\n",
      "Epoch:51 Batch:80 Log-loss0.032648731023073196\n",
      "Epoch:51 Batch:120 Log-loss0.03679773211479187\n",
      "Epoch:51 Batch:160 Log-loss0.037190891802310944\n",
      "Epoch:51 Batch:200 Log-loss0.040229182690382004\n",
      "Epoch:51 Batch:240 Log-loss0.026780465617775917\n",
      "Epoch:51 Batch:280 Log-loss0.033654551953077316\n",
      "Epoch:51 Batch:320 Log-loss0.029010562226176262\n",
      "Epoch:51 Batch:360 Log-loss0.04575179144740105\n",
      "Epoch:51 Batch:400 Log-loss0.039348945021629333\n",
      "Epoch:51 Batch:440 Log-loss0.04538645222783089\n",
      "Epoch:51 Batch:480 Log-loss0.04354294762015343\n",
      "Epoch:51 Batch:520 Log-loss0.03467213734984398\n",
      "Epoch:51 Batch:560 Log-loss0.031412433832883835\n",
      "Epoch average log loss:0.036446957888879944\n",
      "In Epoch51, val_loss:0.03745841053608722, best_val_loss:0.037426348052701425, best_auc:0.9907144276621792\n",
      "Epoch:52 Batch:0 Log-loss0.03826358541846275\n",
      "Epoch:52 Batch:40 Log-loss0.044371943920850754\n",
      "Epoch:52 Batch:80 Log-loss0.03962966427206993\n",
      "Epoch:52 Batch:120 Log-loss0.03893787041306496\n",
      "Epoch:52 Batch:160 Log-loss0.04144815728068352\n",
      "Epoch:52 Batch:200 Log-loss0.020388422533869743\n",
      "Epoch:52 Batch:240 Log-loss0.02690988965332508\n",
      "Epoch:52 Batch:280 Log-loss0.04034512862563133\n",
      "Epoch:52 Batch:320 Log-loss0.038720108568668365\n",
      "Epoch:52 Batch:360 Log-loss0.057048458606004715\n",
      "Epoch:52 Batch:400 Log-loss0.03252666816115379\n",
      "Epoch:52 Batch:440 Log-loss0.03499428555369377\n",
      "Epoch:52 Batch:480 Log-loss0.04152427241206169\n",
      "Epoch:52 Batch:520 Log-loss0.04258376359939575\n",
      "Epoch:52 Batch:560 Log-loss0.026207908987998962\n",
      "Epoch average log loss:0.03638496970358704\n",
      "In Epoch52, val_loss:0.03746114910291993, best_val_loss:0.037426348052701425, best_auc:0.9907144276621792\n",
      "Epoch:53 Batch:0 Log-loss0.03430362790822983\n",
      "Epoch:53 Batch:40 Log-loss0.020273026078939438\n",
      "Epoch:53 Batch:80 Log-loss0.031377702951431274\n",
      "Epoch:53 Batch:120 Log-loss0.05066210404038429\n",
      "Epoch:53 Batch:160 Log-loss0.039031095802783966\n",
      "Epoch:53 Batch:200 Log-loss0.04063449427485466\n",
      "Epoch:53 Batch:240 Log-loss0.027693578973412514\n",
      "Epoch:53 Batch:280 Log-loss0.04241357743740082\n",
      "Epoch:53 Batch:320 Log-loss0.028692694380879402\n",
      "Epoch:53 Batch:360 Log-loss0.03484489396214485\n",
      "Epoch:53 Batch:400 Log-loss0.03895397484302521\n",
      "Epoch:53 Batch:440 Log-loss0.05677283927798271\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:53 Batch:480 Log-loss0.03595719859004021\n",
      "Epoch:53 Batch:520 Log-loss0.0469401516020298\n",
      "Epoch:53 Batch:560 Log-loss0.04815412685275078\n",
      "Epoch average log loss:0.036312858983209094\n",
      "In Epoch53, val_loss:0.03748325602208597, best_val_loss:0.037426348052701425, best_auc:0.9907144276621792\n",
      "Epoch:54 Batch:0 Log-loss0.039839792996644974\n",
      "Epoch:54 Batch:40 Log-loss0.050849948078393936\n",
      "Epoch:54 Batch:80 Log-loss0.04053773358464241\n",
      "Epoch:54 Batch:120 Log-loss0.042080312967300415\n",
      "Epoch:54 Batch:160 Log-loss0.034418996423482895\n",
      "Epoch:54 Batch:200 Log-loss0.03692706301808357\n",
      "Epoch:54 Batch:240 Log-loss0.02783624641597271\n",
      "Epoch:54 Batch:280 Log-loss0.041737109422683716\n",
      "Epoch:54 Batch:320 Log-loss0.0345863401889801\n",
      "Epoch:54 Batch:360 Log-loss0.02867191843688488\n",
      "Epoch:54 Batch:400 Log-loss0.032103750854730606\n",
      "Epoch:54 Batch:440 Log-loss0.042101431638002396\n",
      "Epoch:54 Batch:480 Log-loss0.03534375876188278\n",
      "Epoch:54 Batch:520 Log-loss0.027273863554000854\n",
      "Epoch:54 Batch:560 Log-loss0.03870425745844841\n",
      "Epoch average log loss:0.03641480547043362\n",
      "A pre-trained model at temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt has been loaded.\n",
      "\n",
      "Model has been saved as checkpoints/FASTTXT_BGRU_64_644.pt.\n",
      "\n",
      "Choose the torch base model.\n",
      "Training on fold 5\n",
      "Epoch:1 Batch:0 Log-loss0.6913692951202393\n",
      "Epoch:1 Batch:40 Log-loss0.14444652199745178\n",
      "Epoch:1 Batch:80 Log-loss0.10947064310312271\n",
      "Epoch:1 Batch:120 Log-loss0.05792243406176567\n",
      "Epoch:1 Batch:160 Log-loss0.0693986639380455\n",
      "Epoch:1 Batch:200 Log-loss0.05456719920039177\n",
      "Epoch:1 Batch:240 Log-loss0.060451690107584\n",
      "Epoch:1 Batch:280 Log-loss0.07118817418813705\n",
      "Epoch:1 Batch:320 Log-loss0.052669208496809006\n",
      "Epoch:1 Batch:360 Log-loss0.03945337235927582\n",
      "Epoch:1 Batch:400 Log-loss0.068696029484272\n",
      "Epoch:1 Batch:440 Log-loss0.05660374462604523\n",
      "Epoch:1 Batch:480 Log-loss0.03969307616353035\n",
      "Epoch:1 Batch:520 Log-loss0.04777581989765167\n",
      "Epoch:1 Batch:560 Log-loss0.04436333104968071\n",
      "Epoch average log loss:0.0816867897985503\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch1, val_loss:0.04629929673682523, best_val_loss:0.04629929673682523, best_auc:0.9756436813190651\n",
      "Epoch:2 Batch:0 Log-loss0.06107977405190468\n",
      "Epoch:2 Batch:40 Log-loss0.050138261169195175\n",
      "Epoch:2 Batch:80 Log-loss0.039254363626241684\n",
      "Epoch:2 Batch:120 Log-loss0.04923899471759796\n",
      "Epoch:2 Batch:160 Log-loss0.0463210791349411\n",
      "Epoch:2 Batch:200 Log-loss0.0413607619702816\n",
      "Epoch:2 Batch:240 Log-loss0.04978344216942787\n",
      "Epoch:2 Batch:280 Log-loss0.039901625365018845\n",
      "Epoch:2 Batch:320 Log-loss0.048834413290023804\n",
      "Epoch:2 Batch:360 Log-loss0.0677158311009407\n",
      "Epoch:2 Batch:400 Log-loss0.05961225554347038\n",
      "Epoch:2 Batch:440 Log-loss0.06447271257638931\n",
      "Epoch:2 Batch:480 Log-loss0.04143023490905762\n",
      "Epoch:2 Batch:520 Log-loss0.05634208023548126\n",
      "Epoch:2 Batch:560 Log-loss0.04982476681470871\n",
      "Epoch average log loss:0.05019957511685789\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch2, val_loss:0.042545424168157264, best_val_loss:0.042545424168157264, best_auc:0.9806099729599081\n",
      "Epoch:3 Batch:0 Log-loss0.03931157663464546\n",
      "Epoch:3 Batch:40 Log-loss0.03390613943338394\n",
      "Epoch:3 Batch:80 Log-loss0.04890602454543114\n",
      "Epoch:3 Batch:120 Log-loss0.057574253529310226\n",
      "Epoch:3 Batch:160 Log-loss0.052311282604932785\n",
      "Epoch:3 Batch:200 Log-loss0.056992024183273315\n",
      "Epoch:3 Batch:240 Log-loss0.059587594121694565\n",
      "Epoch:3 Batch:280 Log-loss0.04926019534468651\n",
      "Epoch:3 Batch:320 Log-loss0.059322625398635864\n",
      "Epoch:3 Batch:360 Log-loss0.037799786776304245\n",
      "Epoch:3 Batch:400 Log-loss0.04987514391541481\n",
      "Epoch:3 Batch:440 Log-loss0.02074153907597065\n",
      "Epoch:3 Batch:480 Log-loss0.029640020802617073\n",
      "Epoch:3 Batch:520 Log-loss0.042851272970438004\n",
      "Epoch:3 Batch:560 Log-loss0.046609051525592804\n",
      "Epoch average log loss:0.047488282680777566\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch3, val_loss:0.04209044026205299, best_val_loss:0.04209044026205299, best_auc:0.9836834442808327\n",
      "Epoch:4 Batch:0 Log-loss0.028568459674715996\n",
      "Epoch:4 Batch:40 Log-loss0.031387943774461746\n",
      "Epoch:4 Batch:80 Log-loss0.031625714153051376\n",
      "Epoch:4 Batch:120 Log-loss0.043552085757255554\n",
      "Epoch:4 Batch:160 Log-loss0.033333200961351395\n",
      "Epoch:4 Batch:200 Log-loss0.0482630617916584\n",
      "Epoch:4 Batch:240 Log-loss0.05603400245308876\n",
      "Epoch:4 Batch:280 Log-loss0.035133905708789825\n",
      "Epoch:4 Batch:320 Log-loss0.04838978871703148\n",
      "Epoch:4 Batch:360 Log-loss0.036056917160749435\n",
      "Epoch:4 Batch:400 Log-loss0.04238000512123108\n",
      "Epoch:4 Batch:440 Log-loss0.0338829830288887\n",
      "Epoch:4 Batch:480 Log-loss0.05174681544303894\n",
      "Epoch:4 Batch:520 Log-loss0.03878074511885643\n",
      "Epoch:4 Batch:560 Log-loss0.03338044509291649\n",
      "Epoch average log loss:0.04557684682908335\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch4, val_loss:0.04054394233194187, best_val_loss:0.04054394233194187, best_auc:0.9850965474187272\n",
      "Epoch:5 Batch:0 Log-loss0.03205778822302818\n",
      "Epoch:5 Batch:40 Log-loss0.04731922969222069\n",
      "Epoch:5 Batch:80 Log-loss0.034787628799676895\n",
      "Epoch:5 Batch:120 Log-loss0.04188581928610802\n",
      "Epoch:5 Batch:160 Log-loss0.033788204193115234\n",
      "Epoch:5 Batch:200 Log-loss0.03809262439608574\n",
      "Epoch:5 Batch:240 Log-loss0.05801219120621681\n",
      "Epoch:5 Batch:280 Log-loss0.047392264008522034\n",
      "Epoch:5 Batch:320 Log-loss0.0647178664803505\n",
      "Epoch:5 Batch:360 Log-loss0.057197049260139465\n",
      "Epoch:5 Batch:400 Log-loss0.039803121238946915\n",
      "Epoch:5 Batch:440 Log-loss0.03572804108262062\n",
      "Epoch:5 Batch:480 Log-loss0.04509873315691948\n",
      "Epoch:5 Batch:520 Log-loss0.038569673895835876\n",
      "Epoch:5 Batch:560 Log-loss0.050766922533512115\n",
      "Epoch average log loss:0.04429993190736111\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch5, val_loss:0.03917545164150527, best_val_loss:0.03917545164150527, best_auc:0.9865969082144138\n",
      "Epoch:6 Batch:0 Log-loss0.03529161587357521\n",
      "Epoch:6 Batch:40 Log-loss0.04432624951004982\n",
      "Epoch:6 Batch:80 Log-loss0.0463067926466465\n",
      "Epoch:6 Batch:120 Log-loss0.042198654264211655\n",
      "Epoch:6 Batch:160 Log-loss0.04427763447165489\n",
      "Epoch:6 Batch:200 Log-loss0.04984093829989433\n",
      "Epoch:6 Batch:240 Log-loss0.04092998430132866\n",
      "Epoch:6 Batch:280 Log-loss0.053834009915590286\n",
      "Epoch:6 Batch:320 Log-loss0.039534151554107666\n",
      "Epoch:6 Batch:360 Log-loss0.03941388428211212\n",
      "Epoch:6 Batch:400 Log-loss0.04522527754306793\n",
      "Epoch:6 Batch:440 Log-loss0.03761883080005646\n",
      "Epoch:6 Batch:480 Log-loss0.05459050461649895\n",
      "Epoch:6 Batch:520 Log-loss0.027127278968691826\n",
      "Epoch:6 Batch:560 Log-loss0.0478157140314579\n",
      "Epoch average log loss:0.04330579968435424\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch6, val_loss:0.038748155975315106, best_val_loss:0.038748155975315106, best_auc:0.9876685011164579\n",
      "Epoch:7 Batch:0 Log-loss0.040141645818948746\n",
      "Epoch:7 Batch:40 Log-loss0.0474148690700531\n",
      "Epoch:7 Batch:80 Log-loss0.030340559780597687\n",
      "Epoch:7 Batch:120 Log-loss0.054996196180582047\n",
      "Epoch:7 Batch:160 Log-loss0.034341663122177124\n",
      "Epoch:7 Batch:200 Log-loss0.034530799835920334\n",
      "Epoch:7 Batch:240 Log-loss0.03327740728855133\n",
      "Epoch:7 Batch:280 Log-loss0.033528752624988556\n",
      "Epoch:7 Batch:320 Log-loss0.04266272485256195\n",
      "Epoch:7 Batch:360 Log-loss0.04562912508845329\n",
      "Epoch:7 Batch:400 Log-loss0.04937712848186493\n",
      "Epoch:7 Batch:440 Log-loss0.0354977510869503\n",
      "Epoch:7 Batch:480 Log-loss0.04711533710360527\n",
      "Epoch:7 Batch:520 Log-loss0.04241824150085449\n",
      "Epoch:7 Batch:560 Log-loss0.055129244923591614\n",
      "Epoch average log loss:0.042601203133485144\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch7, val_loss:0.038098151560215364, best_val_loss:0.038098151560215364, best_auc:0.9885057555805622\n",
      "Epoch:8 Batch:0 Log-loss0.03735819831490517\n",
      "Epoch:8 Batch:40 Log-loss0.03460785746574402\n",
      "Epoch:8 Batch:80 Log-loss0.03858811780810356\n",
      "Epoch:8 Batch:120 Log-loss0.02206825651228428\n",
      "Epoch:8 Batch:160 Log-loss0.029797131195664406\n",
      "Epoch:8 Batch:200 Log-loss0.038376469165086746\n",
      "Epoch:8 Batch:240 Log-loss0.03569626063108444\n",
      "Epoch:8 Batch:280 Log-loss0.0453963428735733\n",
      "Epoch:8 Batch:320 Log-loss0.041466351598501205\n",
      "Epoch:8 Batch:360 Log-loss0.0340612418949604\n",
      "Epoch:8 Batch:400 Log-loss0.026447294279932976\n",
      "Epoch:8 Batch:440 Log-loss0.04567592218518257\n",
      "Epoch:8 Batch:480 Log-loss0.02809031493961811\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:8 Batch:520 Log-loss0.03923605754971504\n",
      "Epoch:8 Batch:560 Log-loss0.02341325767338276\n",
      "Epoch average log loss:0.042146239898699735\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch8, val_loss:0.03792250036242368, best_val_loss:0.03792250036242368, best_auc:0.9891503180111494\n",
      "Epoch:9 Batch:0 Log-loss0.04385707154870033\n",
      "Epoch:9 Batch:40 Log-loss0.03290438652038574\n",
      "Epoch:9 Batch:80 Log-loss0.04584120586514473\n",
      "Epoch:9 Batch:120 Log-loss0.04214106500148773\n",
      "Epoch:9 Batch:160 Log-loss0.032752957195043564\n",
      "Epoch:9 Batch:200 Log-loss0.03788946568965912\n",
      "Epoch:9 Batch:240 Log-loss0.04010153189301491\n",
      "Epoch:9 Batch:280 Log-loss0.03367197886109352\n",
      "Epoch:9 Batch:320 Log-loss0.03964441269636154\n",
      "Epoch:9 Batch:360 Log-loss0.04141067713499069\n",
      "Epoch:9 Batch:400 Log-loss0.03334914147853851\n",
      "Epoch:9 Batch:440 Log-loss0.034839730709791183\n",
      "Epoch:9 Batch:480 Log-loss0.04178367182612419\n",
      "Epoch:9 Batch:520 Log-loss0.04252682253718376\n",
      "Epoch:9 Batch:560 Log-loss0.031889401376247406\n",
      "Epoch average log loss:0.041389755798237664\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch9, val_loss:0.037453298187057144, best_val_loss:0.037453298187057144, best_auc:0.9897308217920094\n",
      "Epoch:10 Batch:0 Log-loss0.032970357686281204\n",
      "Epoch:10 Batch:40 Log-loss0.045113515108823776\n",
      "Epoch:10 Batch:80 Log-loss0.03651709109544754\n",
      "Epoch:10 Batch:120 Log-loss0.049396175891160965\n",
      "Epoch:10 Batch:160 Log-loss0.032850101590156555\n",
      "Epoch:10 Batch:200 Log-loss0.03023301251232624\n",
      "Epoch:10 Batch:240 Log-loss0.05723379924893379\n",
      "Epoch:10 Batch:280 Log-loss0.03640115633606911\n",
      "Epoch:10 Batch:320 Log-loss0.05332178995013237\n",
      "Epoch:10 Batch:360 Log-loss0.03541719168424606\n",
      "Epoch:10 Batch:400 Log-loss0.03784845024347305\n",
      "Epoch:10 Batch:440 Log-loss0.041576918214559555\n",
      "Epoch:10 Batch:480 Log-loss0.04402206465601921\n",
      "Epoch:10 Batch:520 Log-loss0.042256999760866165\n",
      "Epoch:10 Batch:560 Log-loss0.03971671313047409\n",
      "Epoch average log loss:0.0408799862854981\n",
      "In Epoch10, val_loss:0.03745658826841907, best_val_loss:0.037453298187057144, best_auc:0.9897308217920094\n",
      "Epoch:11 Batch:0 Log-loss0.038434140384197235\n",
      "Epoch:11 Batch:40 Log-loss0.03463507071137428\n",
      "Epoch:11 Batch:80 Log-loss0.051630448549985886\n",
      "Epoch:11 Batch:120 Log-loss0.03630588576197624\n",
      "Epoch:11 Batch:160 Log-loss0.035894449800252914\n",
      "Epoch:11 Batch:200 Log-loss0.03837592154741287\n",
      "Epoch:11 Batch:240 Log-loss0.04247089847922325\n",
      "Epoch:11 Batch:280 Log-loss0.0315893329679966\n",
      "Epoch:11 Batch:320 Log-loss0.04701871797442436\n",
      "Epoch:11 Batch:360 Log-loss0.03396103158593178\n",
      "Epoch:11 Batch:400 Log-loss0.032991837710142136\n",
      "Epoch:11 Batch:440 Log-loss0.018324866890907288\n",
      "Epoch:11 Batch:480 Log-loss0.038790926337242126\n",
      "Epoch:11 Batch:520 Log-loss0.04092497006058693\n",
      "Epoch:11 Batch:560 Log-loss0.034557897597551346\n",
      "Epoch average log loss:0.04056518297376377\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch11, val_loss:0.03718329137647273, best_val_loss:0.03718329137647273, best_auc:0.9904208733081207\n",
      "Epoch:12 Batch:0 Log-loss0.04741417244076729\n",
      "Epoch:12 Batch:40 Log-loss0.04072757437825203\n",
      "Epoch:12 Batch:80 Log-loss0.04418464004993439\n",
      "Epoch:12 Batch:120 Log-loss0.03663933649659157\n",
      "Epoch:12 Batch:160 Log-loss0.040704257786273956\n",
      "Epoch:12 Batch:200 Log-loss0.039686087518930435\n",
      "Epoch:12 Batch:240 Log-loss0.02325172908604145\n",
      "Epoch:12 Batch:280 Log-loss0.029085159301757812\n",
      "Epoch:12 Batch:320 Log-loss0.03198013827204704\n",
      "Epoch:12 Batch:360 Log-loss0.042363617569208145\n",
      "Epoch:12 Batch:400 Log-loss0.03496985882520676\n",
      "Epoch:12 Batch:440 Log-loss0.03362223505973816\n",
      "Epoch:12 Batch:480 Log-loss0.034775543957948685\n",
      "Epoch:12 Batch:520 Log-loss0.034363653510808945\n",
      "Epoch:12 Batch:560 Log-loss0.03925414755940437\n",
      "Epoch average log loss:0.040313157472493394\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch12, val_loss:0.037067843068628474, best_val_loss:0.037067843068628474, best_auc:0.9902862700188811\n",
      "Epoch:13 Batch:0 Log-loss0.05334747955203056\n",
      "Epoch:13 Batch:40 Log-loss0.050451356917619705\n",
      "Epoch:13 Batch:80 Log-loss0.05181758478283882\n",
      "Epoch:13 Batch:120 Log-loss0.036804910749197006\n",
      "Epoch:13 Batch:160 Log-loss0.047177914530038834\n",
      "Epoch:13 Batch:200 Log-loss0.039658073335886\n",
      "Epoch:13 Batch:240 Log-loss0.033707644790410995\n",
      "Epoch:13 Batch:280 Log-loss0.04427753761410713\n",
      "Epoch:13 Batch:320 Log-loss0.035094987601041794\n",
      "Epoch:13 Batch:360 Log-loss0.041152600198984146\n",
      "Epoch:13 Batch:400 Log-loss0.034882958978414536\n",
      "Epoch:13 Batch:440 Log-loss0.040536440908908844\n",
      "Epoch:13 Batch:480 Log-loss0.048613663762807846\n",
      "Epoch:13 Batch:520 Log-loss0.04113040491938591\n",
      "Epoch:13 Batch:560 Log-loss0.03262479603290558\n",
      "Epoch average log loss:0.03980718266019331\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch13, val_loss:0.037029332856914, best_val_loss:0.037029332856914, best_auc:0.9904556456941354\n",
      "Epoch:14 Batch:0 Log-loss0.050513461232185364\n",
      "Epoch:14 Batch:40 Log-loss0.04050818085670471\n",
      "Epoch:14 Batch:80 Log-loss0.034946877509355545\n",
      "Epoch:14 Batch:120 Log-loss0.0472324900329113\n",
      "Epoch:14 Batch:160 Log-loss0.034612126648426056\n",
      "Epoch:14 Batch:200 Log-loss0.05927233770489693\n",
      "Epoch:14 Batch:240 Log-loss0.04146328195929527\n",
      "Epoch:14 Batch:280 Log-loss0.05025315657258034\n",
      "Epoch:14 Batch:320 Log-loss0.04209606721997261\n",
      "Epoch:14 Batch:360 Log-loss0.04040704667568207\n",
      "Epoch:14 Batch:400 Log-loss0.047708913683891296\n",
      "Epoch:14 Batch:440 Log-loss0.035179901868104935\n",
      "Epoch:14 Batch:480 Log-loss0.036311764270067215\n",
      "Epoch:14 Batch:520 Log-loss0.041579630225896835\n",
      "Epoch:14 Batch:560 Log-loss0.03570771589875221\n",
      "Epoch average log loss:0.03974558477556067\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch14, val_loss:0.036909213026276186, best_val_loss:0.036909213026276186, best_auc:0.9906109783534106\n",
      "Epoch:15 Batch:0 Log-loss0.0294269397854805\n",
      "Epoch:15 Batch:40 Log-loss0.0355643704533577\n",
      "Epoch:15 Batch:80 Log-loss0.04538401961326599\n",
      "Epoch:15 Batch:120 Log-loss0.035511117428541183\n",
      "Epoch:15 Batch:160 Log-loss0.03283995762467384\n",
      "Epoch:15 Batch:200 Log-loss0.03467290848493576\n",
      "Epoch:15 Batch:240 Log-loss0.037975456565618515\n",
      "Epoch:15 Batch:280 Log-loss0.032372910529375076\n",
      "Epoch:15 Batch:320 Log-loss0.026455475017428398\n",
      "Epoch:15 Batch:360 Log-loss0.034744419157505035\n",
      "Epoch:15 Batch:400 Log-loss0.04526063799858093\n",
      "Epoch:15 Batch:440 Log-loss0.03608722612261772\n",
      "Epoch:15 Batch:480 Log-loss0.03617088124155998\n",
      "Epoch:15 Batch:520 Log-loss0.028553633019328117\n",
      "Epoch:15 Batch:560 Log-loss0.05366894602775574\n",
      "Epoch average log loss:0.03931145059510267\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch15, val_loss:0.03679562330277649, best_val_loss:0.03679562330277649, best_auc:0.9904914279309747\n",
      "Epoch:16 Batch:0 Log-loss0.04078557714819908\n",
      "Epoch:16 Batch:40 Log-loss0.03677363321185112\n",
      "Epoch:16 Batch:80 Log-loss0.04410058632493019\n",
      "Epoch:16 Batch:120 Log-loss0.03327653184533119\n",
      "Epoch:16 Batch:160 Log-loss0.040460292249917984\n",
      "Epoch:16 Batch:200 Log-loss0.03667965531349182\n",
      "Epoch:16 Batch:240 Log-loss0.04554145410656929\n",
      "Epoch:16 Batch:280 Log-loss0.04236016795039177\n",
      "Epoch:16 Batch:320 Log-loss0.04350294545292854\n",
      "Epoch:16 Batch:360 Log-loss0.043698500841856\n",
      "Epoch:16 Batch:400 Log-loss0.029119813814759254\n",
      "Epoch:16 Batch:440 Log-loss0.04422798752784729\n",
      "Epoch:16 Batch:480 Log-loss0.03852897509932518\n",
      "Epoch:16 Batch:520 Log-loss0.03799489513039589\n",
      "Epoch:16 Batch:560 Log-loss0.03541090339422226\n",
      "Epoch average log loss:0.03928610598003226\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch16, val_loss:0.03670715160384994, best_val_loss:0.03670715160384994, best_auc:0.9907649454506751\n",
      "Epoch:17 Batch:0 Log-loss0.044251829385757446\n",
      "Epoch:17 Batch:40 Log-loss0.035275597125291824\n",
      "Epoch:17 Batch:80 Log-loss0.04715790972113609\n",
      "Epoch:17 Batch:120 Log-loss0.03454773873090744\n",
      "Epoch:17 Batch:160 Log-loss0.04484742879867554\n",
      "Epoch:17 Batch:200 Log-loss0.044464077800512314\n",
      "Epoch:17 Batch:240 Log-loss0.041220132261514664\n",
      "Epoch:17 Batch:280 Log-loss0.042583346366882324\n",
      "Epoch:17 Batch:320 Log-loss0.04213153198361397\n",
      "Epoch:17 Batch:360 Log-loss0.03142712637782097\n",
      "Epoch:17 Batch:400 Log-loss0.04544190689921379\n",
      "Epoch:17 Batch:440 Log-loss0.05727905407547951\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:17 Batch:480 Log-loss0.03868662193417549\n",
      "Epoch:17 Batch:520 Log-loss0.043887943029403687\n",
      "Epoch:17 Batch:560 Log-loss0.034761834889650345\n",
      "Epoch average log loss:0.03907029921122428\n",
      "In Epoch17, val_loss:0.03674639646263186, best_val_loss:0.03670715160384994, best_auc:0.9907649454506751\n",
      "Epoch:18 Batch:0 Log-loss0.03137710317969322\n",
      "Epoch:18 Batch:40 Log-loss0.03229707479476929\n",
      "Epoch:18 Batch:80 Log-loss0.03881513699889183\n",
      "Epoch:18 Batch:120 Log-loss0.03610445186495781\n",
      "Epoch:18 Batch:160 Log-loss0.03558258339762688\n",
      "Epoch:18 Batch:200 Log-loss0.03684235364198685\n",
      "Epoch:18 Batch:240 Log-loss0.05346217751502991\n",
      "Epoch:18 Batch:280 Log-loss0.045369382947683334\n",
      "Epoch:18 Batch:320 Log-loss0.034988466650247574\n",
      "Epoch:18 Batch:360 Log-loss0.045316021889448166\n",
      "Epoch:18 Batch:400 Log-loss0.044925857335329056\n",
      "Epoch:18 Batch:440 Log-loss0.03924776241183281\n",
      "Epoch:18 Batch:480 Log-loss0.04326817765831947\n",
      "Epoch:18 Batch:520 Log-loss0.03963537514209747\n",
      "Epoch:18 Batch:560 Log-loss0.03409786522388458\n",
      "Epoch average log loss:0.038874478568322955\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch18, val_loss:0.03659744608391271, best_val_loss:0.03659744608391271, best_auc:0.990768694906957\n",
      "Epoch:19 Batch:0 Log-loss0.03217869624495506\n",
      "Epoch:19 Batch:40 Log-loss0.029886724427342415\n",
      "Epoch:19 Batch:80 Log-loss0.04179021716117859\n",
      "Epoch:19 Batch:120 Log-loss0.03561894968152046\n",
      "Epoch:19 Batch:160 Log-loss0.029228450730443\n",
      "Epoch:19 Batch:200 Log-loss0.0314464308321476\n",
      "Epoch:19 Batch:240 Log-loss0.041234951466321945\n",
      "Epoch:19 Batch:280 Log-loss0.03219960629940033\n",
      "Epoch:19 Batch:320 Log-loss0.043573394417762756\n",
      "Epoch:19 Batch:360 Log-loss0.04069763794541359\n",
      "Epoch:19 Batch:400 Log-loss0.03204041346907616\n",
      "Epoch:19 Batch:440 Log-loss0.04305645823478699\n",
      "Epoch:19 Batch:480 Log-loss0.04212762787938118\n",
      "Epoch:19 Batch:520 Log-loss0.05302249267697334\n",
      "Epoch:19 Batch:560 Log-loss0.03657639026641846\n",
      "Epoch average log loss:0.038686054429438496\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch19, val_loss:0.03642981715004291, best_val_loss:0.03642981715004291, best_auc:0.9908758925368805\n",
      "Epoch:20 Batch:0 Log-loss0.039708252996206284\n",
      "Epoch:20 Batch:40 Log-loss0.023286089301109314\n",
      "Epoch:20 Batch:80 Log-loss0.0445268340408802\n",
      "Epoch:20 Batch:120 Log-loss0.06343313306570053\n",
      "Epoch:20 Batch:160 Log-loss0.03654618188738823\n",
      "Epoch:20 Batch:200 Log-loss0.04582531750202179\n",
      "Epoch:20 Batch:240 Log-loss0.037635717540979385\n",
      "Epoch:20 Batch:280 Log-loss0.04367608204483986\n",
      "Epoch:20 Batch:320 Log-loss0.0285947322845459\n",
      "Epoch:20 Batch:360 Log-loss0.041535694152116776\n",
      "Epoch:20 Batch:400 Log-loss0.043910253793001175\n",
      "Epoch:20 Batch:440 Log-loss0.037503745406866074\n",
      "Epoch:20 Batch:480 Log-loss0.02704993449151516\n",
      "Epoch:20 Batch:520 Log-loss0.03190441429615021\n",
      "Epoch:20 Batch:560 Log-loss0.03742451220750809\n",
      "Epoch average log loss:0.038530731157931895\n",
      "In Epoch20, val_loss:0.036629769690830365, best_val_loss:0.03642981715004291, best_auc:0.9908758925368805\n",
      "Epoch:21 Batch:0 Log-loss0.03753868117928505\n",
      "Epoch:21 Batch:40 Log-loss0.03720428794622421\n",
      "Epoch:21 Batch:80 Log-loss0.02885720320045948\n",
      "Epoch:21 Batch:120 Log-loss0.03929968550801277\n",
      "Epoch:21 Batch:160 Log-loss0.026043610647320747\n",
      "Epoch:21 Batch:200 Log-loss0.03776402771472931\n",
      "Epoch:21 Batch:240 Log-loss0.03347480297088623\n",
      "Epoch:21 Batch:280 Log-loss0.023566482588648796\n",
      "Epoch:21 Batch:320 Log-loss0.04587521776556969\n",
      "Epoch:21 Batch:360 Log-loss0.047876764088869095\n",
      "Epoch:21 Batch:400 Log-loss0.052573174238204956\n",
      "Epoch:21 Batch:440 Log-loss0.046729300171136856\n",
      "Epoch:21 Batch:480 Log-loss0.03721090033650398\n",
      "Epoch:21 Batch:520 Log-loss0.036341842263936996\n",
      "Epoch:21 Batch:560 Log-loss0.02943810634315014\n",
      "Epoch average log loss:0.038105244897971194\n",
      "In Epoch21, val_loss:0.0364697462177155, best_val_loss:0.03642981715004291, best_auc:0.9908758925368805\n",
      "Epoch:22 Batch:0 Log-loss0.0370696596801281\n",
      "Epoch:22 Batch:40 Log-loss0.019129648804664612\n",
      "Epoch:22 Batch:80 Log-loss0.04354005679488182\n",
      "Epoch:22 Batch:120 Log-loss0.053019020706415176\n",
      "Epoch:22 Batch:160 Log-loss0.029649635776877403\n",
      "Epoch:22 Batch:200 Log-loss0.0396878756582737\n",
      "Epoch:22 Batch:240 Log-loss0.03361845016479492\n",
      "Epoch:22 Batch:280 Log-loss0.045740287750959396\n",
      "Epoch:22 Batch:320 Log-loss0.04228934645652771\n",
      "Epoch:22 Batch:360 Log-loss0.025532705709338188\n",
      "Epoch:22 Batch:400 Log-loss0.04930688068270683\n",
      "Epoch:22 Batch:440 Log-loss0.0639667883515358\n",
      "Epoch:22 Batch:480 Log-loss0.055798839777708054\n",
      "Epoch:22 Batch:520 Log-loss0.034892212599515915\n",
      "Epoch:22 Batch:560 Log-loss0.04538695886731148\n",
      "Epoch average log loss:0.0382648779844333\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch22, val_loss:0.036379150712878186, best_val_loss:0.036379150712878186, best_auc:0.9909773329175579\n",
      "Epoch:23 Batch:0 Log-loss0.03432147577404976\n",
      "Epoch:23 Batch:40 Log-loss0.027052557095885277\n",
      "Epoch:23 Batch:80 Log-loss0.03296980634331703\n",
      "Epoch:23 Batch:120 Log-loss0.034607674926519394\n",
      "Epoch:23 Batch:160 Log-loss0.035089556127786636\n",
      "Epoch:23 Batch:200 Log-loss0.04084790498018265\n",
      "Epoch:23 Batch:240 Log-loss0.058275263756513596\n",
      "Epoch:23 Batch:280 Log-loss0.055551677942276\n",
      "Epoch:23 Batch:320 Log-loss0.04248499870300293\n",
      "Epoch:23 Batch:360 Log-loss0.03273260220885277\n",
      "Epoch:23 Batch:400 Log-loss0.04227990284562111\n",
      "Epoch:23 Batch:440 Log-loss0.036696311086416245\n",
      "Epoch:23 Batch:480 Log-loss0.03172314539551735\n",
      "Epoch:23 Batch:520 Log-loss0.031045086681842804\n",
      "Epoch:23 Batch:560 Log-loss0.032123975455760956\n",
      "Epoch average log loss:0.03799097093287855\n",
      "In Epoch23, val_loss:0.03659347982805558, best_val_loss:0.036379150712878186, best_auc:0.9909773329175579\n",
      "Epoch:24 Batch:0 Log-loss0.04071586951613426\n",
      "Epoch:24 Batch:40 Log-loss0.04448382556438446\n",
      "Epoch:24 Batch:80 Log-loss0.029495900496840477\n",
      "Epoch:24 Batch:120 Log-loss0.030111854895949364\n",
      "Epoch:24 Batch:160 Log-loss0.04470718279480934\n",
      "Epoch:24 Batch:200 Log-loss0.029170939698815346\n",
      "Epoch:24 Batch:240 Log-loss0.026009241119027138\n",
      "Epoch:24 Batch:280 Log-loss0.05014798417687416\n",
      "Epoch:24 Batch:320 Log-loss0.05187197029590607\n",
      "Epoch:24 Batch:360 Log-loss0.029343420639634132\n",
      "Epoch:24 Batch:400 Log-loss0.025177448987960815\n",
      "Epoch:24 Batch:440 Log-loss0.03217833489179611\n",
      "Epoch:24 Batch:480 Log-loss0.03971670940518379\n",
      "Epoch:24 Batch:520 Log-loss0.03869776427745819\n",
      "Epoch:24 Batch:560 Log-loss0.03463203087449074\n",
      "Epoch average log loss:0.03802948940179444\n",
      "In Epoch24, val_loss:0.036512502376826315, best_val_loss:0.036379150712878186, best_auc:0.9909773329175579\n",
      "Epoch:25 Batch:0 Log-loss0.037379249930381775\n",
      "Epoch:25 Batch:40 Log-loss0.046096399426460266\n",
      "Epoch:25 Batch:80 Log-loss0.02946295589208603\n",
      "Epoch:25 Batch:120 Log-loss0.03472684696316719\n",
      "Epoch:25 Batch:160 Log-loss0.0469161681830883\n",
      "Epoch:25 Batch:200 Log-loss0.047537911683321\n",
      "Epoch:25 Batch:240 Log-loss0.03096240758895874\n",
      "Epoch:25 Batch:280 Log-loss0.03191857412457466\n",
      "Epoch:25 Batch:320 Log-loss0.03826035186648369\n",
      "Epoch:25 Batch:360 Log-loss0.03499506786465645\n",
      "Epoch:25 Batch:400 Log-loss0.02956194430589676\n",
      "Epoch:25 Batch:440 Log-loss0.046338941901922226\n",
      "Epoch:25 Batch:480 Log-loss0.024639254435896873\n",
      "Epoch:25 Batch:520 Log-loss0.03104483149945736\n",
      "Epoch:25 Batch:560 Log-loss0.04590165242552757\n",
      "Epoch average log loss:0.03777582957076707\n",
      "In Epoch25, val_loss:0.03645867929671908, best_val_loss:0.036379150712878186, best_auc:0.9909773329175579\n",
      "Epoch:26 Batch:0 Log-loss0.023717178031802177\n",
      "Epoch:26 Batch:40 Log-loss0.031279560178518295\n",
      "Epoch:26 Batch:80 Log-loss0.049663424491882324\n",
      "Epoch:26 Batch:120 Log-loss0.04088987410068512\n",
      "Epoch:26 Batch:160 Log-loss0.036117300391197205\n",
      "Epoch:26 Batch:200 Log-loss0.045550186187028885\n",
      "Epoch:26 Batch:240 Log-loss0.04243190586566925\n",
      "Epoch:26 Batch:280 Log-loss0.03881818428635597\n",
      "Epoch:26 Batch:320 Log-loss0.021936826407909393\n",
      "Epoch:26 Batch:360 Log-loss0.03752032294869423\n",
      "Epoch:26 Batch:400 Log-loss0.030204566195607185\n",
      "Epoch:26 Batch:440 Log-loss0.023665564134716988\n",
      "Epoch:26 Batch:480 Log-loss0.03674729913473129\n",
      "Epoch:26 Batch:520 Log-loss0.03395231068134308\n",
      "Epoch:26 Batch:560 Log-loss0.04669627919793129\n",
      "Epoch average log loss:0.03794795767670231\n",
      "In Epoch26, val_loss:0.03641097940846635, best_val_loss:0.036379150712878186, best_auc:0.9909773329175579\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:27 Batch:0 Log-loss0.032058022916316986\n",
      "Epoch:27 Batch:40 Log-loss0.037677913904190063\n",
      "Epoch:27 Batch:80 Log-loss0.031161850318312645\n",
      "Epoch:27 Batch:120 Log-loss0.02310800738632679\n",
      "Epoch:27 Batch:160 Log-loss0.028846995905041695\n",
      "Epoch:27 Batch:200 Log-loss0.03471627086400986\n",
      "Epoch:27 Batch:240 Log-loss0.04245747625827789\n",
      "Epoch:27 Batch:280 Log-loss0.03251992538571358\n",
      "Epoch:27 Batch:320 Log-loss0.030567854642868042\n",
      "Epoch:27 Batch:360 Log-loss0.04350472614169121\n",
      "Epoch:27 Batch:400 Log-loss0.046614814549684525\n",
      "Epoch:27 Batch:440 Log-loss0.04512598738074303\n",
      "Epoch:27 Batch:480 Log-loss0.04219217598438263\n",
      "Epoch:27 Batch:520 Log-loss0.030640453100204468\n",
      "Epoch:27 Batch:560 Log-loss0.03488735482096672\n",
      "Epoch average log loss:0.03753504768080477\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch27, val_loss:0.03633666397599476, best_val_loss:0.03633666397599476, best_auc:0.991000214625914\n",
      "Epoch:28 Batch:0 Log-loss0.03391767665743828\n",
      "Epoch:28 Batch:40 Log-loss0.048608291894197464\n",
      "Epoch:28 Batch:80 Log-loss0.04270745441317558\n",
      "Epoch:28 Batch:120 Log-loss0.03446583077311516\n",
      "Epoch:28 Batch:160 Log-loss0.04001234844326973\n",
      "Epoch:28 Batch:200 Log-loss0.02552679367363453\n",
      "Epoch:28 Batch:240 Log-loss0.029403850436210632\n",
      "Epoch:28 Batch:280 Log-loss0.0267080869525671\n",
      "Epoch:28 Batch:320 Log-loss0.03366760164499283\n",
      "Epoch:28 Batch:360 Log-loss0.03755989298224449\n",
      "Epoch:28 Batch:400 Log-loss0.025434399023652077\n",
      "Epoch:28 Batch:440 Log-loss0.03801839426159859\n",
      "Epoch:28 Batch:480 Log-loss0.03964829817414284\n",
      "Epoch:28 Batch:520 Log-loss0.03754177689552307\n",
      "Epoch:28 Batch:560 Log-loss0.03814082592725754\n",
      "Epoch average log loss:0.037597692561601954\n",
      "In Epoch28, val_loss:0.03634371776081106, best_val_loss:0.03633666397599476, best_auc:0.991000214625914\n",
      "Epoch:29 Batch:0 Log-loss0.034195248037576675\n",
      "Epoch:29 Batch:40 Log-loss0.042470455169677734\n",
      "Epoch:29 Batch:80 Log-loss0.0397615060210228\n",
      "Epoch:29 Batch:120 Log-loss0.04871733859181404\n",
      "Epoch:29 Batch:160 Log-loss0.03438761085271835\n",
      "Epoch:29 Batch:200 Log-loss0.0319167785346508\n",
      "Epoch:29 Batch:240 Log-loss0.0251778457313776\n",
      "Epoch:29 Batch:280 Log-loss0.04169158264994621\n",
      "Epoch:29 Batch:320 Log-loss0.02970726229250431\n",
      "Epoch:29 Batch:360 Log-loss0.03333805501461029\n",
      "Epoch:29 Batch:400 Log-loss0.029863806441426277\n",
      "Epoch:29 Batch:440 Log-loss0.03825627639889717\n",
      "Epoch:29 Batch:480 Log-loss0.030297046527266502\n",
      "Epoch:29 Batch:520 Log-loss0.03768713399767876\n",
      "Epoch:29 Batch:560 Log-loss0.03606192767620087\n",
      "Epoch average log loss:0.03731090514588037\n",
      "In Epoch29, val_loss:0.03642986455659979, best_val_loss:0.03633666397599476, best_auc:0.991000214625914\n",
      "Epoch:30 Batch:0 Log-loss0.03719698265194893\n",
      "Epoch:30 Batch:40 Log-loss0.038152649998664856\n",
      "Epoch:30 Batch:80 Log-loss0.02466028742492199\n",
      "Epoch:30 Batch:120 Log-loss0.04051228240132332\n",
      "Epoch:30 Batch:160 Log-loss0.06282468885183334\n",
      "Epoch:30 Batch:200 Log-loss0.02827034331858158\n",
      "Epoch:30 Batch:240 Log-loss0.019250666722655296\n",
      "Epoch:30 Batch:280 Log-loss0.021665168926119804\n",
      "Epoch:30 Batch:320 Log-loss0.04131432995200157\n",
      "Epoch:30 Batch:360 Log-loss0.033020272850990295\n",
      "Epoch:30 Batch:400 Log-loss0.049091797322034836\n",
      "Epoch:30 Batch:440 Log-loss0.030180146917700768\n",
      "Epoch:30 Batch:480 Log-loss0.030071726068854332\n",
      "Epoch:30 Batch:520 Log-loss0.04284058138728142\n",
      "Epoch:30 Batch:560 Log-loss0.038953058421611786\n",
      "Epoch average log loss:0.03728654169410999\n",
      "In Epoch30, val_loss:0.03654276936001021, best_val_loss:0.03633666397599476, best_auc:0.991000214625914\n",
      "Epoch:31 Batch:0 Log-loss0.030518947169184685\n",
      "Epoch:31 Batch:40 Log-loss0.044486407190561295\n",
      "Epoch:31 Batch:80 Log-loss0.037447188049554825\n",
      "Epoch:31 Batch:120 Log-loss0.042931169271469116\n",
      "Epoch:31 Batch:160 Log-loss0.046097349375486374\n",
      "Epoch:31 Batch:200 Log-loss0.02535974234342575\n",
      "Epoch:31 Batch:240 Log-loss0.03509162366390228\n",
      "Epoch:31 Batch:280 Log-loss0.02271062321960926\n",
      "Epoch:31 Batch:320 Log-loss0.03232972323894501\n",
      "Epoch:31 Batch:360 Log-loss0.03358195722103119\n",
      "Epoch:31 Batch:400 Log-loss0.040641866624355316\n",
      "Epoch:31 Batch:440 Log-loss0.028385033831000328\n",
      "Epoch:31 Batch:480 Log-loss0.029309473931789398\n",
      "Epoch:31 Batch:520 Log-loss0.03122629038989544\n",
      "Epoch:31 Batch:560 Log-loss0.03308861702680588\n",
      "Epoch average log loss:0.037326254925158404\n",
      "In Epoch31, val_loss:0.03651767994221627, best_val_loss:0.03633666397599476, best_auc:0.991000214625914\n",
      "Epoch:32 Batch:0 Log-loss0.03049953281879425\n",
      "Epoch:32 Batch:40 Log-loss0.03056311048567295\n",
      "Epoch:32 Batch:80 Log-loss0.0375094935297966\n",
      "Epoch:32 Batch:120 Log-loss0.037101712077856064\n",
      "Epoch:32 Batch:160 Log-loss0.03343871980905533\n",
      "Epoch:32 Batch:200 Log-loss0.03180300444364548\n",
      "Epoch:32 Batch:240 Log-loss0.03307454660534859\n",
      "Epoch:32 Batch:280 Log-loss0.023028669878840446\n",
      "Epoch:32 Batch:320 Log-loss0.03263774514198303\n",
      "Epoch:32 Batch:360 Log-loss0.04265235736966133\n",
      "Epoch:32 Batch:400 Log-loss0.020495368167757988\n",
      "Epoch:32 Batch:440 Log-loss0.031035201624035835\n",
      "Epoch:32 Batch:480 Log-loss0.04176131263375282\n",
      "Epoch:32 Batch:520 Log-loss0.04298768937587738\n",
      "Epoch:32 Batch:560 Log-loss0.04661502316594124\n",
      "Epoch average log loss:0.03714707276750622\n",
      "In Epoch32, val_loss:0.036470302672480394, best_val_loss:0.03633666397599476, best_auc:0.991000214625914\n",
      "Epoch:33 Batch:0 Log-loss0.03548452630639076\n",
      "Epoch:33 Batch:40 Log-loss0.0397958979010582\n",
      "Epoch:33 Batch:80 Log-loss0.04010786861181259\n",
      "Epoch:33 Batch:120 Log-loss0.039589088410139084\n",
      "Epoch:33 Batch:160 Log-loss0.0317985974252224\n",
      "Epoch:33 Batch:200 Log-loss0.03869510814547539\n",
      "Epoch:33 Batch:240 Log-loss0.038690049201250076\n",
      "Epoch:33 Batch:280 Log-loss0.03248303011059761\n",
      "Epoch:33 Batch:320 Log-loss0.029705772176384926\n",
      "Epoch:33 Batch:360 Log-loss0.021687129512429237\n",
      "Epoch:33 Batch:400 Log-loss0.04267691448330879\n",
      "Epoch:33 Batch:440 Log-loss0.03510319069027901\n",
      "Epoch:33 Batch:480 Log-loss0.04524189606308937\n",
      "Epoch:33 Batch:520 Log-loss0.03750737011432648\n",
      "Epoch:33 Batch:560 Log-loss0.03028237819671631\n",
      "Epoch average log loss:0.03711051563067096\n",
      "In Epoch33, val_loss:0.0364845487066399, best_val_loss:0.03633666397599476, best_auc:0.991000214625914\n",
      "Epoch:34 Batch:0 Log-loss0.03942644223570824\n",
      "Epoch:34 Batch:40 Log-loss0.04825478419661522\n",
      "Epoch:34 Batch:80 Log-loss0.034317728132009506\n",
      "Epoch:34 Batch:120 Log-loss0.02157219685614109\n",
      "Epoch:34 Batch:160 Log-loss0.04417578503489494\n",
      "Epoch:34 Batch:200 Log-loss0.03441505506634712\n",
      "Epoch:34 Batch:240 Log-loss0.03971624746918678\n",
      "Epoch:34 Batch:280 Log-loss0.0345669761300087\n",
      "Epoch:34 Batch:320 Log-loss0.03706245869398117\n",
      "Epoch:34 Batch:360 Log-loss0.033735621720552444\n",
      "Epoch:34 Batch:400 Log-loss0.04858924075961113\n",
      "Epoch:34 Batch:440 Log-loss0.04766828194260597\n",
      "Epoch:34 Batch:480 Log-loss0.03791879862546921\n",
      "Epoch:34 Batch:520 Log-loss0.030232930555939674\n",
      "Epoch:34 Batch:560 Log-loss0.040364474058151245\n",
      "Epoch average log loss:0.03701403113414666\n",
      "In Epoch34, val_loss:0.03655818832243745, best_val_loss:0.03633666397599476, best_auc:0.991000214625914\n",
      "Epoch:35 Batch:0 Log-loss0.05198339745402336\n",
      "Epoch:35 Batch:40 Log-loss0.039467643946409225\n",
      "Epoch:35 Batch:80 Log-loss0.026663102209568024\n",
      "Epoch:35 Batch:120 Log-loss0.0373779721558094\n",
      "Epoch:35 Batch:160 Log-loss0.0328947976231575\n",
      "Epoch:35 Batch:200 Log-loss0.0381319597363472\n",
      "Epoch:35 Batch:240 Log-loss0.040760282427072525\n",
      "Epoch:35 Batch:280 Log-loss0.03385462239384651\n",
      "Epoch:35 Batch:320 Log-loss0.0429060123860836\n",
      "Epoch:35 Batch:360 Log-loss0.04227694869041443\n",
      "Epoch:35 Batch:400 Log-loss0.041758447885513306\n",
      "Epoch:35 Batch:440 Log-loss0.03300480172038078\n",
      "Epoch:35 Batch:480 Log-loss0.02616157941520214\n",
      "Epoch:35 Batch:520 Log-loss0.04238637909293175\n",
      "Epoch:35 Batch:560 Log-loss0.04533601179718971\n",
      "Epoch average log loss:0.037070632093985166\n",
      "In Epoch35, val_loss:0.03655125122723163, best_val_loss:0.03633666397599476, best_auc:0.991000214625914\n",
      "Epoch:36 Batch:0 Log-loss0.031355224549770355\n",
      "Epoch:36 Batch:40 Log-loss0.029791107401251793\n",
      "Epoch:36 Batch:80 Log-loss0.04047556594014168\n",
      "Epoch:36 Batch:120 Log-loss0.0337177999317646\n",
      "Epoch:36 Batch:160 Log-loss0.03151476010680199\n",
      "Epoch:36 Batch:200 Log-loss0.031080596148967743\n",
      "Epoch:36 Batch:240 Log-loss0.04592244699597359\n",
      "Epoch:36 Batch:280 Log-loss0.03079073131084442\n",
      "Epoch:36 Batch:320 Log-loss0.038892701268196106\n",
      "Epoch:36 Batch:360 Log-loss0.03379689157009125\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:36 Batch:400 Log-loss0.04799996316432953\n",
      "Epoch:36 Batch:440 Log-loss0.03632586821913719\n",
      "Epoch:36 Batch:480 Log-loss0.032197196036577225\n",
      "Epoch:36 Batch:520 Log-loss0.0477294959127903\n",
      "Epoch:36 Batch:560 Log-loss0.044385313987731934\n",
      "Epoch average log loss:0.037027884367853404\n",
      "In Epoch36, val_loss:0.036511069997664805, best_val_loss:0.03633666397599476, best_auc:0.991000214625914\n",
      "Epoch:37 Batch:0 Log-loss0.033630046993494034\n",
      "Epoch:37 Batch:40 Log-loss0.04341144487261772\n",
      "Epoch:37 Batch:80 Log-loss0.043827518820762634\n",
      "Epoch:37 Batch:120 Log-loss0.03235602378845215\n",
      "Epoch:37 Batch:160 Log-loss0.029233241453766823\n",
      "Epoch:37 Batch:200 Log-loss0.05922232195734978\n",
      "Epoch:37 Batch:240 Log-loss0.040833912789821625\n",
      "Epoch:37 Batch:280 Log-loss0.03064712882041931\n",
      "Epoch:37 Batch:320 Log-loss0.03206590563058853\n",
      "Epoch:37 Batch:360 Log-loss0.0380646176636219\n",
      "Epoch:37 Batch:400 Log-loss0.03685888275504112\n",
      "Epoch:37 Batch:440 Log-loss0.0318370945751667\n",
      "Epoch:37 Batch:480 Log-loss0.03119787573814392\n",
      "Epoch:37 Batch:520 Log-loss0.033530063927173615\n",
      "Epoch:37 Batch:560 Log-loss0.030638350173830986\n",
      "Epoch average log loss:0.037151574323486\n",
      "A pre-trained model at temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt has been loaded.\n",
      "\n",
      "Model has been saved as checkpoints/FASTTXT_BGRU_64_645.pt.\n",
      "\n",
      "Choose the torch base model.\n",
      "Training on fold 6\n",
      "Epoch:1 Batch:0 Log-loss0.713590145111084\n",
      "Epoch:1 Batch:40 Log-loss0.1106889620423317\n",
      "Epoch:1 Batch:80 Log-loss0.10734713077545166\n",
      "Epoch:1 Batch:120 Log-loss0.05269239842891693\n",
      "Epoch:1 Batch:160 Log-loss0.061645712703466415\n",
      "Epoch:1 Batch:200 Log-loss0.04184926673769951\n",
      "Epoch:1 Batch:240 Log-loss0.05214472487568855\n",
      "Epoch:1 Batch:280 Log-loss0.06320098042488098\n",
      "Epoch:1 Batch:320 Log-loss0.06427276134490967\n",
      "Epoch:1 Batch:360 Log-loss0.04161939397454262\n",
      "Epoch:1 Batch:400 Log-loss0.045400649309158325\n",
      "Epoch:1 Batch:440 Log-loss0.05016912892460823\n",
      "Epoch:1 Batch:480 Log-loss0.054204415529966354\n",
      "Epoch:1 Batch:520 Log-loss0.05213052034378052\n",
      "Epoch:1 Batch:560 Log-loss0.05043951794505119\n",
      "Epoch average log loss:0.0852740603465853\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch1, val_loss:0.046820241761913406, best_val_loss:0.046820241761913406, best_auc:0.971250189124116\n",
      "Epoch:2 Batch:0 Log-loss0.045982856303453445\n",
      "Epoch:2 Batch:40 Log-loss0.039628926664590836\n",
      "Epoch:2 Batch:80 Log-loss0.06645912677049637\n",
      "Epoch:2 Batch:120 Log-loss0.05470755323767662\n",
      "Epoch:2 Batch:160 Log-loss0.05234658345580101\n",
      "Epoch:2 Batch:200 Log-loss0.07101254910230637\n",
      "Epoch:2 Batch:240 Log-loss0.05258944630622864\n",
      "Epoch:2 Batch:280 Log-loss0.06393769383430481\n",
      "Epoch:2 Batch:320 Log-loss0.04563343897461891\n",
      "Epoch:2 Batch:360 Log-loss0.02793300338089466\n",
      "Epoch:2 Batch:400 Log-loss0.057922761887311935\n",
      "Epoch:2 Batch:440 Log-loss0.03834327310323715\n",
      "Epoch:2 Batch:480 Log-loss0.04955024644732475\n",
      "Epoch:2 Batch:520 Log-loss0.0388350747525692\n",
      "Epoch:2 Batch:560 Log-loss0.08820310980081558\n",
      "Epoch average log loss:0.05012888586414712\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch2, val_loss:0.04564013739271757, best_val_loss:0.04564013739271757, best_auc:0.9759301352207287\n",
      "Epoch:3 Batch:0 Log-loss0.0512603223323822\n",
      "Epoch:3 Batch:40 Log-loss0.04173871502280235\n",
      "Epoch:3 Batch:80 Log-loss0.049826908856630325\n",
      "Epoch:3 Batch:120 Log-loss0.05173872038722038\n",
      "Epoch:3 Batch:160 Log-loss0.04255513474345207\n",
      "Epoch:3 Batch:200 Log-loss0.05873742699623108\n",
      "Epoch:3 Batch:240 Log-loss0.04764742776751518\n",
      "Epoch:3 Batch:280 Log-loss0.03240246698260307\n",
      "Epoch:3 Batch:320 Log-loss0.053777892142534256\n",
      "Epoch:3 Batch:360 Log-loss0.059904638677835464\n",
      "Epoch:3 Batch:400 Log-loss0.05392204225063324\n",
      "Epoch:3 Batch:440 Log-loss0.05939126014709473\n",
      "Epoch:3 Batch:480 Log-loss0.06824889779090881\n",
      "Epoch:3 Batch:520 Log-loss0.03524573892354965\n",
      "Epoch:3 Batch:560 Log-loss0.052842237055301666\n",
      "Epoch average log loss:0.04787918968192701\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch3, val_loss:0.04278440623796543, best_val_loss:0.04278440623796543, best_auc:0.9788558328019942\n",
      "Epoch:4 Batch:0 Log-loss0.042544323951005936\n",
      "Epoch:4 Batch:40 Log-loss0.04592931270599365\n",
      "Epoch:4 Batch:80 Log-loss0.043889936059713364\n",
      "Epoch:4 Batch:120 Log-loss0.04738106206059456\n",
      "Epoch:4 Batch:160 Log-loss0.03184414282441139\n",
      "Epoch:4 Batch:200 Log-loss0.03580743446946144\n",
      "Epoch:4 Batch:240 Log-loss0.036638177931308746\n",
      "Epoch:4 Batch:280 Log-loss0.04561140760779381\n",
      "Epoch:4 Batch:320 Log-loss0.03674713894724846\n",
      "Epoch:4 Batch:360 Log-loss0.07001523673534393\n",
      "Epoch:4 Batch:400 Log-loss0.04800005629658699\n",
      "Epoch:4 Batch:440 Log-loss0.038442663848400116\n",
      "Epoch:4 Batch:480 Log-loss0.04633896425366402\n",
      "Epoch:4 Batch:520 Log-loss0.048688843846321106\n",
      "Epoch:4 Batch:560 Log-loss0.035890452563762665\n",
      "Epoch average log loss:0.04601062008046678\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch4, val_loss:0.04171983261079501, best_val_loss:0.04171983261079501, best_auc:0.983101779246114\n",
      "Epoch:5 Batch:0 Log-loss0.045562732964754105\n",
      "Epoch:5 Batch:40 Log-loss0.039163216948509216\n",
      "Epoch:5 Batch:80 Log-loss0.027402592822909355\n",
      "Epoch:5 Batch:120 Log-loss0.0633220225572586\n",
      "Epoch:5 Batch:160 Log-loss0.0450931191444397\n",
      "Epoch:5 Batch:200 Log-loss0.04401865229010582\n",
      "Epoch:5 Batch:240 Log-loss0.043939754366874695\n",
      "Epoch:5 Batch:280 Log-loss0.04808100685477257\n",
      "Epoch:5 Batch:320 Log-loss0.06496333330869675\n",
      "Epoch:5 Batch:360 Log-loss0.06231484189629555\n",
      "Epoch:5 Batch:400 Log-loss0.03696885704994202\n",
      "Epoch:5 Batch:440 Log-loss0.0496949702501297\n",
      "Epoch:5 Batch:480 Log-loss0.05220711603760719\n",
      "Epoch:5 Batch:520 Log-loss0.03883444145321846\n",
      "Epoch:5 Batch:560 Log-loss0.039950381964445114\n",
      "Epoch average log loss:0.04446216811026846\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch5, val_loss:0.040750533456780126, best_val_loss:0.040750533456780126, best_auc:0.9835224113513226\n",
      "Epoch:6 Batch:0 Log-loss0.04405249282717705\n",
      "Epoch:6 Batch:40 Log-loss0.053006645292043686\n",
      "Epoch:6 Batch:80 Log-loss0.05237704887986183\n",
      "Epoch:6 Batch:120 Log-loss0.04362265765666962\n",
      "Epoch:6 Batch:160 Log-loss0.04196050390601158\n",
      "Epoch:6 Batch:200 Log-loss0.048561688512563705\n",
      "Epoch:6 Batch:240 Log-loss0.026086358353495598\n",
      "Epoch:6 Batch:280 Log-loss0.03368267044425011\n",
      "Epoch:6 Batch:320 Log-loss0.05708487331867218\n",
      "Epoch:6 Batch:360 Log-loss0.044283851981163025\n",
      "Epoch:6 Batch:400 Log-loss0.035827573388814926\n",
      "Epoch:6 Batch:440 Log-loss0.023172477260231972\n",
      "Epoch:6 Batch:480 Log-loss0.05440214276313782\n",
      "Epoch:6 Batch:520 Log-loss0.045036304742097855\n",
      "Epoch:6 Batch:560 Log-loss0.04088575392961502\n",
      "Epoch average log loss:0.043417578366851166\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch6, val_loss:0.0397565278251075, best_val_loss:0.0397565278251075, best_auc:0.985143615766725\n",
      "Epoch:7 Batch:0 Log-loss0.032453227788209915\n",
      "Epoch:7 Batch:40 Log-loss0.04105307534337044\n",
      "Epoch:7 Batch:80 Log-loss0.04701529070734978\n",
      "Epoch:7 Batch:120 Log-loss0.03594407066702843\n",
      "Epoch:7 Batch:160 Log-loss0.046607855707407\n",
      "Epoch:7 Batch:200 Log-loss0.04215447977185249\n",
      "Epoch:7 Batch:240 Log-loss0.03471658006310463\n",
      "Epoch:7 Batch:280 Log-loss0.0408037044107914\n",
      "Epoch:7 Batch:320 Log-loss0.032285068184137344\n",
      "Epoch:7 Batch:360 Log-loss0.03711121156811714\n",
      "Epoch:7 Batch:400 Log-loss0.05280591547489166\n",
      "Epoch:7 Batch:440 Log-loss0.03870799392461777\n",
      "Epoch:7 Batch:480 Log-loss0.04013942554593086\n",
      "Epoch:7 Batch:520 Log-loss0.041863132268190384\n",
      "Epoch:7 Batch:560 Log-loss0.04663344845175743\n",
      "Epoch average log loss:0.042599252202281994\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch7, val_loss:0.039316828137268535, best_val_loss:0.039316828137268535, best_auc:0.9867531044147252\n",
      "Epoch:8 Batch:0 Log-loss0.03628592565655708\n",
      "Epoch:8 Batch:40 Log-loss0.04624120891094208\n",
      "Epoch:8 Batch:80 Log-loss0.03614461049437523\n",
      "Epoch:8 Batch:120 Log-loss0.0421283096075058\n",
      "Epoch:8 Batch:160 Log-loss0.04939033091068268\n",
      "Epoch:8 Batch:200 Log-loss0.037239935249090195\n",
      "Epoch:8 Batch:240 Log-loss0.04322133585810661\n",
      "Epoch:8 Batch:280 Log-loss0.04245128110051155\n",
      "Epoch:8 Batch:320 Log-loss0.024212362244725227\n",
      "Epoch:8 Batch:360 Log-loss0.037152696400880814\n",
      "Epoch:8 Batch:400 Log-loss0.055634260177612305\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:8 Batch:440 Log-loss0.026999376714229584\n",
      "Epoch:8 Batch:480 Log-loss0.04400629922747612\n",
      "Epoch:8 Batch:520 Log-loss0.03930778056383133\n",
      "Epoch:8 Batch:560 Log-loss0.059055302292108536\n",
      "Epoch average log loss:0.04195517084694334\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch8, val_loss:0.038943949354370344, best_val_loss:0.038943949354370344, best_auc:0.9869054354742267\n",
      "Epoch:9 Batch:0 Log-loss0.04157564416527748\n",
      "Epoch:9 Batch:40 Log-loss0.042970944195985794\n",
      "Epoch:9 Batch:80 Log-loss0.04248839244246483\n",
      "Epoch:9 Batch:120 Log-loss0.04621853306889534\n",
      "Epoch:9 Batch:160 Log-loss0.035649366676807404\n",
      "Epoch:9 Batch:200 Log-loss0.03273468837141991\n",
      "Epoch:9 Batch:240 Log-loss0.045425593852996826\n",
      "Epoch:9 Batch:280 Log-loss0.05068336799740791\n",
      "Epoch:9 Batch:320 Log-loss0.044638875871896744\n",
      "Epoch:9 Batch:360 Log-loss0.04561986029148102\n",
      "Epoch:9 Batch:400 Log-loss0.036621466279029846\n",
      "Epoch:9 Batch:440 Log-loss0.04304751753807068\n",
      "Epoch:9 Batch:480 Log-loss0.04791615903377533\n",
      "Epoch:9 Batch:520 Log-loss0.044932618737220764\n",
      "Epoch:9 Batch:560 Log-loss0.02960882894694805\n",
      "Epoch average log loss:0.04123062624795629\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch9, val_loss:0.0386645393713742, best_val_loss:0.0386645393713742, best_auc:0.987806784041319\n",
      "Epoch:10 Batch:0 Log-loss0.04057114198803902\n",
      "Epoch:10 Batch:40 Log-loss0.04058771952986717\n",
      "Epoch:10 Batch:80 Log-loss0.047433316707611084\n",
      "Epoch:10 Batch:120 Log-loss0.04606245830655098\n",
      "Epoch:10 Batch:160 Log-loss0.03826585039496422\n",
      "Epoch:10 Batch:200 Log-loss0.034447405487298965\n",
      "Epoch:10 Batch:240 Log-loss0.03917120397090912\n",
      "Epoch:10 Batch:280 Log-loss0.03392056003212929\n",
      "Epoch:10 Batch:320 Log-loss0.024944014847278595\n",
      "Epoch:10 Batch:360 Log-loss0.04428640380501747\n",
      "Epoch:10 Batch:400 Log-loss0.0424661822617054\n",
      "Epoch:10 Batch:440 Log-loss0.03647056221961975\n",
      "Epoch:10 Batch:480 Log-loss0.0470363087952137\n",
      "Epoch:10 Batch:520 Log-loss0.03193393722176552\n",
      "Epoch:10 Batch:560 Log-loss0.05164766684174538\n",
      "Epoch average log loss:0.04082497285812029\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch10, val_loss:0.03816732740310581, best_val_loss:0.03816732740310581, best_auc:0.9881323039760205\n",
      "Epoch:11 Batch:0 Log-loss0.04076855629682541\n",
      "Epoch:11 Batch:40 Log-loss0.030017873272299767\n",
      "Epoch:11 Batch:80 Log-loss0.025302261114120483\n",
      "Epoch:11 Batch:120 Log-loss0.04854839667677879\n",
      "Epoch:11 Batch:160 Log-loss0.05927586928009987\n",
      "Epoch:11 Batch:200 Log-loss0.05657466873526573\n",
      "Epoch:11 Batch:240 Log-loss0.036243803799152374\n",
      "Epoch:11 Batch:280 Log-loss0.050344616174697876\n",
      "Epoch:11 Batch:320 Log-loss0.02867060899734497\n",
      "Epoch:11 Batch:360 Log-loss0.027705712243914604\n",
      "Epoch:11 Batch:400 Log-loss0.03910788893699646\n",
      "Epoch:11 Batch:440 Log-loss0.03365608677268028\n",
      "Epoch:11 Batch:480 Log-loss0.026191385462880135\n",
      "Epoch:11 Batch:520 Log-loss0.04815046116709709\n",
      "Epoch:11 Batch:560 Log-loss0.033823128789663315\n",
      "Epoch average log loss:0.040282041426481945\n",
      "In Epoch11, val_loss:0.0381899169131941, best_val_loss:0.03816732740310581, best_auc:0.9881323039760205\n",
      "Epoch:12 Batch:0 Log-loss0.033492155373096466\n",
      "Epoch:12 Batch:40 Log-loss0.0576479397714138\n",
      "Epoch:12 Batch:80 Log-loss0.026296230033040047\n",
      "Epoch:12 Batch:120 Log-loss0.04726553335785866\n",
      "Epoch:12 Batch:160 Log-loss0.05051371827721596\n",
      "Epoch:12 Batch:200 Log-loss0.04665575549006462\n",
      "Epoch:12 Batch:240 Log-loss0.04342197999358177\n",
      "Epoch:12 Batch:280 Log-loss0.02733391337096691\n",
      "Epoch:12 Batch:320 Log-loss0.03330443426966667\n",
      "Epoch:12 Batch:360 Log-loss0.043009355664253235\n",
      "Epoch:12 Batch:400 Log-loss0.04419324919581413\n",
      "Epoch:12 Batch:440 Log-loss0.034859638661146164\n",
      "Epoch:12 Batch:480 Log-loss0.03610836714506149\n",
      "Epoch:12 Batch:520 Log-loss0.04555356502532959\n",
      "Epoch:12 Batch:560 Log-loss0.034884434193372726\n",
      "Epoch average log loss:0.03981817494745233\n",
      "In Epoch12, val_loss:0.038535847183691026, best_val_loss:0.03816732740310581, best_auc:0.9881323039760205\n",
      "Epoch:13 Batch:0 Log-loss0.04316377267241478\n",
      "Epoch:13 Batch:40 Log-loss0.04492279514670372\n",
      "Epoch:13 Batch:80 Log-loss0.030143002048134804\n",
      "Epoch:13 Batch:120 Log-loss0.025759706273674965\n",
      "Epoch:13 Batch:160 Log-loss0.04975590482354164\n",
      "Epoch:13 Batch:200 Log-loss0.04259270802140236\n",
      "Epoch:13 Batch:240 Log-loss0.047845322638750076\n",
      "Epoch:13 Batch:280 Log-loss0.034417737275362015\n",
      "Epoch:13 Batch:320 Log-loss0.04064033553004265\n",
      "Epoch:13 Batch:360 Log-loss0.0338069312274456\n",
      "Epoch:13 Batch:400 Log-loss0.03320325165987015\n",
      "Epoch:13 Batch:440 Log-loss0.02748788706958294\n",
      "Epoch:13 Batch:480 Log-loss0.0487881638109684\n",
      "Epoch:13 Batch:520 Log-loss0.03236749768257141\n",
      "Epoch:13 Batch:560 Log-loss0.03820357099175453\n",
      "Epoch average log loss:0.039724683099692425\n",
      "In Epoch13, val_loss:0.03856137321448395, best_val_loss:0.03816732740310581, best_auc:0.9881323039760205\n",
      "Epoch:14 Batch:0 Log-loss0.034695159643888474\n",
      "Epoch:14 Batch:40 Log-loss0.03557319566607475\n",
      "Epoch:14 Batch:80 Log-loss0.035107869654893875\n",
      "Epoch:14 Batch:120 Log-loss0.035425178706645966\n",
      "Epoch:14 Batch:160 Log-loss0.032241810113191605\n",
      "Epoch:14 Batch:200 Log-loss0.04519977793097496\n",
      "Epoch:14 Batch:240 Log-loss0.04095618799328804\n",
      "Epoch:14 Batch:280 Log-loss0.04275545850396156\n",
      "Epoch:14 Batch:320 Log-loss0.036187995225191116\n",
      "Epoch:14 Batch:360 Log-loss0.026440659537911415\n",
      "Epoch:14 Batch:400 Log-loss0.03263320401310921\n",
      "Epoch:14 Batch:440 Log-loss0.039571940898895264\n",
      "Epoch:14 Batch:480 Log-loss0.04296175017952919\n",
      "Epoch:14 Batch:520 Log-loss0.03737578168511391\n",
      "Epoch:14 Batch:560 Log-loss0.0470861941576004\n",
      "Epoch average log loss:0.03939348565348025\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch14, val_loss:0.038149509922406934, best_val_loss:0.038149509922406934, best_auc:0.9885246273214623\n",
      "Epoch:15 Batch:0 Log-loss0.026786280795931816\n",
      "Epoch:15 Batch:40 Log-loss0.0497170090675354\n",
      "Epoch:15 Batch:80 Log-loss0.03435824438929558\n",
      "Epoch:15 Batch:120 Log-loss0.03141169995069504\n",
      "Epoch:15 Batch:160 Log-loss0.048113349825143814\n",
      "Epoch:15 Batch:200 Log-loss0.0396488681435585\n",
      "Epoch:15 Batch:240 Log-loss0.04640388488769531\n",
      "Epoch:15 Batch:280 Log-loss0.03322567045688629\n",
      "Epoch:15 Batch:320 Log-loss0.035667795687913895\n",
      "Epoch:15 Batch:360 Log-loss0.04897800087928772\n",
      "Epoch:15 Batch:400 Log-loss0.04658129811286926\n",
      "Epoch:15 Batch:440 Log-loss0.047558289021253586\n",
      "Epoch:15 Batch:480 Log-loss0.04271900653839111\n",
      "Epoch:15 Batch:520 Log-loss0.05547885596752167\n",
      "Epoch:15 Batch:560 Log-loss0.030163343995809555\n",
      "Epoch average log loss:0.03905311250793082\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch15, val_loss:0.038025336946415605, best_val_loss:0.038025336946415605, best_auc:0.9889229674990642\n",
      "Epoch:16 Batch:0 Log-loss0.037411805242300034\n",
      "Epoch:16 Batch:40 Log-loss0.044118624180555344\n",
      "Epoch:16 Batch:80 Log-loss0.03267110139131546\n",
      "Epoch:16 Batch:120 Log-loss0.04392203688621521\n",
      "Epoch:16 Batch:160 Log-loss0.04581620171666145\n",
      "Epoch:16 Batch:200 Log-loss0.03727145120501518\n",
      "Epoch:16 Batch:240 Log-loss0.03445836529135704\n",
      "Epoch:16 Batch:280 Log-loss0.04926136136054993\n",
      "Epoch:16 Batch:320 Log-loss0.035294029861688614\n",
      "Epoch:16 Batch:360 Log-loss0.030959973111748695\n",
      "Epoch:16 Batch:400 Log-loss0.03886278346180916\n",
      "Epoch:16 Batch:440 Log-loss0.03647885471582413\n",
      "Epoch:16 Batch:480 Log-loss0.042462367564439774\n",
      "Epoch:16 Batch:520 Log-loss0.04334445670247078\n",
      "Epoch:16 Batch:560 Log-loss0.033805858343839645\n",
      "Epoch average log loss:0.03891468540698822\n",
      "In Epoch16, val_loss:0.038172166002250624, best_val_loss:0.038025336946415605, best_auc:0.9889229674990642\n",
      "Epoch:17 Batch:0 Log-loss0.029925191774964333\n",
      "Epoch:17 Batch:40 Log-loss0.04069669917225838\n",
      "Epoch:17 Batch:80 Log-loss0.04246028885245323\n",
      "Epoch:17 Batch:120 Log-loss0.03629284352064133\n",
      "Epoch:17 Batch:160 Log-loss0.035760797560214996\n",
      "Epoch:17 Batch:200 Log-loss0.03541465848684311\n",
      "Epoch:17 Batch:240 Log-loss0.0392024926841259\n",
      "Epoch:17 Batch:280 Log-loss0.03931111469864845\n",
      "Epoch:17 Batch:320 Log-loss0.03742967173457146\n",
      "Epoch:17 Batch:360 Log-loss0.055017244070768356\n",
      "Epoch:17 Batch:400 Log-loss0.03593927249312401\n",
      "Epoch:17 Batch:440 Log-loss0.034947458654642105\n",
      "Epoch:17 Batch:480 Log-loss0.03988403454422951\n",
      "Epoch:17 Batch:520 Log-loss0.036000605672597885\n",
      "Epoch:17 Batch:560 Log-loss0.041687771677970886\n",
      "Epoch average log loss:0.03879740388199155\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In Epoch17, val_loss:0.03804959634552036, best_val_loss:0.038025336946415605, best_auc:0.9889229674990642\n",
      "Epoch:18 Batch:0 Log-loss0.05093863606452942\n",
      "Epoch:18 Batch:40 Log-loss0.03569447249174118\n",
      "Epoch:18 Batch:80 Log-loss0.030999116599559784\n",
      "Epoch:18 Batch:120 Log-loss0.043984126299619675\n",
      "Epoch:18 Batch:160 Log-loss0.04073851928114891\n",
      "Epoch:18 Batch:200 Log-loss0.0501139871776104\n",
      "Epoch:18 Batch:240 Log-loss0.04075596109032631\n",
      "Epoch:18 Batch:280 Log-loss0.045665789395570755\n",
      "Epoch:18 Batch:320 Log-loss0.043532561510801315\n",
      "Epoch:18 Batch:360 Log-loss0.03232383355498314\n",
      "Epoch:18 Batch:400 Log-loss0.02937909960746765\n",
      "Epoch:18 Batch:440 Log-loss0.030361143872141838\n",
      "Epoch:18 Batch:480 Log-loss0.06526514887809753\n",
      "Epoch:18 Batch:520 Log-loss0.04856736585497856\n",
      "Epoch:18 Batch:560 Log-loss0.04805982857942581\n",
      "Epoch average log loss:0.038626848886321695\n",
      "In Epoch18, val_loss:0.03809442507545027, best_val_loss:0.038025336946415605, best_auc:0.9889229674990642\n",
      "Epoch:19 Batch:0 Log-loss0.03569434955716133\n",
      "Epoch:19 Batch:40 Log-loss0.03236732259392738\n",
      "Epoch:19 Batch:80 Log-loss0.04453149065375328\n",
      "Epoch:19 Batch:120 Log-loss0.03993495926260948\n",
      "Epoch:19 Batch:160 Log-loss0.02994418703019619\n",
      "Epoch:19 Batch:200 Log-loss0.03356366232037544\n",
      "Epoch:19 Batch:240 Log-loss0.044321347028017044\n",
      "Epoch:19 Batch:280 Log-loss0.036475770175457\n",
      "Epoch:19 Batch:320 Log-loss0.02471954934298992\n",
      "Epoch:19 Batch:360 Log-loss0.05248284339904785\n",
      "Epoch:19 Batch:400 Log-loss0.03834311664104462\n",
      "Epoch:19 Batch:440 Log-loss0.050211336463689804\n",
      "Epoch:19 Batch:480 Log-loss0.04595514014363289\n",
      "Epoch:19 Batch:520 Log-loss0.0401889868080616\n",
      "Epoch:19 Batch:560 Log-loss0.03444482013583183\n",
      "Epoch average log loss:0.03852738877259461\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch19, val_loss:0.03772151947195298, best_val_loss:0.03772151947195298, best_auc:0.9891580811088919\n",
      "Epoch:20 Batch:0 Log-loss0.03910957649350166\n",
      "Epoch:20 Batch:40 Log-loss0.03109159506857395\n",
      "Epoch:20 Batch:80 Log-loss0.04258481785655022\n",
      "Epoch:20 Batch:120 Log-loss0.030752966180443764\n",
      "Epoch:20 Batch:160 Log-loss0.04507578909397125\n",
      "Epoch:20 Batch:200 Log-loss0.03114222176373005\n",
      "Epoch:20 Batch:240 Log-loss0.0459633432328701\n",
      "Epoch:20 Batch:280 Log-loss0.02589339204132557\n",
      "Epoch:20 Batch:320 Log-loss0.045983437448740005\n",
      "Epoch:20 Batch:360 Log-loss0.04764269292354584\n",
      "Epoch:20 Batch:400 Log-loss0.03903781250119209\n",
      "Epoch:20 Batch:440 Log-loss0.04937669262290001\n",
      "Epoch:20 Batch:480 Log-loss0.03509402275085449\n",
      "Epoch:20 Batch:520 Log-loss0.0352356843650341\n",
      "Epoch:20 Batch:560 Log-loss0.03234964981675148\n",
      "Epoch average log loss:0.03838847035741699\n",
      "In Epoch20, val_loss:0.038005053326112107, best_val_loss:0.03772151947195298, best_auc:0.9891580811088919\n",
      "Epoch:21 Batch:0 Log-loss0.03158784285187721\n",
      "Epoch:21 Batch:40 Log-loss0.04589429125189781\n",
      "Epoch:21 Batch:80 Log-loss0.045437220484018326\n",
      "Epoch:21 Batch:120 Log-loss0.03050200082361698\n",
      "Epoch:21 Batch:160 Log-loss0.019198445603251457\n",
      "Epoch:21 Batch:200 Log-loss0.03246476873755455\n",
      "Epoch:21 Batch:240 Log-loss0.04273867607116699\n",
      "Epoch:21 Batch:280 Log-loss0.04488910734653473\n",
      "Epoch:21 Batch:320 Log-loss0.02701353281736374\n",
      "Epoch:21 Batch:360 Log-loss0.03907910734415054\n",
      "Epoch:21 Batch:400 Log-loss0.030871383845806122\n",
      "Epoch:21 Batch:440 Log-loss0.03590395674109459\n",
      "Epoch:21 Batch:480 Log-loss0.03835814446210861\n",
      "Epoch:21 Batch:520 Log-loss0.025447160005569458\n",
      "Epoch:21 Batch:560 Log-loss0.05452282354235649\n",
      "Epoch average log loss:0.03810697584412992\n",
      "In Epoch21, val_loss:0.03777641324594276, best_val_loss:0.03772151947195298, best_auc:0.9891580811088919\n",
      "Epoch:22 Batch:0 Log-loss0.03976566344499588\n",
      "Epoch:22 Batch:40 Log-loss0.04194571450352669\n",
      "Epoch:22 Batch:80 Log-loss0.0380387008190155\n",
      "Epoch:22 Batch:120 Log-loss0.04258621111512184\n",
      "Epoch:22 Batch:160 Log-loss0.03659479692578316\n",
      "Epoch:22 Batch:200 Log-loss0.024376576766371727\n",
      "Epoch:22 Batch:240 Log-loss0.03918742388486862\n",
      "Epoch:22 Batch:280 Log-loss0.04734891280531883\n",
      "Epoch:22 Batch:320 Log-loss0.038029611110687256\n",
      "Epoch:22 Batch:360 Log-loss0.03242378681898117\n",
      "Epoch:22 Batch:400 Log-loss0.0395071767270565\n",
      "Epoch:22 Batch:440 Log-loss0.024042969569563866\n",
      "Epoch:22 Batch:480 Log-loss0.04842539504170418\n",
      "Epoch:22 Batch:520 Log-loss0.049196597188711166\n",
      "Epoch:22 Batch:560 Log-loss0.034882497042417526\n",
      "Epoch average log loss:0.03807137168145606\n",
      "In Epoch22, val_loss:0.03790738471407363, best_val_loss:0.03772151947195298, best_auc:0.9891580811088919\n",
      "Epoch:23 Batch:0 Log-loss0.021443037316203117\n",
      "Epoch:23 Batch:40 Log-loss0.01957053318619728\n",
      "Epoch:23 Batch:80 Log-loss0.03797762468457222\n",
      "Epoch:23 Batch:120 Log-loss0.048906613141298294\n",
      "Epoch:23 Batch:160 Log-loss0.05211944505572319\n",
      "Epoch:23 Batch:200 Log-loss0.03918716683983803\n",
      "Epoch:23 Batch:240 Log-loss0.04358925297856331\n",
      "Epoch:23 Batch:280 Log-loss0.03732926398515701\n",
      "Epoch:23 Batch:320 Log-loss0.03473920747637749\n",
      "Epoch:23 Batch:360 Log-loss0.037923943251371384\n",
      "Epoch:23 Batch:400 Log-loss0.048091962933540344\n",
      "Epoch:23 Batch:440 Log-loss0.027429722249507904\n",
      "Epoch:23 Batch:480 Log-loss0.03909976780414581\n",
      "Epoch:23 Batch:520 Log-loss0.05630718544125557\n",
      "Epoch:23 Batch:560 Log-loss0.0572243332862854\n",
      "Epoch average log loss:0.03798077987906124\n",
      "In Epoch23, val_loss:0.03778842842162882, best_val_loss:0.03772151947195298, best_auc:0.9891580811088919\n",
      "Epoch:24 Batch:0 Log-loss0.0385592021048069\n",
      "Epoch:24 Batch:40 Log-loss0.04380302131175995\n",
      "Epoch:24 Batch:80 Log-loss0.03080306388437748\n",
      "Epoch:24 Batch:120 Log-loss0.04823024943470955\n",
      "Epoch:24 Batch:160 Log-loss0.03069177269935608\n",
      "Epoch:24 Batch:200 Log-loss0.03658636286854744\n",
      "Epoch:24 Batch:240 Log-loss0.04174543544650078\n",
      "Epoch:24 Batch:280 Log-loss0.03524552285671234\n",
      "Epoch:24 Batch:320 Log-loss0.044319748878479004\n",
      "Epoch:24 Batch:360 Log-loss0.04036777839064598\n",
      "Epoch:24 Batch:400 Log-loss0.03591692075133324\n",
      "Epoch:24 Batch:440 Log-loss0.020940026268363\n",
      "Epoch:24 Batch:480 Log-loss0.04673019424080849\n",
      "Epoch:24 Batch:520 Log-loss0.022342277690768242\n",
      "Epoch:24 Batch:560 Log-loss0.03843708708882332\n",
      "Epoch average log loss:0.037813038126166376\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch24, val_loss:0.03771431408885065, best_val_loss:0.03771431408885065, best_auc:0.9892261923567189\n",
      "Epoch:25 Batch:0 Log-loss0.03769844397902489\n",
      "Epoch:25 Batch:40 Log-loss0.03784884512424469\n",
      "Epoch:25 Batch:80 Log-loss0.027709836140275\n",
      "Epoch:25 Batch:120 Log-loss0.04009446129202843\n",
      "Epoch:25 Batch:160 Log-loss0.04067336395382881\n",
      "Epoch:25 Batch:200 Log-loss0.024703996255993843\n",
      "Epoch:25 Batch:240 Log-loss0.025807367637753487\n",
      "Epoch:25 Batch:280 Log-loss0.0522785522043705\n",
      "Epoch:25 Batch:320 Log-loss0.03057892806828022\n",
      "Epoch:25 Batch:360 Log-loss0.033298492431640625\n",
      "Epoch:25 Batch:400 Log-loss0.04777484014630318\n",
      "Epoch:25 Batch:440 Log-loss0.03494328260421753\n",
      "Epoch:25 Batch:480 Log-loss0.036174748092889786\n",
      "Epoch:25 Batch:520 Log-loss0.03840385377407074\n",
      "Epoch:25 Batch:560 Log-loss0.05127512291073799\n",
      "Epoch average log loss:0.03776453754731587\n",
      "In Epoch25, val_loss:0.037978108010855306, best_val_loss:0.03771431408885065, best_auc:0.9892261923567189\n",
      "Epoch:26 Batch:0 Log-loss0.045622508972883224\n",
      "Epoch:26 Batch:40 Log-loss0.03550448268651962\n",
      "Epoch:26 Batch:80 Log-loss0.029225900769233704\n",
      "Epoch:26 Batch:120 Log-loss0.0301214549690485\n",
      "Epoch:26 Batch:160 Log-loss0.056276511400938034\n",
      "Epoch:26 Batch:200 Log-loss0.047982607036828995\n",
      "Epoch:26 Batch:240 Log-loss0.03418942168354988\n",
      "Epoch:26 Batch:280 Log-loss0.029364174231886864\n",
      "Epoch:26 Batch:320 Log-loss0.031256671994924545\n",
      "Epoch:26 Batch:360 Log-loss0.03388986364006996\n",
      "Epoch:26 Batch:400 Log-loss0.04327797517180443\n",
      "Epoch:26 Batch:440 Log-loss0.03820193186402321\n",
      "Epoch:26 Batch:480 Log-loss0.0327107310295105\n",
      "Epoch:26 Batch:520 Log-loss0.06550521403551102\n",
      "Epoch:26 Batch:560 Log-loss0.041553620249032974\n",
      "Epoch average log loss:0.03756584736984223\n",
      "In Epoch26, val_loss:0.03772098855767187, best_val_loss:0.03771431408885065, best_auc:0.9892261923567189\n",
      "Epoch:27 Batch:0 Log-loss0.033234164118766785\n",
      "Epoch:27 Batch:40 Log-loss0.04181133583188057\n",
      "Epoch:27 Batch:80 Log-loss0.047862157225608826\n",
      "Epoch:27 Batch:120 Log-loss0.03642099350690842\n",
      "Epoch:27 Batch:160 Log-loss0.04510955885052681\n",
      "Epoch:27 Batch:200 Log-loss0.031799837946891785\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:27 Batch:240 Log-loss0.03731074556708336\n",
      "Epoch:27 Batch:280 Log-loss0.03224267438054085\n",
      "Epoch:27 Batch:320 Log-loss0.019776832312345505\n",
      "Epoch:27 Batch:360 Log-loss0.047354597598314285\n",
      "Epoch:27 Batch:400 Log-loss0.022550620138645172\n",
      "Epoch:27 Batch:440 Log-loss0.032220106571912766\n",
      "Epoch:27 Batch:480 Log-loss0.03721199184656143\n",
      "Epoch:27 Batch:520 Log-loss0.042260248214006424\n",
      "Epoch:27 Batch:560 Log-loss0.02921636588871479\n",
      "Epoch average log loss:0.03739128584108715\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch27, val_loss:0.037555375900341165, best_val_loss:0.037555375900341165, best_auc:0.9894698424386442\n",
      "Epoch:28 Batch:0 Log-loss0.03778574988245964\n",
      "Epoch:28 Batch:40 Log-loss0.03254866227507591\n",
      "Epoch:28 Batch:80 Log-loss0.04492509365081787\n",
      "Epoch:28 Batch:120 Log-loss0.03577310964465141\n",
      "Epoch:28 Batch:160 Log-loss0.03733382001519203\n",
      "Epoch:28 Batch:200 Log-loss0.0500914566218853\n",
      "Epoch:28 Batch:240 Log-loss0.024008147418498993\n",
      "Epoch:28 Batch:280 Log-loss0.029459349811077118\n",
      "Epoch:28 Batch:320 Log-loss0.038698721677064896\n",
      "Epoch:28 Batch:360 Log-loss0.035206567496061325\n",
      "Epoch:28 Batch:400 Log-loss0.04279014468193054\n",
      "Epoch:28 Batch:440 Log-loss0.05540575459599495\n",
      "Epoch:28 Batch:480 Log-loss0.04033886268734932\n",
      "Epoch:28 Batch:520 Log-loss0.02529505081474781\n",
      "Epoch:28 Batch:560 Log-loss0.030648251995444298\n",
      "Epoch average log loss:0.03743311275861093\n",
      "In Epoch28, val_loss:0.03771825281201019, best_val_loss:0.037555375900341165, best_auc:0.9894698424386442\n",
      "Epoch:29 Batch:0 Log-loss0.0466795451939106\n",
      "Epoch:29 Batch:40 Log-loss0.033533595502376556\n",
      "Epoch:29 Batch:80 Log-loss0.03435247018933296\n",
      "Epoch:29 Batch:120 Log-loss0.038672395050525665\n",
      "Epoch:29 Batch:160 Log-loss0.03730140998959541\n",
      "Epoch:29 Batch:200 Log-loss0.04359757527709007\n",
      "Epoch:29 Batch:240 Log-loss0.03215237334370613\n",
      "Epoch:29 Batch:280 Log-loss0.04442046210169792\n",
      "Epoch:29 Batch:320 Log-loss0.038238491863012314\n",
      "Epoch:29 Batch:360 Log-loss0.03745714947581291\n",
      "Epoch:29 Batch:400 Log-loss0.039512164890766144\n",
      "Epoch:29 Batch:440 Log-loss0.040831491351127625\n",
      "Epoch:29 Batch:480 Log-loss0.03355679288506508\n",
      "Epoch:29 Batch:520 Log-loss0.04843128100037575\n",
      "Epoch:29 Batch:560 Log-loss0.03092803619801998\n",
      "Epoch average log loss:0.037300980463624\n",
      "In Epoch29, val_loss:0.03771452481318651, best_val_loss:0.037555375900341165, best_auc:0.9894698424386442\n",
      "Epoch:30 Batch:0 Log-loss0.02644296921789646\n",
      "Epoch:30 Batch:40 Log-loss0.03449507802724838\n",
      "Epoch:30 Batch:80 Log-loss0.03022320382297039\n",
      "Epoch:30 Batch:120 Log-loss0.037109728902578354\n",
      "Epoch:30 Batch:160 Log-loss0.03640471771359444\n",
      "Epoch:30 Batch:200 Log-loss0.03178868815302849\n",
      "Epoch:30 Batch:240 Log-loss0.04053446277976036\n",
      "Epoch:30 Batch:280 Log-loss0.02931044064462185\n",
      "Epoch:30 Batch:320 Log-loss0.019387319684028625\n",
      "Epoch:30 Batch:360 Log-loss0.0363294817507267\n",
      "Epoch:30 Batch:400 Log-loss0.02461492456495762\n",
      "Epoch:30 Batch:440 Log-loss0.03324706107378006\n",
      "Epoch:30 Batch:480 Log-loss0.0457061231136322\n",
      "Epoch:30 Batch:520 Log-loss0.05551464855670929\n",
      "Epoch:30 Batch:560 Log-loss0.03095335140824318\n",
      "Epoch average log loss:0.03713758735019448\n",
      "In Epoch30, val_loss:0.0377749771026785, best_val_loss:0.037555375900341165, best_auc:0.9894698424386442\n",
      "Epoch:31 Batch:0 Log-loss0.03083851747214794\n",
      "Epoch:31 Batch:40 Log-loss0.02654462307691574\n",
      "Epoch:31 Batch:80 Log-loss0.03750127926468849\n",
      "Epoch:31 Batch:120 Log-loss0.02796895056962967\n",
      "Epoch:31 Batch:160 Log-loss0.02037198655307293\n",
      "Epoch:31 Batch:200 Log-loss0.028169387951493263\n",
      "Epoch:31 Batch:240 Log-loss0.03898784518241882\n",
      "Epoch:31 Batch:280 Log-loss0.045826490968465805\n",
      "Epoch:31 Batch:320 Log-loss0.040080584585666656\n",
      "Epoch:31 Batch:360 Log-loss0.04244023934006691\n",
      "Epoch:31 Batch:400 Log-loss0.03367849439382553\n",
      "Epoch:31 Batch:440 Log-loss0.02948400378227234\n",
      "Epoch:31 Batch:480 Log-loss0.041485466063022614\n",
      "Epoch:31 Batch:520 Log-loss0.03910782188177109\n",
      "Epoch:31 Batch:560 Log-loss0.052070632576942444\n",
      "Epoch average log loss:0.037139952388991204\n",
      "In Epoch31, val_loss:0.0378377744370578, best_val_loss:0.037555375900341165, best_auc:0.9894698424386442\n",
      "Epoch:32 Batch:0 Log-loss0.0402878075838089\n",
      "Epoch:32 Batch:40 Log-loss0.02944844402372837\n",
      "Epoch:32 Batch:80 Log-loss0.039866853505373\n",
      "Epoch:32 Batch:120 Log-loss0.03710690885782242\n",
      "Epoch:32 Batch:160 Log-loss0.045365024358034134\n",
      "Epoch:32 Batch:200 Log-loss0.02814488299190998\n",
      "Epoch:32 Batch:240 Log-loss0.03718370944261551\n",
      "Epoch:32 Batch:280 Log-loss0.03810292109847069\n",
      "Epoch:32 Batch:320 Log-loss0.05537784472107887\n",
      "Epoch:32 Batch:360 Log-loss0.02545783668756485\n",
      "Epoch:32 Batch:400 Log-loss0.03560284897685051\n",
      "Epoch:32 Batch:440 Log-loss0.03322187811136246\n",
      "Epoch:32 Batch:480 Log-loss0.03235450014472008\n",
      "Epoch:32 Batch:520 Log-loss0.026612935587763786\n",
      "Epoch:32 Batch:560 Log-loss0.05925809219479561\n",
      "Epoch average log loss:0.037000206987639624\n",
      "In Epoch32, val_loss:0.03773979040263386, best_val_loss:0.037555375900341165, best_auc:0.9894698424386442\n",
      "Epoch:33 Batch:0 Log-loss0.0318661704659462\n",
      "Epoch:33 Batch:40 Log-loss0.03324947878718376\n",
      "Epoch:33 Batch:80 Log-loss0.03674418479204178\n",
      "Epoch:33 Batch:120 Log-loss0.050412919372320175\n",
      "Epoch:33 Batch:160 Log-loss0.04977835714817047\n",
      "Epoch:33 Batch:200 Log-loss0.050772037357091904\n",
      "Epoch:33 Batch:240 Log-loss0.023204758763313293\n",
      "Epoch:33 Batch:280 Log-loss0.0334150567650795\n",
      "Epoch:33 Batch:320 Log-loss0.0348062738776207\n",
      "Epoch:33 Batch:360 Log-loss0.027573898434638977\n",
      "Epoch:33 Batch:400 Log-loss0.03674648329615593\n",
      "Epoch:33 Batch:440 Log-loss0.04304824396967888\n",
      "Epoch:33 Batch:480 Log-loss0.04908953234553337\n",
      "Epoch:33 Batch:520 Log-loss0.021091684699058533\n",
      "Epoch:33 Batch:560 Log-loss0.03963451087474823\n",
      "Epoch average log loss:0.037094993378767475\n",
      "In Epoch33, val_loss:0.03776682616790279, best_val_loss:0.037555375900341165, best_auc:0.9894698424386442\n",
      "Epoch:34 Batch:0 Log-loss0.0280045997351408\n",
      "Epoch:34 Batch:40 Log-loss0.04228108748793602\n",
      "Epoch:34 Batch:80 Log-loss0.03143279626965523\n",
      "Epoch:34 Batch:120 Log-loss0.03969597443938255\n",
      "Epoch:34 Batch:160 Log-loss0.0340283177793026\n",
      "Epoch:34 Batch:200 Log-loss0.036743033677339554\n",
      "Epoch:34 Batch:240 Log-loss0.03682767227292061\n",
      "Epoch:34 Batch:280 Log-loss0.04535515233874321\n",
      "Epoch:34 Batch:320 Log-loss0.03966126963496208\n",
      "Epoch:34 Batch:360 Log-loss0.02497132122516632\n",
      "Epoch:34 Batch:400 Log-loss0.03785395994782448\n",
      "Epoch:34 Batch:440 Log-loss0.03047323226928711\n",
      "Epoch:34 Batch:480 Log-loss0.043604884296655655\n",
      "Epoch:34 Batch:520 Log-loss0.03762846812605858\n",
      "Epoch:34 Batch:560 Log-loss0.035075731575489044\n",
      "Epoch average log loss:0.037029578114327576\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch34, val_loss:0.03753689442131411, best_val_loss:0.03753689442131411, best_auc:0.9894980425088349\n",
      "Epoch:35 Batch:0 Log-loss0.043506186455488205\n",
      "Epoch:35 Batch:40 Log-loss0.029366252943873405\n",
      "Epoch:35 Batch:80 Log-loss0.04491110146045685\n",
      "Epoch:35 Batch:120 Log-loss0.03476570546627045\n",
      "Epoch:35 Batch:160 Log-loss0.05384251847863197\n",
      "Epoch:35 Batch:200 Log-loss0.03821564465761185\n",
      "Epoch:35 Batch:240 Log-loss0.028437113389372826\n",
      "Epoch:35 Batch:280 Log-loss0.04121262580156326\n",
      "Epoch:35 Batch:320 Log-loss0.03227711841464043\n",
      "Epoch:35 Batch:360 Log-loss0.030910709872841835\n",
      "Epoch:35 Batch:400 Log-loss0.039235979318618774\n",
      "Epoch:35 Batch:440 Log-loss0.0230764988809824\n",
      "Epoch:35 Batch:480 Log-loss0.04893194139003754\n",
      "Epoch:35 Batch:520 Log-loss0.02819262631237507\n",
      "Epoch:35 Batch:560 Log-loss0.03947911411523819\n",
      "Epoch average log loss:0.03706115557080401\n",
      "In Epoch35, val_loss:0.03774015272791608, best_val_loss:0.03753689442131411, best_auc:0.9894980425088349\n",
      "Epoch:36 Batch:0 Log-loss0.04084985330700874\n",
      "Epoch:36 Batch:40 Log-loss0.04503254592418671\n",
      "Epoch:36 Batch:80 Log-loss0.026873936876654625\n",
      "Epoch:36 Batch:120 Log-loss0.02931053750216961\n",
      "Epoch:36 Batch:160 Log-loss0.034571271389722824\n",
      "Epoch:36 Batch:200 Log-loss0.04191166162490845\n",
      "Epoch:36 Batch:240 Log-loss0.03603973612189293\n",
      "Epoch:36 Batch:280 Log-loss0.02257745899260044\n",
      "Epoch:36 Batch:320 Log-loss0.021698692813515663\n",
      "Epoch:36 Batch:360 Log-loss0.056309301406145096\n",
      "Epoch:36 Batch:400 Log-loss0.03680960088968277\n",
      "Epoch:36 Batch:440 Log-loss0.03159823641180992\n",
      "Epoch:36 Batch:480 Log-loss0.025632344186306\n",
      "Epoch:36 Batch:520 Log-loss0.030117934569716454\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:36 Batch:560 Log-loss0.043779853731393814\n",
      "Epoch average log loss:0.03689869058850621\n",
      "In Epoch36, val_loss:0.037692696456019144, best_val_loss:0.03753689442131411, best_auc:0.9894980425088349\n",
      "Epoch:37 Batch:0 Log-loss0.033113252371549606\n",
      "Epoch:37 Batch:40 Log-loss0.0506560243666172\n",
      "Epoch:37 Batch:80 Log-loss0.032049696892499924\n",
      "Epoch:37 Batch:120 Log-loss0.0572362057864666\n",
      "Epoch:37 Batch:160 Log-loss0.039986517280340195\n",
      "Epoch:37 Batch:200 Log-loss0.03667522221803665\n",
      "Epoch:37 Batch:240 Log-loss0.029182152822613716\n",
      "Epoch:37 Batch:280 Log-loss0.03598419204354286\n",
      "Epoch:37 Batch:320 Log-loss0.04507964849472046\n",
      "Epoch:37 Batch:360 Log-loss0.029649024829268456\n",
      "Epoch:37 Batch:400 Log-loss0.03400484099984169\n",
      "Epoch:37 Batch:440 Log-loss0.03319002315402031\n",
      "Epoch:37 Batch:480 Log-loss0.03851920738816261\n",
      "Epoch:37 Batch:520 Log-loss0.042954761534929276\n",
      "Epoch:37 Batch:560 Log-loss0.05137614160776138\n",
      "Epoch average log loss:0.036965755850542334\n",
      "In Epoch37, val_loss:0.037767706182749584, best_val_loss:0.03753689442131411, best_auc:0.9894980425088349\n",
      "Epoch:38 Batch:0 Log-loss0.048654843121767044\n",
      "Epoch:38 Batch:40 Log-loss0.049928929656744\n",
      "Epoch:38 Batch:80 Log-loss0.044960688799619675\n",
      "Epoch:38 Batch:120 Log-loss0.035125698894262314\n",
      "Epoch:38 Batch:160 Log-loss0.03873629495501518\n",
      "Epoch:38 Batch:200 Log-loss0.032465990632772446\n",
      "Epoch:38 Batch:240 Log-loss0.038918402045965195\n",
      "Epoch:38 Batch:280 Log-loss0.030073771253228188\n",
      "Epoch:38 Batch:320 Log-loss0.022205866873264313\n",
      "Epoch:38 Batch:360 Log-loss0.04950452968478203\n",
      "Epoch:38 Batch:400 Log-loss0.04408600553870201\n",
      "Epoch:38 Batch:440 Log-loss0.02645478956401348\n",
      "Epoch:38 Batch:480 Log-loss0.03408680483698845\n",
      "Epoch:38 Batch:520 Log-loss0.022644633427262306\n",
      "Epoch:38 Batch:560 Log-loss0.04475206509232521\n",
      "Epoch average log loss:0.03685937927304102\n",
      "In Epoch38, val_loss:0.037604021005762475, best_val_loss:0.03753689442131411, best_auc:0.9894980425088349\n",
      "Epoch:39 Batch:0 Log-loss0.03061969019472599\n",
      "Epoch:39 Batch:40 Log-loss0.03872482106089592\n",
      "Epoch:39 Batch:80 Log-loss0.02852599509060383\n",
      "Epoch:39 Batch:120 Log-loss0.04446747899055481\n",
      "Epoch:39 Batch:160 Log-loss0.02959521673619747\n",
      "Epoch:39 Batch:200 Log-loss0.038492102175951004\n",
      "Epoch:39 Batch:240 Log-loss0.05709138512611389\n",
      "Epoch:39 Batch:280 Log-loss0.04564033821225166\n",
      "Epoch:39 Batch:320 Log-loss0.03853593394160271\n",
      "Epoch:39 Batch:360 Log-loss0.03231348469853401\n",
      "Epoch:39 Batch:400 Log-loss0.01946382038295269\n",
      "Epoch:39 Batch:440 Log-loss0.02363642305135727\n",
      "Epoch:39 Batch:480 Log-loss0.03224242478609085\n",
      "Epoch:39 Batch:520 Log-loss0.04703443869948387\n",
      "Epoch:39 Batch:560 Log-loss0.0312652625143528\n",
      "Epoch average log loss:0.036976306038975185\n",
      "In Epoch39, val_loss:0.03772815061728477, best_val_loss:0.03753689442131411, best_auc:0.9894980425088349\n",
      "Epoch:40 Batch:0 Log-loss0.03548114001750946\n",
      "Epoch:40 Batch:40 Log-loss0.034386325627565384\n",
      "Epoch:40 Batch:80 Log-loss0.026391135528683662\n",
      "Epoch:40 Batch:120 Log-loss0.03455324470996857\n",
      "Epoch:40 Batch:160 Log-loss0.042666781693696976\n",
      "Epoch:40 Batch:200 Log-loss0.03863406181335449\n",
      "Epoch:40 Batch:240 Log-loss0.04342355951666832\n",
      "Epoch:40 Batch:280 Log-loss0.03966986760497093\n",
      "Epoch:40 Batch:320 Log-loss0.03505094721913338\n",
      "Epoch:40 Batch:360 Log-loss0.027551891282200813\n",
      "Epoch:40 Batch:400 Log-loss0.04603028669953346\n",
      "Epoch:40 Batch:440 Log-loss0.03712254390120506\n",
      "Epoch:40 Batch:480 Log-loss0.03708238527178764\n",
      "Epoch:40 Batch:520 Log-loss0.03264297544956207\n",
      "Epoch:40 Batch:560 Log-loss0.03677695244550705\n",
      "Epoch average log loss:0.03670608214701393\n",
      "In Epoch40, val_loss:0.03764903466339769, best_val_loss:0.03753689442131411, best_auc:0.9894980425088349\n",
      "Epoch:41 Batch:0 Log-loss0.037765927612781525\n",
      "Epoch:41 Batch:40 Log-loss0.0475633405148983\n",
      "Epoch:41 Batch:80 Log-loss0.047460440546274185\n",
      "Epoch:41 Batch:120 Log-loss0.02981652319431305\n",
      "Epoch:41 Batch:160 Log-loss0.023361995816230774\n",
      "Epoch:41 Batch:200 Log-loss0.029265277087688446\n",
      "Epoch:41 Batch:240 Log-loss0.05399094149470329\n",
      "Epoch:41 Batch:280 Log-loss0.034952279180288315\n",
      "Epoch:41 Batch:320 Log-loss0.02462487667798996\n",
      "Epoch:41 Batch:360 Log-loss0.03377920016646385\n",
      "Epoch:41 Batch:400 Log-loss0.03317504748702049\n",
      "Epoch:41 Batch:440 Log-loss0.03708808869123459\n",
      "Epoch:41 Batch:480 Log-loss0.04225854575634003\n",
      "Epoch:41 Batch:520 Log-loss0.029448403045535088\n",
      "Epoch:41 Batch:560 Log-loss0.038518451154232025\n",
      "Epoch average log loss:0.036707201550182485\n",
      "In Epoch41, val_loss:0.03767776227740063, best_val_loss:0.03753689442131411, best_auc:0.9894980425088349\n",
      "Epoch:42 Batch:0 Log-loss0.039363715797662735\n",
      "Epoch:42 Batch:40 Log-loss0.05171319842338562\n",
      "Epoch:42 Batch:80 Log-loss0.03208129480481148\n",
      "Epoch:42 Batch:120 Log-loss0.03710998222231865\n",
      "Epoch:42 Batch:160 Log-loss0.021049099043011665\n",
      "Epoch:42 Batch:200 Log-loss0.028903663158416748\n",
      "Epoch:42 Batch:240 Log-loss0.039156634360551834\n",
      "Epoch:42 Batch:280 Log-loss0.04213179275393486\n",
      "Epoch:42 Batch:320 Log-loss0.03574173524975777\n",
      "Epoch:42 Batch:360 Log-loss0.026704445481300354\n",
      "Epoch:42 Batch:400 Log-loss0.04010618105530739\n",
      "Epoch:42 Batch:440 Log-loss0.027346424758434296\n",
      "Epoch:42 Batch:480 Log-loss0.033717963844537735\n",
      "Epoch:42 Batch:520 Log-loss0.0528632253408432\n",
      "Epoch:42 Batch:560 Log-loss0.038464415818452835\n",
      "Epoch average log loss:0.03671360237058252\n",
      "In Epoch42, val_loss:0.03767394971456159, best_val_loss:0.03753689442131411, best_auc:0.9894980425088349\n",
      "Epoch:43 Batch:0 Log-loss0.02312597446143627\n",
      "Epoch:43 Batch:40 Log-loss0.03775925934314728\n",
      "Epoch:43 Batch:80 Log-loss0.03516390547156334\n",
      "Epoch:43 Batch:120 Log-loss0.04188178852200508\n",
      "Epoch:43 Batch:160 Log-loss0.03546628728508949\n",
      "Epoch:43 Batch:200 Log-loss0.031580813229084015\n",
      "Epoch:43 Batch:240 Log-loss0.04410004988312721\n",
      "Epoch:43 Batch:280 Log-loss0.0388767383992672\n",
      "Epoch:43 Batch:320 Log-loss0.050801832228899\n",
      "Epoch:43 Batch:360 Log-loss0.039315979927778244\n",
      "Epoch:43 Batch:400 Log-loss0.03318842872977257\n",
      "Epoch:43 Batch:440 Log-loss0.03019747883081436\n",
      "Epoch:43 Batch:480 Log-loss0.027662863954901695\n",
      "Epoch:43 Batch:520 Log-loss0.052824702113866806\n",
      "Epoch:43 Batch:560 Log-loss0.05535495653748512\n",
      "Epoch average log loss:0.03671900650952011\n",
      "In Epoch43, val_loss:0.03770568957084002, best_val_loss:0.03753689442131411, best_auc:0.9894980425088349\n",
      "Epoch:44 Batch:0 Log-loss0.02887733094394207\n",
      "Epoch:44 Batch:40 Log-loss0.037985965609550476\n",
      "Epoch:44 Batch:80 Log-loss0.030382707715034485\n",
      "Epoch:44 Batch:120 Log-loss0.030499646440148354\n",
      "Epoch:44 Batch:160 Log-loss0.03230500593781471\n",
      "Epoch:44 Batch:200 Log-loss0.043089136481285095\n",
      "Epoch:44 Batch:240 Log-loss0.04415172338485718\n",
      "Epoch:44 Batch:280 Log-loss0.03845372423529625\n",
      "Epoch:44 Batch:320 Log-loss0.0304496418684721\n",
      "Epoch:44 Batch:360 Log-loss0.03880935162305832\n",
      "Epoch:44 Batch:400 Log-loss0.042133789509534836\n",
      "Epoch:44 Batch:440 Log-loss0.04179553687572479\n",
      "Epoch:44 Batch:480 Log-loss0.0318288616836071\n",
      "Epoch:44 Batch:520 Log-loss0.04021209105849266\n",
      "Epoch:44 Batch:560 Log-loss0.021646549925208092\n",
      "Epoch average log loss:0.03656954340237592\n",
      "A pre-trained model at temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt has been loaded.\n",
      "\n",
      "Model has been saved as checkpoints/FASTTXT_BGRU_64_646.pt.\n",
      "\n",
      "Choose the torch base model.\n",
      "Training on fold 7\n",
      "Epoch:1 Batch:0 Log-loss0.6847388744354248\n",
      "Epoch:1 Batch:40 Log-loss0.1473647803068161\n",
      "Epoch:1 Batch:80 Log-loss0.10918299108743668\n",
      "Epoch:1 Batch:120 Log-loss0.08407030254602432\n",
      "Epoch:1 Batch:160 Log-loss0.07367322593927383\n",
      "Epoch:1 Batch:200 Log-loss0.04902592673897743\n",
      "Epoch:1 Batch:240 Log-loss0.05660560354590416\n",
      "Epoch:1 Batch:280 Log-loss0.060514554381370544\n",
      "Epoch:1 Batch:320 Log-loss0.05570743605494499\n",
      "Epoch:1 Batch:360 Log-loss0.057314351201057434\n",
      "Epoch:1 Batch:400 Log-loss0.054464537650346756\n",
      "Epoch:1 Batch:440 Log-loss0.0801781490445137\n",
      "Epoch:1 Batch:480 Log-loss0.04890446364879608\n",
      "Epoch:1 Batch:520 Log-loss0.05452659726142883\n",
      "Epoch:1 Batch:560 Log-loss0.04016212373971939\n",
      "Epoch average log loss:0.0816530954092741\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch1, val_loss:0.048032087319946154, best_val_loss:0.048032087319946154, best_auc:0.9731544870694885\n",
      "Epoch:2 Batch:0 Log-loss0.04950941726565361\n",
      "Epoch:2 Batch:40 Log-loss0.045124366879463196\n",
      "Epoch:2 Batch:80 Log-loss0.046773865818977356\n",
      "Epoch:2 Batch:120 Log-loss0.057001661509275436\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:2 Batch:160 Log-loss0.051143571734428406\n",
      "Epoch:2 Batch:200 Log-loss0.04185109958052635\n",
      "Epoch:2 Batch:240 Log-loss0.038973793387413025\n",
      "Epoch:2 Batch:280 Log-loss0.046875178813934326\n",
      "Epoch:2 Batch:320 Log-loss0.0638737753033638\n",
      "Epoch:2 Batch:360 Log-loss0.04933242127299309\n",
      "Epoch:2 Batch:400 Log-loss0.045571219176054\n",
      "Epoch:2 Batch:440 Log-loss0.056573331356048584\n",
      "Epoch:2 Batch:480 Log-loss0.03366239741444588\n",
      "Epoch:2 Batch:520 Log-loss0.05586392804980278\n",
      "Epoch:2 Batch:560 Log-loss0.05873476713895798\n",
      "Epoch average log loss:0.04966696346444743\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch2, val_loss:0.04754312649880641, best_val_loss:0.04754312649880641, best_auc:0.9785247130291529\n",
      "Epoch:3 Batch:0 Log-loss0.03936924785375595\n",
      "Epoch:3 Batch:40 Log-loss0.03871582821011543\n",
      "Epoch:3 Batch:80 Log-loss0.03771987184882164\n",
      "Epoch:3 Batch:120 Log-loss0.04217385873198509\n",
      "Epoch:3 Batch:160 Log-loss0.05575795844197273\n",
      "Epoch:3 Batch:200 Log-loss0.05685500428080559\n",
      "Epoch:3 Batch:240 Log-loss0.04755816236138344\n",
      "Epoch:3 Batch:280 Log-loss0.040668241679668427\n",
      "Epoch:3 Batch:320 Log-loss0.04804227128624916\n",
      "Epoch:3 Batch:360 Log-loss0.04405558109283447\n",
      "Epoch:3 Batch:400 Log-loss0.04735381901264191\n",
      "Epoch:3 Batch:440 Log-loss0.041143160313367844\n",
      "Epoch:3 Batch:480 Log-loss0.046727847307920456\n",
      "Epoch:3 Batch:520 Log-loss0.04980112239718437\n",
      "Epoch:3 Batch:560 Log-loss0.03199201077222824\n",
      "Epoch average log loss:0.047510899819566736\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch3, val_loss:0.04427264511854132, best_val_loss:0.04427264511854132, best_auc:0.9805429774350252\n",
      "Epoch:4 Batch:0 Log-loss0.04492677375674248\n",
      "Epoch:4 Batch:40 Log-loss0.032988786697387695\n",
      "Epoch:4 Batch:80 Log-loss0.035446759313344955\n",
      "Epoch:4 Batch:120 Log-loss0.04756909981369972\n",
      "Epoch:4 Batch:160 Log-loss0.05936150625348091\n",
      "Epoch:4 Batch:200 Log-loss0.05519605055451393\n",
      "Epoch:4 Batch:240 Log-loss0.06507119536399841\n",
      "Epoch:4 Batch:280 Log-loss0.03558405116200447\n",
      "Epoch:4 Batch:320 Log-loss0.03329894319176674\n",
      "Epoch:4 Batch:360 Log-loss0.03620460629463196\n",
      "Epoch:4 Batch:400 Log-loss0.045777589082717896\n",
      "Epoch:4 Batch:440 Log-loss0.03110656328499317\n",
      "Epoch:4 Batch:480 Log-loss0.03991588205099106\n",
      "Epoch:4 Batch:520 Log-loss0.051307205110788345\n",
      "Epoch:4 Batch:560 Log-loss0.04328859969973564\n",
      "Epoch average log loss:0.0458797722422917\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch4, val_loss:0.04322860118945677, best_val_loss:0.04322860118945677, best_auc:0.9834331145558436\n",
      "Epoch:5 Batch:0 Log-loss0.039448074996471405\n",
      "Epoch:5 Batch:40 Log-loss0.04773596301674843\n",
      "Epoch:5 Batch:80 Log-loss0.04806764796376228\n",
      "Epoch:5 Batch:120 Log-loss0.034408967941999435\n",
      "Epoch:5 Batch:160 Log-loss0.040873344987630844\n",
      "Epoch:5 Batch:200 Log-loss0.034418437629938126\n",
      "Epoch:5 Batch:240 Log-loss0.035816434770822525\n",
      "Epoch:5 Batch:280 Log-loss0.05378564074635506\n",
      "Epoch:5 Batch:320 Log-loss0.03485655412077904\n",
      "Epoch:5 Batch:360 Log-loss0.06958527863025665\n",
      "Epoch:5 Batch:400 Log-loss0.040308233350515366\n",
      "Epoch:5 Batch:440 Log-loss0.04724821820855141\n",
      "Epoch:5 Batch:480 Log-loss0.03277435526251793\n",
      "Epoch:5 Batch:520 Log-loss0.046959783881902695\n",
      "Epoch:5 Batch:560 Log-loss0.04808307811617851\n",
      "Epoch average log loss:0.0443968327017501\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch5, val_loss:0.04237847278600023, best_val_loss:0.04237847278600023, best_auc:0.9843133012897165\n",
      "Epoch:6 Batch:0 Log-loss0.04949997738003731\n",
      "Epoch:6 Batch:40 Log-loss0.040857259184122086\n",
      "Epoch:6 Batch:80 Log-loss0.045181479305028915\n",
      "Epoch:6 Batch:120 Log-loss0.05285532400012016\n",
      "Epoch:6 Batch:160 Log-loss0.04006911441683769\n",
      "Epoch:6 Batch:200 Log-loss0.040553390979766846\n",
      "Epoch:6 Batch:240 Log-loss0.03548148274421692\n",
      "Epoch:6 Batch:280 Log-loss0.04006924107670784\n",
      "Epoch:6 Batch:320 Log-loss0.06191598251461983\n",
      "Epoch:6 Batch:360 Log-loss0.046756211668252945\n",
      "Epoch:6 Batch:400 Log-loss0.03450215235352516\n",
      "Epoch:6 Batch:440 Log-loss0.051299866288900375\n",
      "Epoch:6 Batch:480 Log-loss0.0450579971075058\n",
      "Epoch:6 Batch:520 Log-loss0.03467212989926338\n",
      "Epoch:6 Batch:560 Log-loss0.03378592059016228\n",
      "Epoch average log loss:0.04363558208237269\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch6, val_loss:0.041906960523872244, best_val_loss:0.041906960523872244, best_auc:0.9851793229322444\n",
      "Epoch:7 Batch:0 Log-loss0.049568068236112595\n",
      "Epoch:7 Batch:40 Log-loss0.03400542959570885\n",
      "Epoch:7 Batch:80 Log-loss0.04222223162651062\n",
      "Epoch:7 Batch:120 Log-loss0.03519877791404724\n",
      "Epoch:7 Batch:160 Log-loss0.03653337433934212\n",
      "Epoch:7 Batch:200 Log-loss0.03464621677994728\n",
      "Epoch:7 Batch:240 Log-loss0.033927809447050095\n",
      "Epoch:7 Batch:280 Log-loss0.02859663963317871\n",
      "Epoch:7 Batch:320 Log-loss0.03760647401213646\n",
      "Epoch:7 Batch:360 Log-loss0.04378640279173851\n",
      "Epoch:7 Batch:400 Log-loss0.04700281843543053\n",
      "Epoch:7 Batch:440 Log-loss0.04349257051944733\n",
      "Epoch:7 Batch:480 Log-loss0.044203098863363266\n",
      "Epoch:7 Batch:520 Log-loss0.058504607528448105\n",
      "Epoch:7 Batch:560 Log-loss0.03123592957854271\n",
      "Epoch average log loss:0.04283586843604488\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch7, val_loss:0.04108687899208241, best_val_loss:0.04108687899208241, best_auc:0.9861230320817183\n",
      "Epoch:8 Batch:0 Log-loss0.04602302238345146\n",
      "Epoch:8 Batch:40 Log-loss0.029139554128050804\n",
      "Epoch:8 Batch:80 Log-loss0.03362478315830231\n",
      "Epoch:8 Batch:120 Log-loss0.03719116374850273\n",
      "Epoch:8 Batch:160 Log-loss0.05996053293347359\n",
      "Epoch:8 Batch:200 Log-loss0.03772072494029999\n",
      "Epoch:8 Batch:240 Log-loss0.05645588040351868\n",
      "Epoch:8 Batch:280 Log-loss0.03802279755473137\n",
      "Epoch:8 Batch:320 Log-loss0.04481768608093262\n",
      "Epoch:8 Batch:360 Log-loss0.05121537670493126\n",
      "Epoch:8 Batch:400 Log-loss0.05782018601894379\n",
      "Epoch:8 Batch:440 Log-loss0.04293329641222954\n",
      "Epoch:8 Batch:480 Log-loss0.045943018049001694\n",
      "Epoch:8 Batch:520 Log-loss0.0499676875770092\n",
      "Epoch:8 Batch:560 Log-loss0.03215639665722847\n",
      "Epoch average log loss:0.042180228818740166\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch8, val_loss:0.04087073731392081, best_val_loss:0.04087073731392081, best_auc:0.9863827956360538\n",
      "Epoch:9 Batch:0 Log-loss0.04199642315506935\n",
      "Epoch:9 Batch:40 Log-loss0.02954779751598835\n",
      "Epoch:9 Batch:80 Log-loss0.043502237647771835\n",
      "Epoch:9 Batch:120 Log-loss0.05036960169672966\n",
      "Epoch:9 Batch:160 Log-loss0.04271744564175606\n",
      "Epoch:9 Batch:200 Log-loss0.03125210851430893\n",
      "Epoch:9 Batch:240 Log-loss0.046955760568380356\n",
      "Epoch:9 Batch:280 Log-loss0.03551499918103218\n",
      "Epoch:9 Batch:320 Log-loss0.024914344772696495\n",
      "Epoch:9 Batch:360 Log-loss0.041943375021219254\n",
      "Epoch:9 Batch:400 Log-loss0.03518382087349892\n",
      "Epoch:9 Batch:440 Log-loss0.048602182418107986\n",
      "Epoch:9 Batch:480 Log-loss0.05110188201069832\n",
      "Epoch:9 Batch:520 Log-loss0.03744601830840111\n",
      "Epoch:9 Batch:560 Log-loss0.03331148624420166\n",
      "Epoch average log loss:0.041618434186758736\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch9, val_loss:0.040868447937694445, best_val_loss:0.040868447937694445, best_auc:0.986687567033676\n",
      "Epoch:10 Batch:0 Log-loss0.034436069428920746\n",
      "Epoch:10 Batch:40 Log-loss0.040412791073322296\n",
      "Epoch:10 Batch:80 Log-loss0.04336996749043465\n",
      "Epoch:10 Batch:120 Log-loss0.04576236009597778\n",
      "Epoch:10 Batch:160 Log-loss0.03795764595270157\n",
      "Epoch:10 Batch:200 Log-loss0.04015057161450386\n",
      "Epoch:10 Batch:240 Log-loss0.03336483612656593\n",
      "Epoch:10 Batch:280 Log-loss0.054260287433862686\n",
      "Epoch:10 Batch:320 Log-loss0.03661545738577843\n",
      "Epoch:10 Batch:360 Log-loss0.035661276429891586\n",
      "Epoch:10 Batch:400 Log-loss0.034003034234046936\n",
      "Epoch:10 Batch:440 Log-loss0.03794648125767708\n",
      "Epoch:10 Batch:480 Log-loss0.04771532118320465\n",
      "Epoch:10 Batch:520 Log-loss0.043758079409599304\n",
      "Epoch:10 Batch:560 Log-loss0.049488503485918045\n",
      "Epoch average log loss:0.041249259587909495\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch10, val_loss:0.04039875872206635, best_val_loss:0.04039875872206635, best_auc:0.9872619199110989\n",
      "Epoch:11 Batch:0 Log-loss0.032437849789857864\n",
      "Epoch:11 Batch:40 Log-loss0.042168352752923965\n",
      "Epoch:11 Batch:80 Log-loss0.03140078857541084\n",
      "Epoch:11 Batch:120 Log-loss0.04715253412723541\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:11 Batch:160 Log-loss0.03412267565727234\n",
      "Epoch:11 Batch:200 Log-loss0.039331838488578796\n",
      "Epoch:11 Batch:240 Log-loss0.04977966845035553\n",
      "Epoch:11 Batch:280 Log-loss0.04404750093817711\n",
      "Epoch:11 Batch:320 Log-loss0.03475538268685341\n",
      "Epoch:11 Batch:360 Log-loss0.041808854788541794\n",
      "Epoch:11 Batch:400 Log-loss0.03803091496229172\n",
      "Epoch:11 Batch:440 Log-loss0.039786387234926224\n",
      "Epoch:11 Batch:480 Log-loss0.04521549865603447\n",
      "Epoch:11 Batch:520 Log-loss0.04675464704632759\n",
      "Epoch:11 Batch:560 Log-loss0.03984994441270828\n",
      "Epoch average log loss:0.04075316271212484\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch11, val_loss:0.040111199292105215, best_val_loss:0.040111199292105215, best_auc:0.9874906535293957\n",
      "Epoch:12 Batch:0 Log-loss0.02083994261920452\n",
      "Epoch:12 Batch:40 Log-loss0.03963178023695946\n",
      "Epoch:12 Batch:80 Log-loss0.02447463758289814\n",
      "Epoch:12 Batch:120 Log-loss0.05673886463046074\n",
      "Epoch:12 Batch:160 Log-loss0.03424261882901192\n",
      "Epoch:12 Batch:200 Log-loss0.04518589749932289\n",
      "Epoch:12 Batch:240 Log-loss0.04733797907829285\n",
      "Epoch:12 Batch:280 Log-loss0.021866247057914734\n",
      "Epoch:12 Batch:320 Log-loss0.0491948239505291\n",
      "Epoch:12 Batch:360 Log-loss0.03176405653357506\n",
      "Epoch:12 Batch:400 Log-loss0.039708878844976425\n",
      "Epoch:12 Batch:440 Log-loss0.03515198454260826\n",
      "Epoch:12 Batch:480 Log-loss0.029206983745098114\n",
      "Epoch:12 Batch:520 Log-loss0.03467964380979538\n",
      "Epoch:12 Batch:560 Log-loss0.03360031172633171\n",
      "Epoch average log loss:0.04034730960348887\n",
      "In Epoch12, val_loss:0.04031601479888507, best_val_loss:0.040111199292105215, best_auc:0.9874906535293957\n",
      "Epoch:13 Batch:0 Log-loss0.049732472747564316\n",
      "Epoch:13 Batch:40 Log-loss0.043813470751047134\n",
      "Epoch:13 Batch:80 Log-loss0.04548491910099983\n",
      "Epoch:13 Batch:120 Log-loss0.041368838399648666\n",
      "Epoch:13 Batch:160 Log-loss0.03876333311200142\n",
      "Epoch:13 Batch:200 Log-loss0.042603060603141785\n",
      "Epoch:13 Batch:240 Log-loss0.042474210262298584\n",
      "Epoch:13 Batch:280 Log-loss0.03552115336060524\n",
      "Epoch:13 Batch:320 Log-loss0.03323916718363762\n",
      "Epoch:13 Batch:360 Log-loss0.04472581669688225\n",
      "Epoch:13 Batch:400 Log-loss0.05754939839243889\n",
      "Epoch:13 Batch:440 Log-loss0.04095358029007912\n",
      "Epoch:13 Batch:480 Log-loss0.05136605724692345\n",
      "Epoch:13 Batch:520 Log-loss0.04078381136059761\n",
      "Epoch:13 Batch:560 Log-loss0.03836187347769737\n",
      "Epoch average log loss:0.039814921505084\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch13, val_loss:0.03989247372793165, best_val_loss:0.03989247372793165, best_auc:0.9879203925283071\n",
      "Epoch:14 Batch:0 Log-loss0.04019859433174133\n",
      "Epoch:14 Batch:40 Log-loss0.0352318100631237\n",
      "Epoch:14 Batch:80 Log-loss0.035979874432086945\n",
      "Epoch:14 Batch:120 Log-loss0.04270473122596741\n",
      "Epoch:14 Batch:160 Log-loss0.04769240692257881\n",
      "Epoch:14 Batch:200 Log-loss0.05482206866145134\n",
      "Epoch:14 Batch:240 Log-loss0.0429607592523098\n",
      "Epoch:14 Batch:280 Log-loss0.053824055939912796\n",
      "Epoch:14 Batch:320 Log-loss0.034033406525850296\n",
      "Epoch:14 Batch:360 Log-loss0.03510909900069237\n",
      "Epoch:14 Batch:400 Log-loss0.027408627793192863\n",
      "Epoch:14 Batch:440 Log-loss0.0507691465318203\n",
      "Epoch:14 Batch:480 Log-loss0.037000346928834915\n",
      "Epoch:14 Batch:520 Log-loss0.051704343408346176\n",
      "Epoch:14 Batch:560 Log-loss0.04446380212903023\n",
      "Epoch average log loss:0.0396334093463208\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch14, val_loss:0.039772748815156765, best_val_loss:0.039772748815156765, best_auc:0.9881501688284429\n",
      "Epoch:15 Batch:0 Log-loss0.03728616610169411\n",
      "Epoch:15 Batch:40 Log-loss0.04505695775151253\n",
      "Epoch:15 Batch:80 Log-loss0.05133683606982231\n",
      "Epoch:15 Batch:120 Log-loss0.04582890868186951\n",
      "Epoch:15 Batch:160 Log-loss0.033522140234708786\n",
      "Epoch:15 Batch:200 Log-loss0.029200874269008636\n",
      "Epoch:15 Batch:240 Log-loss0.0447346456348896\n",
      "Epoch:15 Batch:280 Log-loss0.03732946142554283\n",
      "Epoch:15 Batch:320 Log-loss0.03983616456389427\n",
      "Epoch:15 Batch:360 Log-loss0.0369112528860569\n",
      "Epoch:15 Batch:400 Log-loss0.029855087399482727\n",
      "Epoch:15 Batch:440 Log-loss0.030023589730262756\n",
      "Epoch:15 Batch:480 Log-loss0.04149103909730911\n",
      "Epoch:15 Batch:520 Log-loss0.04554137587547302\n",
      "Epoch:15 Batch:560 Log-loss0.04639372229576111\n",
      "Epoch average log loss:0.03926125509232017\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch15, val_loss:0.03973446351337286, best_val_loss:0.03973446351337286, best_auc:0.9884409719882359\n",
      "Epoch:16 Batch:0 Log-loss0.0315152108669281\n",
      "Epoch:16 Batch:40 Log-loss0.04021463915705681\n",
      "Epoch:16 Batch:80 Log-loss0.038582202047109604\n",
      "Epoch:16 Batch:120 Log-loss0.030607929453253746\n",
      "Epoch:16 Batch:160 Log-loss0.042898669838905334\n",
      "Epoch:16 Batch:200 Log-loss0.034938182681798935\n",
      "Epoch:16 Batch:240 Log-loss0.04480034485459328\n",
      "Epoch:16 Batch:280 Log-loss0.058243054896593094\n",
      "Epoch:16 Batch:320 Log-loss0.03376993164420128\n",
      "Epoch:16 Batch:360 Log-loss0.02988635189831257\n",
      "Epoch:16 Batch:400 Log-loss0.025555117055773735\n",
      "Epoch:16 Batch:440 Log-loss0.04248053953051567\n",
      "Epoch:16 Batch:480 Log-loss0.03917837142944336\n",
      "Epoch:16 Batch:520 Log-loss0.02460014820098877\n",
      "Epoch:16 Batch:560 Log-loss0.04443420097231865\n",
      "Epoch average log loss:0.0392405097572399\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch16, val_loss:0.039686360333489774, best_val_loss:0.039686360333489774, best_auc:0.9885667094306049\n",
      "Epoch:17 Batch:0 Log-loss0.039256032556295395\n",
      "Epoch:17 Batch:40 Log-loss0.03491366282105446\n",
      "Epoch:17 Batch:80 Log-loss0.04872455820441246\n",
      "Epoch:17 Batch:120 Log-loss0.037624653428792953\n",
      "Epoch:17 Batch:160 Log-loss0.04627665504813194\n",
      "Epoch:17 Batch:200 Log-loss0.04818331077694893\n",
      "Epoch:17 Batch:240 Log-loss0.02473907172679901\n",
      "Epoch:17 Batch:280 Log-loss0.04435533285140991\n",
      "Epoch:17 Batch:320 Log-loss0.042872313410043716\n",
      "Epoch:17 Batch:360 Log-loss0.05459792912006378\n",
      "Epoch:17 Batch:400 Log-loss0.040151722729206085\n",
      "Epoch:17 Batch:440 Log-loss0.039331045001745224\n",
      "Epoch:17 Batch:480 Log-loss0.041756272315979004\n",
      "Epoch:17 Batch:520 Log-loss0.03794824704527855\n",
      "Epoch:17 Batch:560 Log-loss0.031517330557107925\n",
      "Epoch average log loss:0.03878694464053426\n",
      "In Epoch17, val_loss:0.039765135773607414, best_val_loss:0.039686360333489774, best_auc:0.9885667094306049\n",
      "Epoch:18 Batch:0 Log-loss0.03937380388379097\n",
      "Epoch:18 Batch:40 Log-loss0.03302646055817604\n",
      "Epoch:18 Batch:80 Log-loss0.04253016412258148\n",
      "Epoch:18 Batch:120 Log-loss0.03590817376971245\n",
      "Epoch:18 Batch:160 Log-loss0.037942711263895035\n",
      "Epoch:18 Batch:200 Log-loss0.04351206123828888\n",
      "Epoch:18 Batch:240 Log-loss0.025547757744789124\n",
      "Epoch:18 Batch:280 Log-loss0.027362361550331116\n",
      "Epoch:18 Batch:320 Log-loss0.037260524928569794\n",
      "Epoch:18 Batch:360 Log-loss0.03249051421880722\n",
      "Epoch:18 Batch:400 Log-loss0.03187504783272743\n",
      "Epoch:18 Batch:440 Log-loss0.04035716876387596\n",
      "Epoch:18 Batch:480 Log-loss0.044282544404268265\n",
      "Epoch:18 Batch:520 Log-loss0.05516554415225983\n",
      "Epoch:18 Batch:560 Log-loss0.029926523566246033\n",
      "Epoch average log loss:0.038711824105121195\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch18, val_loss:0.03960255559537662, best_val_loss:0.03960255559537662, best_auc:0.9886193675289764\n",
      "Epoch:19 Batch:0 Log-loss0.026531225070357323\n",
      "Epoch:19 Batch:40 Log-loss0.0346241369843483\n",
      "Epoch:19 Batch:80 Log-loss0.04183223471045494\n",
      "Epoch:19 Batch:120 Log-loss0.044679660350084305\n",
      "Epoch:19 Batch:160 Log-loss0.03694837540388107\n",
      "Epoch:19 Batch:200 Log-loss0.04285956546664238\n",
      "Epoch:19 Batch:240 Log-loss0.025239592418074608\n",
      "Epoch:19 Batch:280 Log-loss0.044177114963531494\n",
      "Epoch:19 Batch:320 Log-loss0.05213463678956032\n",
      "Epoch:19 Batch:360 Log-loss0.04077412188053131\n",
      "Epoch:19 Batch:400 Log-loss0.04737050458788872\n",
      "Epoch:19 Batch:440 Log-loss0.042110029608011246\n",
      "Epoch:19 Batch:480 Log-loss0.039172131568193436\n",
      "Epoch:19 Batch:520 Log-loss0.05762694776058197\n",
      "Epoch:19 Batch:560 Log-loss0.02871011011302471\n",
      "Epoch average log loss:0.03860504872032574\n",
      "In Epoch19, val_loss:0.03961382637019391, best_val_loss:0.03960255559537662, best_auc:0.9886193675289764\n",
      "Epoch:20 Batch:0 Log-loss0.0491015799343586\n",
      "Epoch:20 Batch:40 Log-loss0.05325661972165108\n",
      "Epoch:20 Batch:80 Log-loss0.046926435083150864\n",
      "Epoch:20 Batch:120 Log-loss0.035613179206848145\n",
      "Epoch:20 Batch:160 Log-loss0.02528190053999424\n",
      "Epoch:20 Batch:200 Log-loss0.04381493106484413\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:20 Batch:240 Log-loss0.03927800431847572\n",
      "Epoch:20 Batch:280 Log-loss0.04753885045647621\n",
      "Epoch:20 Batch:320 Log-loss0.02503562718629837\n",
      "Epoch:20 Batch:360 Log-loss0.028513098135590553\n",
      "Epoch:20 Batch:400 Log-loss0.050185661762952805\n",
      "Epoch:20 Batch:440 Log-loss0.029106413945555687\n",
      "Epoch:20 Batch:480 Log-loss0.04420166090130806\n",
      "Epoch:20 Batch:520 Log-loss0.044173773378133774\n",
      "Epoch:20 Batch:560 Log-loss0.031179916113615036\n",
      "Epoch average log loss:0.03824161816654461\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch20, val_loss:0.039440354297872064, best_val_loss:0.039440354297872064, best_auc:0.9885961692987003\n",
      "Epoch:21 Batch:0 Log-loss0.04731267690658569\n",
      "Epoch:21 Batch:40 Log-loss0.021234488114714622\n",
      "Epoch:21 Batch:80 Log-loss0.035927463322877884\n",
      "Epoch:21 Batch:120 Log-loss0.03808920457959175\n",
      "Epoch:21 Batch:160 Log-loss0.04243460297584534\n",
      "Epoch:21 Batch:200 Log-loss0.04232342168688774\n",
      "Epoch:21 Batch:240 Log-loss0.032635707408189774\n",
      "Epoch:21 Batch:280 Log-loss0.0266655832529068\n",
      "Epoch:21 Batch:320 Log-loss0.0302253570407629\n",
      "Epoch:21 Batch:360 Log-loss0.029830416664481163\n",
      "Epoch:21 Batch:400 Log-loss0.028758889064192772\n",
      "Epoch:21 Batch:440 Log-loss0.032335978001356125\n",
      "Epoch:21 Batch:480 Log-loss0.026150211691856384\n",
      "Epoch:21 Batch:520 Log-loss0.035958822816610336\n",
      "Epoch:21 Batch:560 Log-loss0.033255137503147125\n",
      "Epoch average log loss:0.0383606812850173\n",
      "In Epoch21, val_loss:0.03952089401503181, best_val_loss:0.039440354297872064, best_auc:0.9885961692987003\n",
      "Epoch:22 Batch:0 Log-loss0.03665093705058098\n",
      "Epoch:22 Batch:40 Log-loss0.052704911679029465\n",
      "Epoch:22 Batch:80 Log-loss0.031391043215990067\n",
      "Epoch:22 Batch:120 Log-loss0.04058124125003815\n",
      "Epoch:22 Batch:160 Log-loss0.03770093992352486\n",
      "Epoch:22 Batch:200 Log-loss0.029224850237369537\n",
      "Epoch:22 Batch:240 Log-loss0.037364412099123\n",
      "Epoch:22 Batch:280 Log-loss0.023387828841805458\n",
      "Epoch:22 Batch:320 Log-loss0.03347298502922058\n",
      "Epoch:22 Batch:360 Log-loss0.0437176413834095\n",
      "Epoch:22 Batch:400 Log-loss0.037191618233919144\n",
      "Epoch:22 Batch:440 Log-loss0.03565965220332146\n",
      "Epoch:22 Batch:480 Log-loss0.04028230905532837\n",
      "Epoch:22 Batch:520 Log-loss0.041453417390584946\n",
      "Epoch:22 Batch:560 Log-loss0.05172073841094971\n",
      "Epoch average log loss:0.037933830664094004\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch22, val_loss:0.039413484886547466, best_val_loss:0.039413484886547466, best_auc:0.9884393984993415\n",
      "Epoch:23 Batch:0 Log-loss0.04405422508716583\n",
      "Epoch:23 Batch:40 Log-loss0.051001325249671936\n",
      "Epoch:23 Batch:80 Log-loss0.03429703786969185\n",
      "Epoch:23 Batch:120 Log-loss0.048494573682546616\n",
      "Epoch:23 Batch:160 Log-loss0.02432120591402054\n",
      "Epoch:23 Batch:200 Log-loss0.048480451107025146\n",
      "Epoch:23 Batch:240 Log-loss0.02940003015100956\n",
      "Epoch:23 Batch:280 Log-loss0.024316176772117615\n",
      "Epoch:23 Batch:320 Log-loss0.05347920581698418\n",
      "Epoch:23 Batch:360 Log-loss0.02159636653959751\n",
      "Epoch:23 Batch:400 Log-loss0.03326866030693054\n",
      "Epoch:23 Batch:440 Log-loss0.05402945354580879\n",
      "Epoch:23 Batch:480 Log-loss0.02963229827582836\n",
      "Epoch:23 Batch:520 Log-loss0.016961870715022087\n",
      "Epoch:23 Batch:560 Log-loss0.030568372458219528\n",
      "Epoch average log loss:0.03775619514219995\n",
      "In Epoch23, val_loss:0.039462631396260106, best_val_loss:0.039413484886547466, best_auc:0.9884393984993415\n",
      "Epoch:24 Batch:0 Log-loss0.039017204195261\n",
      "Epoch:24 Batch:40 Log-loss0.03489994257688522\n",
      "Epoch:24 Batch:80 Log-loss0.03910868614912033\n",
      "Epoch:24 Batch:120 Log-loss0.03384656831622124\n",
      "Epoch:24 Batch:160 Log-loss0.029902061447501183\n",
      "Epoch:24 Batch:200 Log-loss0.042303960770368576\n",
      "Epoch:24 Batch:240 Log-loss0.02799932099878788\n",
      "Epoch:24 Batch:280 Log-loss0.04171399399638176\n",
      "Epoch:24 Batch:320 Log-loss0.046121805906295776\n",
      "Epoch:24 Batch:360 Log-loss0.02842424064874649\n",
      "Epoch:24 Batch:400 Log-loss0.03243046998977661\n",
      "Epoch:24 Batch:440 Log-loss0.04530530795454979\n",
      "Epoch:24 Batch:480 Log-loss0.03888637199997902\n",
      "Epoch:24 Batch:520 Log-loss0.02212744951248169\n",
      "Epoch:24 Batch:560 Log-loss0.0559396930038929\n",
      "Epoch average log loss:0.03781300634478352\n",
      "In Epoch24, val_loss:0.039470987509573056, best_val_loss:0.039413484886547466, best_auc:0.9884393984993415\n",
      "Epoch:25 Batch:0 Log-loss0.01906864531338215\n",
      "Epoch:25 Batch:40 Log-loss0.03377547115087509\n",
      "Epoch:25 Batch:80 Log-loss0.03789268434047699\n",
      "Epoch:25 Batch:120 Log-loss0.029045993462204933\n",
      "Epoch:25 Batch:160 Log-loss0.04039779677987099\n",
      "Epoch:25 Batch:200 Log-loss0.04047254100441933\n",
      "Epoch:25 Batch:240 Log-loss0.04624016955494881\n",
      "Epoch:25 Batch:280 Log-loss0.04204067960381508\n",
      "Epoch:25 Batch:320 Log-loss0.04643465578556061\n",
      "Epoch:25 Batch:360 Log-loss0.039752744138240814\n",
      "Epoch:25 Batch:400 Log-loss0.03436814621090889\n",
      "Epoch:25 Batch:440 Log-loss0.03298741951584816\n",
      "Epoch:25 Batch:480 Log-loss0.028882628306746483\n",
      "Epoch:25 Batch:520 Log-loss0.030398204922676086\n",
      "Epoch:25 Batch:560 Log-loss0.03229491785168648\n",
      "Epoch average log loss:0.03749528487013387\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch25, val_loss:0.03918212631970267, best_val_loss:0.03918212631970267, best_auc:0.9887729979000054\n",
      "Epoch:26 Batch:0 Log-loss0.05015937611460686\n",
      "Epoch:26 Batch:40 Log-loss0.03670068457722664\n",
      "Epoch:26 Batch:80 Log-loss0.04086395353078842\n",
      "Epoch:26 Batch:120 Log-loss0.03361522778868675\n",
      "Epoch:26 Batch:160 Log-loss0.03797445818781853\n",
      "Epoch:26 Batch:200 Log-loss0.043344441801309586\n",
      "Epoch:26 Batch:240 Log-loss0.046405401080846786\n",
      "Epoch:26 Batch:280 Log-loss0.03704940155148506\n",
      "Epoch:26 Batch:320 Log-loss0.04635630175471306\n",
      "Epoch:26 Batch:360 Log-loss0.029954485595226288\n",
      "Epoch:26 Batch:400 Log-loss0.0364869050681591\n",
      "Epoch:26 Batch:440 Log-loss0.030687032267451286\n",
      "Epoch:26 Batch:480 Log-loss0.038769613951444626\n",
      "Epoch:26 Batch:520 Log-loss0.03551938757300377\n",
      "Epoch:26 Batch:560 Log-loss0.04280678927898407\n",
      "Epoch average log loss:0.03748133263184822\n",
      "In Epoch26, val_loss:0.039334599456462706, best_val_loss:0.03918212631970267, best_auc:0.9887729979000054\n",
      "Epoch:27 Batch:0 Log-loss0.030283361673355103\n",
      "Epoch:27 Batch:40 Log-loss0.0404316745698452\n",
      "Epoch:27 Batch:80 Log-loss0.03071501851081848\n",
      "Epoch:27 Batch:120 Log-loss0.024257497861981392\n",
      "Epoch:27 Batch:160 Log-loss0.034674860537052155\n",
      "Epoch:27 Batch:200 Log-loss0.04102038964629173\n",
      "Epoch:27 Batch:240 Log-loss0.03162980452179909\n",
      "Epoch:27 Batch:280 Log-loss0.05012653395533562\n",
      "Epoch:27 Batch:320 Log-loss0.026050930842757225\n",
      "Epoch:27 Batch:360 Log-loss0.036929309368133545\n",
      "Epoch:27 Batch:400 Log-loss0.04486754536628723\n",
      "Epoch:27 Batch:440 Log-loss0.041950877755880356\n",
      "Epoch:27 Batch:480 Log-loss0.04496055468916893\n",
      "Epoch:27 Batch:520 Log-loss0.05548066273331642\n",
      "Epoch:27 Batch:560 Log-loss0.05055062472820282\n",
      "Epoch average log loss:0.03738111284494932\n",
      "In Epoch27, val_loss:0.039214625609726717, best_val_loss:0.03918212631970267, best_auc:0.9887729979000054\n",
      "Epoch:28 Batch:0 Log-loss0.029902255162596703\n",
      "Epoch:28 Batch:40 Log-loss0.035896386951208115\n",
      "Epoch:28 Batch:80 Log-loss0.04631885886192322\n",
      "Epoch:28 Batch:120 Log-loss0.026410913094878197\n",
      "Epoch:28 Batch:160 Log-loss0.0433260053396225\n",
      "Epoch:28 Batch:200 Log-loss0.0375828854739666\n",
      "Epoch:28 Batch:240 Log-loss0.03594312071800232\n",
      "Epoch:28 Batch:280 Log-loss0.02600180357694626\n",
      "Epoch:28 Batch:320 Log-loss0.030849816277623177\n",
      "Epoch:28 Batch:360 Log-loss0.04336516559123993\n",
      "Epoch:28 Batch:400 Log-loss0.03653218597173691\n",
      "Epoch:28 Batch:440 Log-loss0.05697781220078468\n",
      "Epoch:28 Batch:480 Log-loss0.020524198189377785\n",
      "Epoch:28 Batch:520 Log-loss0.032962046563625336\n",
      "Epoch:28 Batch:560 Log-loss0.02858838438987732\n",
      "Epoch average log loss:0.03732610216164695\n",
      "In Epoch28, val_loss:0.039214507694115185, best_val_loss:0.03918212631970267, best_auc:0.9887729979000054\n",
      "Epoch:29 Batch:0 Log-loss0.03990308940410614\n",
      "Epoch:29 Batch:40 Log-loss0.02657437138259411\n",
      "Epoch:29 Batch:80 Log-loss0.03569754958152771\n",
      "Epoch:29 Batch:120 Log-loss0.03587372973561287\n",
      "Epoch:29 Batch:160 Log-loss0.029935874044895172\n",
      "Epoch:29 Batch:200 Log-loss0.038436610251665115\n",
      "Epoch:29 Batch:240 Log-loss0.04730182886123657\n",
      "Epoch:29 Batch:280 Log-loss0.02537563443183899\n",
      "Epoch:29 Batch:320 Log-loss0.03725668415427208\n",
      "Epoch:29 Batch:360 Log-loss0.024169614538550377\n",
      "Epoch:29 Batch:400 Log-loss0.0476381778717041\n",
      "Epoch:29 Batch:440 Log-loss0.03552936390042305\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:29 Batch:480 Log-loss0.04226592183113098\n",
      "Epoch:29 Batch:520 Log-loss0.0438595749437809\n",
      "Epoch:29 Batch:560 Log-loss0.04386654496192932\n",
      "Epoch average log loss:0.037408116340105026\n",
      "In Epoch29, val_loss:0.039461284095747166, best_val_loss:0.03918212631970267, best_auc:0.9887729979000054\n",
      "Epoch:30 Batch:0 Log-loss0.050351519137620926\n",
      "Epoch:30 Batch:40 Log-loss0.043535325676202774\n",
      "Epoch:30 Batch:80 Log-loss0.030129151418805122\n",
      "Epoch:30 Batch:120 Log-loss0.04064665734767914\n",
      "Epoch:30 Batch:160 Log-loss0.03896826133131981\n",
      "Epoch:30 Batch:200 Log-loss0.03554191812872887\n",
      "Epoch:30 Batch:240 Log-loss0.020142018795013428\n",
      "Epoch:30 Batch:280 Log-loss0.028736582025885582\n",
      "Epoch:30 Batch:320 Log-loss0.03584105893969536\n",
      "Epoch:30 Batch:360 Log-loss0.0378265306353569\n",
      "Epoch:30 Batch:400 Log-loss0.04393140971660614\n",
      "Epoch:30 Batch:440 Log-loss0.039167147129774094\n",
      "Epoch:30 Batch:480 Log-loss0.02146512269973755\n",
      "Epoch:30 Batch:520 Log-loss0.03640872612595558\n",
      "Epoch:30 Batch:560 Log-loss0.03652501478791237\n",
      "Epoch average log loss:0.037215454560438434\n",
      "In Epoch30, val_loss:0.039390880152355895, best_val_loss:0.03918212631970267, best_auc:0.9887729979000054\n",
      "Epoch:31 Batch:0 Log-loss0.04025515913963318\n",
      "Epoch:31 Batch:40 Log-loss0.028995631262660027\n",
      "Epoch:31 Batch:80 Log-loss0.03300989791750908\n",
      "Epoch:31 Batch:120 Log-loss0.034989260137081146\n",
      "Epoch:31 Batch:160 Log-loss0.03687821701169014\n",
      "Epoch:31 Batch:200 Log-loss0.043009478598833084\n",
      "Epoch:31 Batch:240 Log-loss0.03524161875247955\n",
      "Epoch:31 Batch:280 Log-loss0.04201192036271095\n",
      "Epoch:31 Batch:320 Log-loss0.030934125185012817\n",
      "Epoch:31 Batch:360 Log-loss0.03602949157357216\n",
      "Epoch:31 Batch:400 Log-loss0.040865179151296616\n",
      "Epoch:31 Batch:440 Log-loss0.05257543548941612\n",
      "Epoch:31 Batch:480 Log-loss0.04014355689287186\n",
      "Epoch:31 Batch:520 Log-loss0.048832010477781296\n",
      "Epoch:31 Batch:560 Log-loss0.029530469328165054\n",
      "Epoch average log loss:0.03726452357534851\n",
      "In Epoch31, val_loss:0.0392040389586343, best_val_loss:0.03918212631970267, best_auc:0.9887729979000054\n",
      "Epoch:32 Batch:0 Log-loss0.03195927292108536\n",
      "Epoch:32 Batch:40 Log-loss0.034355416893959045\n",
      "Epoch:32 Batch:80 Log-loss0.04779862239956856\n",
      "Epoch:32 Batch:120 Log-loss0.03529566526412964\n",
      "Epoch:32 Batch:160 Log-loss0.02844388037919998\n",
      "Epoch:32 Batch:200 Log-loss0.03961807116866112\n",
      "Epoch:32 Batch:240 Log-loss0.04545123502612114\n",
      "Epoch:32 Batch:280 Log-loss0.038101356476545334\n",
      "Epoch:32 Batch:320 Log-loss0.048623401671648026\n",
      "Epoch:32 Batch:360 Log-loss0.03624889627099037\n",
      "Epoch:32 Batch:400 Log-loss0.03428098186850548\n",
      "Epoch:32 Batch:440 Log-loss0.03705834969878197\n",
      "Epoch:32 Batch:480 Log-loss0.029787225648760796\n",
      "Epoch:32 Batch:520 Log-loss0.028221942484378815\n",
      "Epoch:32 Batch:560 Log-loss0.03296661004424095\n",
      "Epoch average log loss:0.03716330928062754\n",
      "In Epoch32, val_loss:0.039228833360179176, best_val_loss:0.03918212631970267, best_auc:0.9887729979000054\n",
      "Epoch:33 Batch:0 Log-loss0.03658837080001831\n",
      "Epoch:33 Batch:40 Log-loss0.01893739216029644\n",
      "Epoch:33 Batch:80 Log-loss0.03606293722987175\n",
      "Epoch:33 Batch:120 Log-loss0.03153231739997864\n",
      "Epoch:33 Batch:160 Log-loss0.03971758484840393\n",
      "Epoch:33 Batch:200 Log-loss0.03533118590712547\n",
      "Epoch:33 Batch:240 Log-loss0.03928961232304573\n",
      "Epoch:33 Batch:280 Log-loss0.017387043684720993\n",
      "Epoch:33 Batch:320 Log-loss0.04970294237136841\n",
      "Epoch:33 Batch:360 Log-loss0.03388490900397301\n",
      "Epoch:33 Batch:400 Log-loss0.043097883462905884\n",
      "Epoch:33 Batch:440 Log-loss0.04014674946665764\n",
      "Epoch:33 Batch:480 Log-loss0.04746508225798607\n",
      "Epoch:33 Batch:520 Log-loss0.04196473956108093\n",
      "Epoch:33 Batch:560 Log-loss0.032397277653217316\n",
      "Epoch average log loss:0.03704280859258558\n",
      "In Epoch33, val_loss:0.03923982740419036, best_val_loss:0.03918212631970267, best_auc:0.9887729979000054\n",
      "Epoch:34 Batch:0 Log-loss0.029924700036644936\n",
      "Epoch:34 Batch:40 Log-loss0.031687211245298386\n",
      "Epoch:34 Batch:80 Log-loss0.026144005358219147\n",
      "Epoch:34 Batch:120 Log-loss0.025737645104527473\n",
      "Epoch:34 Batch:160 Log-loss0.0350053496658802\n",
      "Epoch:34 Batch:200 Log-loss0.03998386859893799\n",
      "Epoch:34 Batch:240 Log-loss0.04357084631919861\n",
      "Epoch:34 Batch:280 Log-loss0.04466677084565163\n",
      "Epoch:34 Batch:320 Log-loss0.06460622698068619\n",
      "Epoch:34 Batch:360 Log-loss0.0441577322781086\n",
      "Epoch:34 Batch:400 Log-loss0.03368588909506798\n",
      "Epoch:34 Batch:440 Log-loss0.039497558027505875\n",
      "Epoch:34 Batch:480 Log-loss0.035518936812877655\n",
      "Epoch:34 Batch:520 Log-loss0.0348350889980793\n",
      "Epoch:34 Batch:560 Log-loss0.020677538588643074\n",
      "Epoch average log loss:0.0370565920136869\n",
      "In Epoch34, val_loss:0.039245300455503826, best_val_loss:0.03918212631970267, best_auc:0.9887729979000054\n",
      "Epoch:35 Batch:0 Log-loss0.029520681127905846\n",
      "Epoch:35 Batch:40 Log-loss0.0361725278198719\n",
      "Epoch:35 Batch:80 Log-loss0.04238084331154823\n",
      "Epoch:35 Batch:120 Log-loss0.05381317436695099\n",
      "Epoch:35 Batch:160 Log-loss0.03163266181945801\n",
      "Epoch:35 Batch:200 Log-loss0.03313982114195824\n",
      "Epoch:35 Batch:240 Log-loss0.037323638796806335\n",
      "Epoch:35 Batch:280 Log-loss0.053008418530225754\n",
      "Epoch:35 Batch:320 Log-loss0.028686227276921272\n",
      "Epoch:35 Batch:360 Log-loss0.04282316938042641\n",
      "Epoch:35 Batch:400 Log-loss0.03841964527964592\n",
      "Epoch:35 Batch:440 Log-loss0.03410785645246506\n",
      "Epoch:35 Batch:480 Log-loss0.03560377657413483\n",
      "Epoch:35 Batch:520 Log-loss0.03278009220957756\n",
      "Epoch:35 Batch:560 Log-loss0.04520709440112114\n",
      "Epoch average log loss:0.03708336045632937\n",
      "A pre-trained model at temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt has been loaded.\n",
      "\n",
      "Model has been saved as checkpoints/FASTTXT_BGRU_64_647.pt.\n",
      "\n",
      "Choose the torch base model.\n",
      "Training on fold 8\n",
      "Epoch:1 Batch:0 Log-loss0.7159834504127502\n",
      "Epoch:1 Batch:40 Log-loss0.1737450510263443\n",
      "Epoch:1 Batch:80 Log-loss0.12732715904712677\n",
      "Epoch:1 Batch:120 Log-loss0.05739487707614899\n",
      "Epoch:1 Batch:160 Log-loss0.072841577231884\n",
      "Epoch:1 Batch:200 Log-loss0.03947355970740318\n",
      "Epoch:1 Batch:240 Log-loss0.06868664175271988\n",
      "Epoch:1 Batch:280 Log-loss0.058306898921728134\n",
      "Epoch:1 Batch:320 Log-loss0.06096668913960457\n",
      "Epoch:1 Batch:360 Log-loss0.04807475209236145\n",
      "Epoch:1 Batch:400 Log-loss0.06641850620508194\n",
      "Epoch:1 Batch:440 Log-loss0.06771082431077957\n",
      "Epoch:1 Batch:480 Log-loss0.05163849517703056\n",
      "Epoch:1 Batch:520 Log-loss0.037381868809461594\n",
      "Epoch:1 Batch:560 Log-loss0.043999359011650085\n",
      "Epoch average log loss:0.08410625256198857\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch1, val_loss:0.047031237947955724, best_val_loss:0.047031237947955724, best_auc:0.9765077606928781\n",
      "Epoch:2 Batch:0 Log-loss0.05025814101099968\n",
      "Epoch:2 Batch:40 Log-loss0.04937128350138664\n",
      "Epoch:2 Batch:80 Log-loss0.03804296627640724\n",
      "Epoch:2 Batch:120 Log-loss0.0692998468875885\n",
      "Epoch:2 Batch:160 Log-loss0.04504923149943352\n",
      "Epoch:2 Batch:200 Log-loss0.05187910795211792\n",
      "Epoch:2 Batch:240 Log-loss0.04926518723368645\n",
      "Epoch:2 Batch:280 Log-loss0.050162628293037415\n",
      "Epoch:2 Batch:320 Log-loss0.040824998170137405\n",
      "Epoch:2 Batch:360 Log-loss0.04806411638855934\n",
      "Epoch:2 Batch:400 Log-loss0.05353475734591484\n",
      "Epoch:2 Batch:440 Log-loss0.040612637996673584\n",
      "Epoch:2 Batch:480 Log-loss0.03585008159279823\n",
      "Epoch:2 Batch:520 Log-loss0.04250263050198555\n",
      "Epoch:2 Batch:560 Log-loss0.044719550758600235\n",
      "Epoch average log loss:0.04973161562478968\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch2, val_loss:0.04352150370446634, best_val_loss:0.04352150370446634, best_auc:0.9817291822092286\n",
      "Epoch:3 Batch:0 Log-loss0.06154675409197807\n",
      "Epoch:3 Batch:40 Log-loss0.04938409850001335\n",
      "Epoch:3 Batch:80 Log-loss0.04441682621836662\n",
      "Epoch:3 Batch:120 Log-loss0.036932557821273804\n",
      "Epoch:3 Batch:160 Log-loss0.04919986426830292\n",
      "Epoch:3 Batch:200 Log-loss0.041835155338048935\n",
      "Epoch:3 Batch:240 Log-loss0.030105091631412506\n",
      "Epoch:3 Batch:280 Log-loss0.04807701334357262\n",
      "Epoch:3 Batch:320 Log-loss0.06327836960554123\n",
      "Epoch:3 Batch:360 Log-loss0.057227518409490585\n",
      "Epoch:3 Batch:400 Log-loss0.05276729166507721\n",
      "Epoch:3 Batch:440 Log-loss0.04560430347919464\n",
      "Epoch:3 Batch:480 Log-loss0.04162225127220154\n",
      "Epoch:3 Batch:520 Log-loss0.06076871231198311\n",
      "Epoch:3 Batch:560 Log-loss0.04931176081299782\n",
      "Epoch average log loss:0.0475721745419183\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch3, val_loss:0.04235944787988185, best_val_loss:0.04235944787988185, best_auc:0.9839234044759565\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:4 Batch:0 Log-loss0.04857441410422325\n",
      "Epoch:4 Batch:40 Log-loss0.0388202890753746\n",
      "Epoch:4 Batch:80 Log-loss0.034085892140865326\n",
      "Epoch:4 Batch:120 Log-loss0.058511149138212204\n",
      "Epoch:4 Batch:160 Log-loss0.046584855765104294\n",
      "Epoch:4 Batch:200 Log-loss0.04567112401127815\n",
      "Epoch:4 Batch:240 Log-loss0.04985649883747101\n",
      "Epoch:4 Batch:280 Log-loss0.04126705601811409\n",
      "Epoch:4 Batch:320 Log-loss0.04631468281149864\n",
      "Epoch:4 Batch:360 Log-loss0.07630687206983566\n",
      "Epoch:4 Batch:400 Log-loss0.04593697190284729\n",
      "Epoch:4 Batch:440 Log-loss0.05655166879296303\n",
      "Epoch:4 Batch:480 Log-loss0.04430748149752617\n",
      "Epoch:4 Batch:520 Log-loss0.04134054854512215\n",
      "Epoch:4 Batch:560 Log-loss0.044468868523836136\n",
      "Epoch average log loss:0.04551009874724384\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch4, val_loss:0.040792865033664825, best_val_loss:0.040792865033664825, best_auc:0.9865633028123061\n",
      "Epoch:5 Batch:0 Log-loss0.03637802228331566\n",
      "Epoch:5 Batch:40 Log-loss0.05392134189605713\n",
      "Epoch:5 Batch:80 Log-loss0.05715477466583252\n",
      "Epoch:5 Batch:120 Log-loss0.04299914836883545\n",
      "Epoch:5 Batch:160 Log-loss0.055256232619285583\n",
      "Epoch:5 Batch:200 Log-loss0.055297356098890305\n",
      "Epoch:5 Batch:240 Log-loss0.03885792940855026\n",
      "Epoch:5 Batch:280 Log-loss0.03219599276781082\n",
      "Epoch:5 Batch:320 Log-loss0.04246653988957405\n",
      "Epoch:5 Batch:360 Log-loss0.04977395012974739\n",
      "Epoch:5 Batch:400 Log-loss0.04198045656085014\n",
      "Epoch:5 Batch:440 Log-loss0.04818504676222801\n",
      "Epoch:5 Batch:480 Log-loss0.028663819655776024\n",
      "Epoch:5 Batch:520 Log-loss0.04024674370884895\n",
      "Epoch:5 Batch:560 Log-loss0.02948191948235035\n",
      "Epoch average log loss:0.04427311914374254\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch5, val_loss:0.04014161584617907, best_val_loss:0.04014161584617907, best_auc:0.9870377609434221\n",
      "Epoch:6 Batch:0 Log-loss0.03528156131505966\n",
      "Epoch:6 Batch:40 Log-loss0.03975541144609451\n",
      "Epoch:6 Batch:80 Log-loss0.04612524434924126\n",
      "Epoch:6 Batch:120 Log-loss0.05269394442439079\n",
      "Epoch:6 Batch:160 Log-loss0.05312766134738922\n",
      "Epoch:6 Batch:200 Log-loss0.044631075114011765\n",
      "Epoch:6 Batch:240 Log-loss0.04904970899224281\n",
      "Epoch:6 Batch:280 Log-loss0.045015085488557816\n",
      "Epoch:6 Batch:320 Log-loss0.05450628697872162\n",
      "Epoch:6 Batch:360 Log-loss0.05297081544995308\n",
      "Epoch:6 Batch:400 Log-loss0.035784762352705\n",
      "Epoch:6 Batch:440 Log-loss0.039812225848436356\n",
      "Epoch:6 Batch:480 Log-loss0.040592242032289505\n",
      "Epoch:6 Batch:520 Log-loss0.051883235573768616\n",
      "Epoch:6 Batch:560 Log-loss0.049664244055747986\n",
      "Epoch average log loss:0.04319683102824326\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch6, val_loss:0.03967976279219041, best_val_loss:0.03967976279219041, best_auc:0.9866171245411389\n",
      "Epoch:7 Batch:0 Log-loss0.034034743905067444\n",
      "Epoch:7 Batch:40 Log-loss0.04603477939963341\n",
      "Epoch:7 Batch:80 Log-loss0.04285693168640137\n",
      "Epoch:7 Batch:120 Log-loss0.04206661880016327\n",
      "Epoch:7 Batch:160 Log-loss0.0475534088909626\n",
      "Epoch:7 Batch:200 Log-loss0.024747267365455627\n",
      "Epoch:7 Batch:240 Log-loss0.06481324881315231\n",
      "Epoch:7 Batch:280 Log-loss0.04704727604985237\n",
      "Epoch:7 Batch:320 Log-loss0.05943211540579796\n",
      "Epoch:7 Batch:360 Log-loss0.032667797058820724\n",
      "Epoch:7 Batch:400 Log-loss0.03645746409893036\n",
      "Epoch:7 Batch:440 Log-loss0.04720528423786163\n",
      "Epoch:7 Batch:480 Log-loss0.04988894984126091\n",
      "Epoch:7 Batch:520 Log-loss0.02691088430583477\n",
      "Epoch:7 Batch:560 Log-loss0.059371691197156906\n",
      "Epoch average log loss:0.042436490656941064\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch7, val_loss:0.03903243801247839, best_val_loss:0.03903243801247839, best_auc:0.9878677359229707\n",
      "Epoch:8 Batch:0 Log-loss0.04105726629495621\n",
      "Epoch:8 Batch:40 Log-loss0.04444850608706474\n",
      "Epoch:8 Batch:80 Log-loss0.04096689447760582\n",
      "Epoch:8 Batch:120 Log-loss0.04039299115538597\n",
      "Epoch:8 Batch:160 Log-loss0.03308219835162163\n",
      "Epoch:8 Batch:200 Log-loss0.03771097585558891\n",
      "Epoch:8 Batch:240 Log-loss0.03279855474829674\n",
      "Epoch:8 Batch:280 Log-loss0.060421478003263474\n",
      "Epoch:8 Batch:320 Log-loss0.034058112651109695\n",
      "Epoch:8 Batch:360 Log-loss0.0298226997256279\n",
      "Epoch:8 Batch:400 Log-loss0.03880004584789276\n",
      "Epoch:8 Batch:440 Log-loss0.045527633279561996\n",
      "Epoch:8 Batch:480 Log-loss0.032693859189748764\n",
      "Epoch:8 Batch:520 Log-loss0.03936700150370598\n",
      "Epoch:8 Batch:560 Log-loss0.04838794469833374\n",
      "Epoch average log loss:0.04197749418672174\n",
      "In Epoch8, val_loss:0.03910763543816257, best_val_loss:0.03903243801247839, best_auc:0.9878677359229707\n",
      "Epoch:9 Batch:0 Log-loss0.032825227826833725\n",
      "Epoch:9 Batch:40 Log-loss0.03658854216337204\n",
      "Epoch:9 Batch:80 Log-loss0.03488708660006523\n",
      "Epoch:9 Batch:120 Log-loss0.03385379537940025\n",
      "Epoch:9 Batch:160 Log-loss0.036216773092746735\n",
      "Epoch:9 Batch:200 Log-loss0.05459311231970787\n",
      "Epoch:9 Batch:240 Log-loss0.036822687834501266\n",
      "Epoch:9 Batch:280 Log-loss0.034488745033741\n",
      "Epoch:9 Batch:320 Log-loss0.034555837512016296\n",
      "Epoch:9 Batch:360 Log-loss0.03697606548666954\n",
      "Epoch:9 Batch:400 Log-loss0.0360347144305706\n",
      "Epoch:9 Batch:440 Log-loss0.034781187772750854\n",
      "Epoch:9 Batch:480 Log-loss0.045812126249074936\n",
      "Epoch:9 Batch:520 Log-loss0.03498566523194313\n",
      "Epoch:9 Batch:560 Log-loss0.03627384454011917\n",
      "Epoch average log loss:0.04132521484778928\n",
      "In Epoch9, val_loss:0.039373018292593306, best_val_loss:0.03903243801247839, best_auc:0.9878677359229707\n",
      "Epoch:10 Batch:0 Log-loss0.036329593509435654\n",
      "Epoch:10 Batch:40 Log-loss0.03397580608725548\n",
      "Epoch:10 Batch:80 Log-loss0.06285807490348816\n",
      "Epoch:10 Batch:120 Log-loss0.08099215477705002\n",
      "Epoch:10 Batch:160 Log-loss0.04451969638466835\n",
      "Epoch:10 Batch:200 Log-loss0.03001917339861393\n",
      "Epoch:10 Batch:240 Log-loss0.04049527272582054\n",
      "Epoch:10 Batch:280 Log-loss0.030624395236372948\n",
      "Epoch:10 Batch:320 Log-loss0.04097941145300865\n",
      "Epoch:10 Batch:360 Log-loss0.042754143476486206\n",
      "Epoch:10 Batch:400 Log-loss0.037723883986473083\n",
      "Epoch:10 Batch:440 Log-loss0.04194476082921028\n",
      "Epoch:10 Batch:480 Log-loss0.022992461919784546\n",
      "Epoch:10 Batch:520 Log-loss0.031441688537597656\n",
      "Epoch:10 Batch:560 Log-loss0.03456113860011101\n",
      "Epoch average log loss:0.04073975131780441\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch10, val_loss:0.03809650519363438, best_val_loss:0.03809650519363438, best_auc:0.9883203300539964\n",
      "Epoch:11 Batch:0 Log-loss0.038593653589487076\n",
      "Epoch:11 Batch:40 Log-loss0.0270706620067358\n",
      "Epoch:11 Batch:80 Log-loss0.03377605974674225\n",
      "Epoch:11 Batch:120 Log-loss0.04913032427430153\n",
      "Epoch:11 Batch:160 Log-loss0.03615599498152733\n",
      "Epoch:11 Batch:200 Log-loss0.0514947772026062\n",
      "Epoch:11 Batch:240 Log-loss0.03236778452992439\n",
      "Epoch:11 Batch:280 Log-loss0.03474494442343712\n",
      "Epoch:11 Batch:320 Log-loss0.06013891100883484\n",
      "Epoch:11 Batch:360 Log-loss0.03802955895662308\n",
      "Epoch:11 Batch:400 Log-loss0.031286124140024185\n",
      "Epoch:11 Batch:440 Log-loss0.06256408989429474\n",
      "Epoch:11 Batch:480 Log-loss0.04490112140774727\n",
      "Epoch:11 Batch:520 Log-loss0.04607159271836281\n",
      "Epoch:11 Batch:560 Log-loss0.028095869347453117\n",
      "Epoch average log loss:0.04060614268029375\n",
      "In Epoch11, val_loss:0.03841326055416092, best_val_loss:0.03809650519363438, best_auc:0.9883203300539964\n",
      "Epoch:12 Batch:0 Log-loss0.027106262743473053\n",
      "Epoch:12 Batch:40 Log-loss0.03197503089904785\n",
      "Epoch:12 Batch:80 Log-loss0.03864133730530739\n",
      "Epoch:12 Batch:120 Log-loss0.04038150981068611\n",
      "Epoch:12 Batch:160 Log-loss0.05095447599887848\n",
      "Epoch:12 Batch:200 Log-loss0.044335704296827316\n",
      "Epoch:12 Batch:240 Log-loss0.033678844571113586\n",
      "Epoch:12 Batch:280 Log-loss0.027660004794597626\n",
      "Epoch:12 Batch:320 Log-loss0.04021655395627022\n",
      "Epoch:12 Batch:360 Log-loss0.04735824093222618\n",
      "Epoch:12 Batch:400 Log-loss0.03892191872000694\n",
      "Epoch:12 Batch:440 Log-loss0.038978587836027145\n",
      "Epoch:12 Batch:480 Log-loss0.03348822891712189\n",
      "Epoch:12 Batch:520 Log-loss0.034800201654434204\n",
      "Epoch:12 Batch:560 Log-loss0.03847157210111618\n",
      "Epoch average log loss:0.04020809261793537\n",
      "In Epoch12, val_loss:0.038539937186548516, best_val_loss:0.03809650519363438, best_auc:0.9883203300539964\n",
      "Epoch:13 Batch:0 Log-loss0.050839539617300034\n",
      "Epoch:13 Batch:40 Log-loss0.055919159203767776\n",
      "Epoch:13 Batch:80 Log-loss0.03733773157000542\n",
      "Epoch:13 Batch:120 Log-loss0.04876786097884178\n",
      "Epoch:13 Batch:160 Log-loss0.028713593259453773\n",
      "Epoch:13 Batch:200 Log-loss0.04860207065939903\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:13 Batch:240 Log-loss0.03666029870510101\n",
      "Epoch:13 Batch:280 Log-loss0.03794027492403984\n",
      "Epoch:13 Batch:320 Log-loss0.03209575638175011\n",
      "Epoch:13 Batch:360 Log-loss0.027211545035243034\n",
      "Epoch:13 Batch:400 Log-loss0.039986204355955124\n",
      "Epoch:13 Batch:440 Log-loss0.032385874539613724\n",
      "Epoch:13 Batch:480 Log-loss0.042384516447782516\n",
      "Epoch:13 Batch:520 Log-loss0.03023792989552021\n",
      "Epoch:13 Batch:560 Log-loss0.04209998995065689\n",
      "Epoch average log loss:0.04007272729317524\n",
      "In Epoch13, val_loss:0.03827162573478809, best_val_loss:0.03809650519363438, best_auc:0.9883203300539964\n",
      "Epoch:14 Batch:0 Log-loss0.05613327398896217\n",
      "Epoch:14 Batch:40 Log-loss0.02607407234609127\n",
      "Epoch:14 Batch:80 Log-loss0.037494733929634094\n",
      "Epoch:14 Batch:120 Log-loss0.04086542874574661\n",
      "Epoch:14 Batch:160 Log-loss0.042488064616918564\n",
      "Epoch:14 Batch:200 Log-loss0.033579111099243164\n",
      "Epoch:14 Batch:240 Log-loss0.04034866392612457\n",
      "Epoch:14 Batch:280 Log-loss0.03760915622115135\n",
      "Epoch:14 Batch:320 Log-loss0.04818230867385864\n",
      "Epoch:14 Batch:360 Log-loss0.018556082621216774\n",
      "Epoch:14 Batch:400 Log-loss0.04591047763824463\n",
      "Epoch:14 Batch:440 Log-loss0.0345747284591198\n",
      "Epoch:14 Batch:480 Log-loss0.04879818484187126\n",
      "Epoch:14 Batch:520 Log-loss0.029916949570178986\n",
      "Epoch:14 Batch:560 Log-loss0.03504589572548866\n",
      "Epoch average log loss:0.03965433707593807\n",
      "In Epoch14, val_loss:0.038234268630378675, best_val_loss:0.03809650519363438, best_auc:0.9883203300539964\n",
      "Epoch:15 Batch:0 Log-loss0.032437872141599655\n",
      "Epoch:15 Batch:40 Log-loss0.0510646253824234\n",
      "Epoch:15 Batch:80 Log-loss0.033335570245981216\n",
      "Epoch:15 Batch:120 Log-loss0.029026446864008904\n",
      "Epoch:15 Batch:160 Log-loss0.06016123667359352\n",
      "Epoch:15 Batch:200 Log-loss0.0289973933249712\n",
      "Epoch:15 Batch:240 Log-loss0.04878391698002815\n",
      "Epoch:15 Batch:280 Log-loss0.04475365951657295\n",
      "Epoch:15 Batch:320 Log-loss0.03715169057250023\n",
      "Epoch:15 Batch:360 Log-loss0.0416150726377964\n",
      "Epoch:15 Batch:400 Log-loss0.029904818162322044\n",
      "Epoch:15 Batch:440 Log-loss0.03822535648941994\n",
      "Epoch:15 Batch:480 Log-loss0.02440122701227665\n",
      "Epoch:15 Batch:520 Log-loss0.05432389676570892\n",
      "Epoch:15 Batch:560 Log-loss0.02072293311357498\n",
      "Epoch average log loss:0.03926188463810831\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch15, val_loss:0.03773010818155371, best_val_loss:0.03773010818155371, best_auc:0.9886115327225923\n",
      "Epoch:16 Batch:0 Log-loss0.03238993510603905\n",
      "Epoch:16 Batch:40 Log-loss0.038835976272821426\n",
      "Epoch:16 Batch:80 Log-loss0.04906176030635834\n",
      "Epoch:16 Batch:120 Log-loss0.03411758318543434\n",
      "Epoch:16 Batch:160 Log-loss0.028121301904320717\n",
      "Epoch:16 Batch:200 Log-loss0.03132477030158043\n",
      "Epoch:16 Batch:240 Log-loss0.03403988108038902\n",
      "Epoch:16 Batch:280 Log-loss0.04617295786738396\n",
      "Epoch:16 Batch:320 Log-loss0.028649315237998962\n",
      "Epoch:16 Batch:360 Log-loss0.03443606570363045\n",
      "Epoch:16 Batch:400 Log-loss0.034606192260980606\n",
      "Epoch:16 Batch:440 Log-loss0.031048625707626343\n",
      "Epoch:16 Batch:480 Log-loss0.028200283646583557\n",
      "Epoch:16 Batch:520 Log-loss0.021874278783798218\n",
      "Epoch:16 Batch:560 Log-loss0.04928867891430855\n",
      "Epoch average log loss:0.03911329621914774\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch16, val_loss:0.03754172324352814, best_val_loss:0.03754172324352814, best_auc:0.9890237911411538\n",
      "Epoch:17 Batch:0 Log-loss0.02628447115421295\n",
      "Epoch:17 Batch:40 Log-loss0.03567551448941231\n",
      "Epoch:17 Batch:80 Log-loss0.02586827240884304\n",
      "Epoch:17 Batch:120 Log-loss0.0240404661744833\n",
      "Epoch:17 Batch:160 Log-loss0.027138834819197655\n",
      "Epoch:17 Batch:200 Log-loss0.03598582372069359\n",
      "Epoch:17 Batch:240 Log-loss0.05403206869959831\n",
      "Epoch:17 Batch:280 Log-loss0.0399974025785923\n",
      "Epoch:17 Batch:320 Log-loss0.037992462515830994\n",
      "Epoch:17 Batch:360 Log-loss0.03954767808318138\n",
      "Epoch:17 Batch:400 Log-loss0.03549304977059364\n",
      "Epoch:17 Batch:440 Log-loss0.03754180297255516\n",
      "Epoch:17 Batch:480 Log-loss0.032725509256124496\n",
      "Epoch:17 Batch:520 Log-loss0.03892924636602402\n",
      "Epoch:17 Batch:560 Log-loss0.03619138151407242\n",
      "Epoch average log loss:0.038998307420739106\n",
      "In Epoch17, val_loss:0.037881276841671384, best_val_loss:0.03754172324352814, best_auc:0.9890237911411538\n",
      "Epoch:18 Batch:0 Log-loss0.027576781809329987\n",
      "Epoch:18 Batch:40 Log-loss0.039820585399866104\n",
      "Epoch:18 Batch:80 Log-loss0.04533747211098671\n",
      "Epoch:18 Batch:120 Log-loss0.039189014583826065\n",
      "Epoch:18 Batch:160 Log-loss0.029501868411898613\n",
      "Epoch:18 Batch:200 Log-loss0.04544581100344658\n",
      "Epoch:18 Batch:240 Log-loss0.0476706437766552\n",
      "Epoch:18 Batch:280 Log-loss0.03454592451453209\n",
      "Epoch:18 Batch:320 Log-loss0.05615970119833946\n",
      "Epoch:18 Batch:360 Log-loss0.021196478977799416\n",
      "Epoch:18 Batch:400 Log-loss0.02785450406372547\n",
      "Epoch:18 Batch:440 Log-loss0.05216888710856438\n",
      "Epoch:18 Batch:480 Log-loss0.03663700446486473\n",
      "Epoch:18 Batch:520 Log-loss0.03772762417793274\n",
      "Epoch:18 Batch:560 Log-loss0.042274683713912964\n",
      "Epoch average log loss:0.038599224179051816\n",
      "In Epoch18, val_loss:0.03809824478296884, best_val_loss:0.03754172324352814, best_auc:0.9890237911411538\n",
      "Epoch:19 Batch:0 Log-loss0.027302145957946777\n",
      "Epoch:19 Batch:40 Log-loss0.03127019479870796\n",
      "Epoch:19 Batch:80 Log-loss0.04860580340027809\n",
      "Epoch:19 Batch:120 Log-loss0.03701022267341614\n",
      "Epoch:19 Batch:160 Log-loss0.03702758252620697\n",
      "Epoch:19 Batch:200 Log-loss0.03110695444047451\n",
      "Epoch:19 Batch:240 Log-loss0.047169268131256104\n",
      "Epoch:19 Batch:280 Log-loss0.03868363797664642\n",
      "Epoch:19 Batch:320 Log-loss0.055937495082616806\n",
      "Epoch:19 Batch:360 Log-loss0.04088526591658592\n",
      "Epoch:19 Batch:400 Log-loss0.03920480236411095\n",
      "Epoch:19 Batch:440 Log-loss0.03901318833231926\n",
      "Epoch:19 Batch:480 Log-loss0.04249653220176697\n",
      "Epoch:19 Batch:520 Log-loss0.03821108862757683\n",
      "Epoch:19 Batch:560 Log-loss0.04318378493189812\n",
      "Epoch average log loss:0.038488734250755184\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch19, val_loss:0.03750593372882481, best_val_loss:0.03750593372882481, best_auc:0.9889360910311962\n",
      "Epoch:20 Batch:0 Log-loss0.040346164256334305\n",
      "Epoch:20 Batch:40 Log-loss0.038335587829351425\n",
      "Epoch:20 Batch:80 Log-loss0.042143434286117554\n",
      "Epoch:20 Batch:120 Log-loss0.037355467677116394\n",
      "Epoch:20 Batch:160 Log-loss0.0367610827088356\n",
      "Epoch:20 Batch:200 Log-loss0.035122375935316086\n",
      "Epoch:20 Batch:240 Log-loss0.044152531772851944\n",
      "Epoch:20 Batch:280 Log-loss0.037687577307224274\n",
      "Epoch:20 Batch:320 Log-loss0.032773714512586594\n",
      "Epoch:20 Batch:360 Log-loss0.0315123125910759\n",
      "Epoch:20 Batch:400 Log-loss0.037870846688747406\n",
      "Epoch:20 Batch:440 Log-loss0.05827366188168526\n",
      "Epoch:20 Batch:480 Log-loss0.059124916791915894\n",
      "Epoch:20 Batch:520 Log-loss0.05034298822283745\n",
      "Epoch:20 Batch:560 Log-loss0.04069828242063522\n",
      "Epoch average log loss:0.03841680142961975\n",
      "In Epoch20, val_loss:0.037734186327027165, best_val_loss:0.03750593372882481, best_auc:0.9889360910311962\n",
      "Epoch:21 Batch:0 Log-loss0.061003003269433975\n",
      "Epoch:21 Batch:40 Log-loss0.03545662388205528\n",
      "Epoch:21 Batch:80 Log-loss0.04379378631711006\n",
      "Epoch:21 Batch:120 Log-loss0.035823073238134384\n",
      "Epoch:21 Batch:160 Log-loss0.03040201961994171\n",
      "Epoch:21 Batch:200 Log-loss0.03208356723189354\n",
      "Epoch:21 Batch:240 Log-loss0.04036736115813255\n",
      "Epoch:21 Batch:280 Log-loss0.024640550836920738\n",
      "Epoch:21 Batch:320 Log-loss0.030955666676163673\n",
      "Epoch:21 Batch:360 Log-loss0.02596364915370941\n",
      "Epoch:21 Batch:400 Log-loss0.027885762974619865\n",
      "Epoch:21 Batch:440 Log-loss0.059332337230443954\n",
      "Epoch:21 Batch:480 Log-loss0.03015785664319992\n",
      "Epoch:21 Batch:520 Log-loss0.03820761293172836\n",
      "Epoch:21 Batch:560 Log-loss0.029466664418578148\n",
      "Epoch average log loss:0.03805206756972309\n",
      "In Epoch21, val_loss:0.03781623890908084, best_val_loss:0.03750593372882481, best_auc:0.9889360910311962\n",
      "Epoch:22 Batch:0 Log-loss0.036321092396974564\n",
      "Epoch:22 Batch:40 Log-loss0.05611078813672066\n",
      "Epoch:22 Batch:80 Log-loss0.03546832129359245\n",
      "Epoch:22 Batch:120 Log-loss0.02866569347679615\n",
      "Epoch:22 Batch:160 Log-loss0.03630736097693443\n",
      "Epoch:22 Batch:200 Log-loss0.04152630642056465\n",
      "Epoch:22 Batch:240 Log-loss0.0420013964176178\n",
      "Epoch:22 Batch:280 Log-loss0.035083089023828506\n",
      "Epoch:22 Batch:320 Log-loss0.03933700919151306\n",
      "Epoch:22 Batch:360 Log-loss0.04500250890851021\n",
      "Epoch:22 Batch:400 Log-loss0.04481178894639015\n",
      "Epoch:22 Batch:440 Log-loss0.04585220292210579\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:22 Batch:480 Log-loss0.04583213850855827\n",
      "Epoch:22 Batch:520 Log-loss0.023703603073954582\n",
      "Epoch:22 Batch:560 Log-loss0.03026270680129528\n",
      "Epoch average log loss:0.03812316501113985\n",
      "In Epoch22, val_loss:0.037584361195042916, best_val_loss:0.03750593372882481, best_auc:0.9889360910311962\n",
      "Epoch:23 Batch:0 Log-loss0.037214744836091995\n",
      "Epoch:23 Batch:40 Log-loss0.04588218033313751\n",
      "Epoch:23 Batch:80 Log-loss0.04655660688877106\n",
      "Epoch:23 Batch:120 Log-loss0.03538404032588005\n",
      "Epoch:23 Batch:160 Log-loss0.043606821447610855\n",
      "Epoch:23 Batch:200 Log-loss0.028064822778105736\n",
      "Epoch:23 Batch:240 Log-loss0.029773995280265808\n",
      "Epoch:23 Batch:280 Log-loss0.05827256664633751\n",
      "Epoch:23 Batch:320 Log-loss0.042706605046987534\n",
      "Epoch:23 Batch:360 Log-loss0.04082326963543892\n",
      "Epoch:23 Batch:400 Log-loss0.03643850237131119\n",
      "Epoch:23 Batch:440 Log-loss0.04934261739253998\n",
      "Epoch:23 Batch:480 Log-loss0.029192691668868065\n",
      "Epoch:23 Batch:520 Log-loss0.042013172060251236\n",
      "Epoch:23 Batch:560 Log-loss0.03500572219491005\n",
      "Epoch average log loss:0.037886357992621404\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch23, val_loss:0.03741658122718051, best_val_loss:0.03741658122718051, best_auc:0.9889405895507194\n",
      "Epoch:24 Batch:0 Log-loss0.031235069036483765\n",
      "Epoch:24 Batch:40 Log-loss0.0435156375169754\n",
      "Epoch:24 Batch:80 Log-loss0.029163384810090065\n",
      "Epoch:24 Batch:120 Log-loss0.05142667517066002\n",
      "Epoch:24 Batch:160 Log-loss0.046295326203107834\n",
      "Epoch:24 Batch:200 Log-loss0.03254657983779907\n",
      "Epoch:24 Batch:240 Log-loss0.03282509371638298\n",
      "Epoch:24 Batch:280 Log-loss0.03594803437590599\n",
      "Epoch:24 Batch:320 Log-loss0.0404357872903347\n",
      "Epoch:24 Batch:360 Log-loss0.049868226051330566\n",
      "Epoch:24 Batch:400 Log-loss0.034136995673179626\n",
      "Epoch:24 Batch:440 Log-loss0.03001795895397663\n",
      "Epoch:24 Batch:480 Log-loss0.041049886494874954\n",
      "Epoch:24 Batch:520 Log-loss0.03785763308405876\n",
      "Epoch:24 Batch:560 Log-loss0.03349993750452995\n",
      "Epoch average log loss:0.03791639679444155\n",
      "In Epoch24, val_loss:0.03764091384789666, best_val_loss:0.03741658122718051, best_auc:0.9889405895507194\n",
      "Epoch:25 Batch:0 Log-loss0.03973551467061043\n",
      "Epoch:25 Batch:40 Log-loss0.02925843559205532\n",
      "Epoch:25 Batch:80 Log-loss0.043649185448884964\n",
      "Epoch:25 Batch:120 Log-loss0.05197209119796753\n",
      "Epoch:25 Batch:160 Log-loss0.045266006141901016\n",
      "Epoch:25 Batch:200 Log-loss0.029215024784207344\n",
      "Epoch:25 Batch:240 Log-loss0.050669897347688675\n",
      "Epoch:25 Batch:280 Log-loss0.037389662116765976\n",
      "Epoch:25 Batch:320 Log-loss0.04788606986403465\n",
      "Epoch:25 Batch:360 Log-loss0.035077013075351715\n",
      "Epoch:25 Batch:400 Log-loss0.03945589065551758\n",
      "Epoch:25 Batch:440 Log-loss0.027119778096675873\n",
      "Epoch:25 Batch:480 Log-loss0.052936356514692307\n",
      "Epoch:25 Batch:520 Log-loss0.03725657984614372\n",
      "Epoch:25 Batch:560 Log-loss0.03240514546632767\n",
      "Epoch average log loss:0.03776888405970697\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch25, val_loss:0.03733482266701688, best_val_loss:0.03733482266701688, best_auc:0.9891807204146933\n",
      "Epoch:26 Batch:0 Log-loss0.032317280769348145\n",
      "Epoch:26 Batch:40 Log-loss0.054939642548561096\n",
      "Epoch:26 Batch:80 Log-loss0.03624802827835083\n",
      "Epoch:26 Batch:120 Log-loss0.04012062028050423\n",
      "Epoch:26 Batch:160 Log-loss0.03578798845410347\n",
      "Epoch:26 Batch:200 Log-loss0.03759680688381195\n",
      "Epoch:26 Batch:240 Log-loss0.04116688296198845\n",
      "Epoch:26 Batch:280 Log-loss0.05156002938747406\n",
      "Epoch:26 Batch:320 Log-loss0.025927335023880005\n",
      "Epoch:26 Batch:360 Log-loss0.038931459188461304\n",
      "Epoch:26 Batch:400 Log-loss0.04402075707912445\n",
      "Epoch:26 Batch:440 Log-loss0.0603916235268116\n",
      "Epoch:26 Batch:480 Log-loss0.029959989711642265\n",
      "Epoch:26 Batch:520 Log-loss0.031952016055583954\n",
      "Epoch:26 Batch:560 Log-loss0.03917781263589859\n",
      "Epoch average log loss:0.03771874695750219\n",
      "In Epoch26, val_loss:0.03749604049163027, best_val_loss:0.03733482266701688, best_auc:0.9891807204146933\n",
      "Epoch:27 Batch:0 Log-loss0.04675152525305748\n",
      "Epoch:27 Batch:40 Log-loss0.040210410952568054\n",
      "Epoch:27 Batch:80 Log-loss0.044276099652051926\n",
      "Epoch:27 Batch:120 Log-loss0.034098606556653976\n",
      "Epoch:27 Batch:160 Log-loss0.03889470919966698\n",
      "Epoch:27 Batch:200 Log-loss0.0313912071287632\n",
      "Epoch:27 Batch:240 Log-loss0.03453155606985092\n",
      "Epoch:27 Batch:280 Log-loss0.030280953273177147\n",
      "Epoch:27 Batch:320 Log-loss0.04118373990058899\n",
      "Epoch:27 Batch:360 Log-loss0.030283836647868156\n",
      "Epoch:27 Batch:400 Log-loss0.039450425654649734\n",
      "Epoch:27 Batch:440 Log-loss0.02769283391535282\n",
      "Epoch:27 Batch:480 Log-loss0.03282198682427406\n",
      "Epoch:27 Batch:520 Log-loss0.04547950252890587\n",
      "Epoch:27 Batch:560 Log-loss0.023265350610017776\n",
      "Epoch average log loss:0.03766658665346248\n",
      "In Epoch27, val_loss:0.037501611004300495, best_val_loss:0.03733482266701688, best_auc:0.9891807204146933\n",
      "Epoch:28 Batch:0 Log-loss0.04118536785244942\n",
      "Epoch:28 Batch:40 Log-loss0.03218836337327957\n",
      "Epoch:28 Batch:80 Log-loss0.045742396265268326\n",
      "Epoch:28 Batch:120 Log-loss0.029432691633701324\n",
      "Epoch:28 Batch:160 Log-loss0.04153880104422569\n",
      "Epoch:28 Batch:200 Log-loss0.04068535193800926\n",
      "Epoch:28 Batch:240 Log-loss0.026821961626410484\n",
      "Epoch:28 Batch:280 Log-loss0.028480909764766693\n",
      "Epoch:28 Batch:320 Log-loss0.032482851296663284\n",
      "Epoch:28 Batch:360 Log-loss0.04482010379433632\n",
      "Epoch:28 Batch:400 Log-loss0.023657813668251038\n",
      "Epoch:28 Batch:440 Log-loss0.04696473479270935\n",
      "Epoch:28 Batch:480 Log-loss0.04306816682219505\n",
      "Epoch:28 Batch:520 Log-loss0.035321757197380066\n",
      "Epoch:28 Batch:560 Log-loss0.06264135241508484\n",
      "Epoch average log loss:0.03754935007808464\n",
      "In Epoch28, val_loss:0.0374158894630759, best_val_loss:0.03733482266701688, best_auc:0.9891807204146933\n",
      "Epoch:29 Batch:0 Log-loss0.04081025719642639\n",
      "Epoch:29 Batch:40 Log-loss0.042275574058294296\n",
      "Epoch:29 Batch:80 Log-loss0.02816469967365265\n",
      "Epoch:29 Batch:120 Log-loss0.03602433577179909\n",
      "Epoch:29 Batch:160 Log-loss0.044451210647821426\n",
      "Epoch:29 Batch:200 Log-loss0.03911854699254036\n",
      "Epoch:29 Batch:240 Log-loss0.0542285293340683\n",
      "Epoch:29 Batch:280 Log-loss0.031594518572092056\n",
      "Epoch:29 Batch:320 Log-loss0.025183824822306633\n",
      "Epoch:29 Batch:360 Log-loss0.029537402093410492\n",
      "Epoch:29 Batch:400 Log-loss0.057713333517313004\n",
      "Epoch:29 Batch:440 Log-loss0.03755534812808037\n",
      "Epoch:29 Batch:480 Log-loss0.05049026384949684\n",
      "Epoch:29 Batch:520 Log-loss0.0325910821557045\n",
      "Epoch:29 Batch:560 Log-loss0.0447113960981369\n",
      "Epoch average log loss:0.037373624928295615\n",
      "In Epoch29, val_loss:0.03734068636775508, best_val_loss:0.03733482266701688, best_auc:0.9891807204146933\n",
      "Epoch:30 Batch:0 Log-loss0.03937250003218651\n",
      "Epoch:30 Batch:40 Log-loss0.029161972925066948\n",
      "Epoch:30 Batch:80 Log-loss0.03263365849852562\n",
      "Epoch:30 Batch:120 Log-loss0.03029371239244938\n",
      "Epoch:30 Batch:160 Log-loss0.026867644861340523\n",
      "Epoch:30 Batch:200 Log-loss0.041316669434309006\n",
      "Epoch:30 Batch:240 Log-loss0.043451834470033646\n",
      "Epoch:30 Batch:280 Log-loss0.046559423208236694\n",
      "Epoch:30 Batch:320 Log-loss0.028353847563266754\n",
      "Epoch:30 Batch:360 Log-loss0.026548631489276886\n",
      "Epoch:30 Batch:400 Log-loss0.028754549100995064\n",
      "Epoch:30 Batch:440 Log-loss0.03933136165142059\n",
      "Epoch:30 Batch:480 Log-loss0.0318344384431839\n",
      "Epoch:30 Batch:520 Log-loss0.027782054618000984\n",
      "Epoch:30 Batch:560 Log-loss0.03751451149582863\n",
      "Epoch average log loss:0.03741872648242861\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch30, val_loss:0.03732899875139137, best_val_loss:0.03732899875139137, best_auc:0.9890425745442356\n",
      "Epoch:31 Batch:0 Log-loss0.04051419720053673\n",
      "Epoch:31 Batch:40 Log-loss0.039691898971796036\n",
      "Epoch:31 Batch:80 Log-loss0.0346277616918087\n",
      "Epoch:31 Batch:120 Log-loss0.03591184318065643\n",
      "Epoch:31 Batch:160 Log-loss0.04546809569001198\n",
      "Epoch:31 Batch:200 Log-loss0.041774775832891464\n",
      "Epoch:31 Batch:240 Log-loss0.03111187554895878\n",
      "Epoch:31 Batch:280 Log-loss0.03959900513291359\n",
      "Epoch:31 Batch:320 Log-loss0.026567762717604637\n",
      "Epoch:31 Batch:360 Log-loss0.032844822853803635\n",
      "Epoch:31 Batch:400 Log-loss0.04191766306757927\n",
      "Epoch:31 Batch:440 Log-loss0.039985429495573044\n",
      "Epoch:31 Batch:480 Log-loss0.030769670382142067\n",
      "Epoch:31 Batch:520 Log-loss0.026361554861068726\n",
      "Epoch:31 Batch:560 Log-loss0.04033121466636658\n",
      "Epoch average log loss:0.037188925657288305\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch31, val_loss:0.037280518824741884, best_val_loss:0.037280518824741884, best_auc:0.9891612523457282\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:32 Batch:0 Log-loss0.02555924467742443\n",
      "Epoch:32 Batch:40 Log-loss0.043173015117645264\n",
      "Epoch:32 Batch:80 Log-loss0.03214317187666893\n",
      "Epoch:32 Batch:120 Log-loss0.04182949289679527\n",
      "Epoch:32 Batch:160 Log-loss0.048973049968481064\n",
      "Epoch:32 Batch:200 Log-loss0.027037186548113823\n",
      "Epoch:32 Batch:240 Log-loss0.03429853543639183\n",
      "Epoch:32 Batch:280 Log-loss0.04188264533877373\n",
      "Epoch:32 Batch:320 Log-loss0.027212508022785187\n",
      "Epoch:32 Batch:360 Log-loss0.05423672869801521\n",
      "Epoch:32 Batch:400 Log-loss0.026018239557743073\n",
      "Epoch:32 Batch:440 Log-loss0.03591928258538246\n",
      "Epoch:32 Batch:480 Log-loss0.029280902817845345\n",
      "Epoch:32 Batch:520 Log-loss0.03492751345038414\n",
      "Epoch:32 Batch:560 Log-loss0.04193948954343796\n",
      "Epoch average log loss:0.03728586017553295\n",
      "In Epoch32, val_loss:0.03737022754721287, best_val_loss:0.037280518824741884, best_auc:0.9891612523457282\n",
      "Epoch:33 Batch:0 Log-loss0.054315660148859024\n",
      "Epoch:33 Batch:40 Log-loss0.04033203795552254\n",
      "Epoch:33 Batch:80 Log-loss0.030183130875229836\n",
      "Epoch:33 Batch:120 Log-loss0.0322297029197216\n",
      "Epoch:33 Batch:160 Log-loss0.043465498834848404\n",
      "Epoch:33 Batch:200 Log-loss0.043879445642232895\n",
      "Epoch:33 Batch:240 Log-loss0.042952507734298706\n",
      "Epoch:33 Batch:280 Log-loss0.034070711582899094\n",
      "Epoch:33 Batch:320 Log-loss0.0443364717066288\n",
      "Epoch:33 Batch:360 Log-loss0.04191993549466133\n",
      "Epoch:33 Batch:400 Log-loss0.038383595645427704\n",
      "Epoch:33 Batch:440 Log-loss0.03415733575820923\n",
      "Epoch:33 Batch:480 Log-loss0.04168145731091499\n",
      "Epoch:33 Batch:520 Log-loss0.056622136384248734\n",
      "Epoch:33 Batch:560 Log-loss0.04547705873847008\n",
      "Epoch average log loss:0.03703521355387888\n",
      "In Epoch33, val_loss:0.0373681116061955, best_val_loss:0.037280518824741884, best_auc:0.9891612523457282\n",
      "Epoch:34 Batch:0 Log-loss0.035542163997888565\n",
      "Epoch:34 Batch:40 Log-loss0.05068262293934822\n",
      "Epoch:34 Batch:80 Log-loss0.029585182666778564\n",
      "Epoch:34 Batch:120 Log-loss0.03425804525613785\n",
      "Epoch:34 Batch:160 Log-loss0.04044454172253609\n",
      "Epoch:34 Batch:200 Log-loss0.03100859560072422\n",
      "Epoch:34 Batch:240 Log-loss0.03750937059521675\n",
      "Epoch:34 Batch:280 Log-loss0.03432982787489891\n",
      "Epoch:34 Batch:320 Log-loss0.04016968607902527\n",
      "Epoch:34 Batch:360 Log-loss0.04137187451124191\n",
      "Epoch:34 Batch:400 Log-loss0.039030011743307114\n",
      "Epoch:34 Batch:440 Log-loss0.03829947113990784\n",
      "Epoch:34 Batch:480 Log-loss0.037445537745952606\n",
      "Epoch:34 Batch:520 Log-loss0.03140982612967491\n",
      "Epoch:34 Batch:560 Log-loss0.0319150909781456\n",
      "Epoch average log loss:0.03708627867246313\n",
      "In Epoch34, val_loss:0.03735566996021334, best_val_loss:0.037280518824741884, best_auc:0.9891612523457282\n",
      "Epoch:35 Batch:0 Log-loss0.040992435067892075\n",
      "Epoch:35 Batch:40 Log-loss0.039444390684366226\n",
      "Epoch:35 Batch:80 Log-loss0.03171803802251816\n",
      "Epoch:35 Batch:120 Log-loss0.02853487618267536\n",
      "Epoch:35 Batch:160 Log-loss0.033874258399009705\n",
      "Epoch:35 Batch:200 Log-loss0.02402404509484768\n",
      "Epoch:35 Batch:240 Log-loss0.02929956279695034\n",
      "Epoch:35 Batch:280 Log-loss0.0454569011926651\n",
      "Epoch:35 Batch:320 Log-loss0.03148229047656059\n",
      "Epoch:35 Batch:360 Log-loss0.03766770660877228\n",
      "Epoch:35 Batch:400 Log-loss0.04817536100745201\n",
      "Epoch:35 Batch:440 Log-loss0.03583447262644768\n",
      "Epoch:35 Batch:480 Log-loss0.04011039063334465\n",
      "Epoch:35 Batch:520 Log-loss0.04511226341128349\n",
      "Epoch:35 Batch:560 Log-loss0.039899565279483795\n",
      "Epoch average log loss:0.03715980926728142\n",
      "In Epoch35, val_loss:0.037343013598350384, best_val_loss:0.037280518824741884, best_auc:0.9891612523457282\n",
      "Epoch:36 Batch:0 Log-loss0.027390748262405396\n",
      "Epoch:36 Batch:40 Log-loss0.03682770952582359\n",
      "Epoch:36 Batch:80 Log-loss0.044804323464632034\n",
      "Epoch:36 Batch:120 Log-loss0.041138533502817154\n",
      "Epoch:36 Batch:160 Log-loss0.03126450628042221\n",
      "Epoch:36 Batch:200 Log-loss0.02871652878820896\n",
      "Epoch:36 Batch:240 Log-loss0.016614744439721107\n",
      "Epoch:36 Batch:280 Log-loss0.03286256641149521\n",
      "Epoch:36 Batch:320 Log-loss0.04188508912920952\n",
      "Epoch:36 Batch:360 Log-loss0.0405288003385067\n",
      "Epoch:36 Batch:400 Log-loss0.040382370352745056\n",
      "Epoch:36 Batch:440 Log-loss0.026983385905623436\n",
      "Epoch:36 Batch:480 Log-loss0.03270965814590454\n",
      "Epoch:36 Batch:520 Log-loss0.023365361616015434\n",
      "Epoch:36 Batch:560 Log-loss0.04872070252895355\n",
      "Epoch average log loss:0.03679906792240217\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch36, val_loss:0.03722852222087956, best_val_loss:0.03722852222087956, best_auc:0.9890564514341756\n",
      "Epoch:37 Batch:0 Log-loss0.034658219665288925\n",
      "Epoch:37 Batch:40 Log-loss0.026480356231331825\n",
      "Epoch:37 Batch:80 Log-loss0.0394851379096508\n",
      "Epoch:37 Batch:120 Log-loss0.049479857087135315\n",
      "Epoch:37 Batch:160 Log-loss0.041176099330186844\n",
      "Epoch:37 Batch:200 Log-loss0.03891998529434204\n",
      "Epoch:37 Batch:240 Log-loss0.02449328638613224\n",
      "Epoch:37 Batch:280 Log-loss0.027745775878429413\n",
      "Epoch:37 Batch:320 Log-loss0.04008052125573158\n",
      "Epoch:37 Batch:360 Log-loss0.044503260403871536\n",
      "Epoch:37 Batch:400 Log-loss0.044721320271492004\n",
      "Epoch:37 Batch:440 Log-loss0.0413668192923069\n",
      "Epoch:37 Batch:480 Log-loss0.04073168337345123\n",
      "Epoch:37 Batch:520 Log-loss0.03495747968554497\n",
      "Epoch:37 Batch:560 Log-loss0.03812623769044876\n",
      "Epoch average log loss:0.036785230482928456\n",
      "In Epoch37, val_loss:0.037322914544773024, best_val_loss:0.03722852222087956, best_auc:0.9890564514341756\n",
      "Epoch:38 Batch:0 Log-loss0.027000943198800087\n",
      "Epoch:38 Batch:40 Log-loss0.04061673954129219\n",
      "Epoch:38 Batch:80 Log-loss0.05059562996029854\n",
      "Epoch:38 Batch:120 Log-loss0.03693980723619461\n",
      "Epoch:38 Batch:160 Log-loss0.03739790990948677\n",
      "Epoch:38 Batch:200 Log-loss0.03573295846581459\n",
      "Epoch:38 Batch:240 Log-loss0.04197469726204872\n",
      "Epoch:38 Batch:280 Log-loss0.029526004567742348\n",
      "Epoch:38 Batch:320 Log-loss0.03616681322455406\n",
      "Epoch:38 Batch:360 Log-loss0.049979496747255325\n",
      "Epoch:38 Batch:400 Log-loss0.031815409660339355\n",
      "Epoch:38 Batch:440 Log-loss0.04396064952015877\n",
      "Epoch:38 Batch:480 Log-loss0.023452656343579292\n",
      "Epoch:38 Batch:520 Log-loss0.03942645341157913\n",
      "Epoch:38 Batch:560 Log-loss0.038625720888376236\n",
      "Epoch average log loss:0.036995707784912416\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch38, val_loss:0.03719077182150346, best_val_loss:0.03719077182150346, best_auc:0.9890165718722078\n",
      "Epoch:39 Batch:0 Log-loss0.04016615077853203\n",
      "Epoch:39 Batch:40 Log-loss0.040844857692718506\n",
      "Epoch:39 Batch:80 Log-loss0.038754504173994064\n",
      "Epoch:39 Batch:120 Log-loss0.03784547001123428\n",
      "Epoch:39 Batch:160 Log-loss0.03219325467944145\n",
      "Epoch:39 Batch:200 Log-loss0.04241184517741203\n",
      "Epoch:39 Batch:240 Log-loss0.043886762112379074\n",
      "Epoch:39 Batch:280 Log-loss0.05087236687541008\n",
      "Epoch:39 Batch:320 Log-loss0.03512367233633995\n",
      "Epoch:39 Batch:360 Log-loss0.03659537807106972\n",
      "Epoch:39 Batch:400 Log-loss0.022997314110398293\n",
      "Epoch:39 Batch:440 Log-loss0.0248101819306612\n",
      "Epoch:39 Batch:480 Log-loss0.04040003940463066\n",
      "Epoch:39 Batch:520 Log-loss0.04615933075547218\n",
      "Epoch:39 Batch:560 Log-loss0.025107108056545258\n",
      "Epoch average log loss:0.03700028236489743\n",
      "In Epoch39, val_loss:0.03727926171658792, best_val_loss:0.03719077182150346, best_auc:0.9890165718722078\n",
      "Epoch:40 Batch:0 Log-loss0.022937921807169914\n",
      "Epoch:40 Batch:40 Log-loss0.030566314235329628\n",
      "Epoch:40 Batch:80 Log-loss0.04522509500384331\n",
      "Epoch:40 Batch:120 Log-loss0.029762016609311104\n",
      "Epoch:40 Batch:160 Log-loss0.03157923370599747\n",
      "Epoch:40 Batch:200 Log-loss0.03123619593679905\n",
      "Epoch:40 Batch:240 Log-loss0.02937442995607853\n",
      "Epoch:40 Batch:280 Log-loss0.0335443951189518\n",
      "Epoch:40 Batch:320 Log-loss0.031075915321707726\n",
      "Epoch:40 Batch:360 Log-loss0.043952494859695435\n",
      "Epoch:40 Batch:400 Log-loss0.03846026584506035\n",
      "Epoch:40 Batch:440 Log-loss0.037237461656332016\n",
      "Epoch:40 Batch:480 Log-loss0.03888985514640808\n",
      "Epoch:40 Batch:520 Log-loss0.041369106620550156\n",
      "Epoch:40 Batch:560 Log-loss0.0361940823495388\n",
      "Epoch average log loss:0.03696970444538498\n",
      "Model has been saved as temporary_checkpoints/FASTTXT_BGRU_64_64-TEMP.pt.\n",
      "\n",
      "In Epoch40, val_loss:0.037187047823038145, best_val_loss:0.037187047823038145, best_auc:0.9889335917214264\n",
      "Epoch:41 Batch:0 Log-loss0.046189576387405396\n",
      "Epoch:41 Batch:40 Log-loss0.04938426986336708\n",
      "Epoch:41 Batch:80 Log-loss0.03842392936348915\n",
      "Epoch:41 Batch:120 Log-loss0.048611417412757874\n",
      "Epoch:41 Batch:160 Log-loss0.0379195399582386\n",
      "Epoch:41 Batch:200 Log-loss0.040140461176633835\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:41 Batch:240 Log-loss0.04956863448023796\n",
      "Epoch:41 Batch:280 Log-loss0.028649361804127693\n",
      "Epoch:41 Batch:320 Log-loss0.041014574468135834\n",
      "Epoch:41 Batch:360 Log-loss0.030673174187541008\n",
      "Epoch:41 Batch:400 Log-loss0.0355849415063858\n",
      "Epoch:41 Batch:440 Log-loss0.026770837604999542\n",
      "Epoch:41 Batch:480 Log-loss0.029718542471528053\n",
      "Epoch:41 Batch:520 Log-loss0.03486945107579231\n",
      "Epoch:41 Batch:560 Log-loss0.04453378543257713\n",
      "Epoch average log loss:0.036813383070485935\n",
      "In Epoch41, val_loss:0.03725409934025623, best_val_loss:0.037187047823038145, best_auc:0.9889335917214264\n",
      "Epoch:42 Batch:0 Log-loss0.032268550246953964\n",
      "Epoch:42 Batch:40 Log-loss0.03819175437092781\n",
      "Epoch:42 Batch:80 Log-loss0.026156602427363396\n",
      "Epoch:42 Batch:120 Log-loss0.035541970282793045\n",
      "Epoch:42 Batch:160 Log-loss0.037008222192525864\n",
      "Epoch:42 Batch:200 Log-loss0.03382289037108421\n",
      "Epoch:42 Batch:240 Log-loss0.049914002418518066\n",
      "Epoch:42 Batch:280 Log-loss0.04340128228068352\n",
      "Epoch:42 Batch:320 Log-loss0.03338106349110603\n",
      "Epoch:42 Batch:360 Log-loss0.031241053715348244\n",
      "Epoch:42 Batch:400 Log-loss0.034176282584667206\n",
      "Epoch:42 Batch:440 Log-loss0.05386197194457054\n",
      "Epoch:42 Batch:480 Log-loss0.03476850688457489\n",
      "Epoch:42 Batch:520 Log-loss0.03774731606245041\n",
      "Epoch:42 Batch:560 Log-loss0.016875572502613068\n",
      "Epoch average log loss:0.03660768243051799\n",
      "In Epoch42, val_loss:0.03732168410835348, best_val_loss:0.037187047823038145, best_auc:0.9889335917214264\n",
      "Epoch:43 Batch:0 Log-loss0.03515872731804848\n",
      "Epoch:43 Batch:40 Log-loss0.026994481682777405\n",
      "Epoch:43 Batch:80 Log-loss0.03507150709629059\n",
      "Epoch:43 Batch:120 Log-loss0.052563104778528214\n",
      "Epoch:43 Batch:160 Log-loss0.04921707138419151\n",
      "Epoch:43 Batch:200 Log-loss0.02705453336238861\n",
      "Epoch:43 Batch:240 Log-loss0.03229468688368797\n",
      "Epoch:43 Batch:280 Log-loss0.03737211227416992\n",
      "Epoch:43 Batch:320 Log-loss0.0402056910097599\n",
      "Epoch:43 Batch:360 Log-loss0.03167347237467766\n",
      "Epoch:43 Batch:400 Log-loss0.03287142515182495\n",
      "Epoch:43 Batch:440 Log-loss0.03742274269461632\n",
      "Epoch:43 Batch:480 Log-loss0.03631872311234474\n",
      "Epoch:43 Batch:520 Log-loss0.028033895418047905\n",
      "Epoch:43 Batch:560 Log-loss0.03672155365347862\n",
      "Epoch average log loss:0.036791893686833124\n",
      "In Epoch43, val_loss:0.037288376450489484, best_val_loss:0.037187047823038145, best_auc:0.9889335917214264\n",
      "Epoch:44 Batch:0 Log-loss0.031550969928503036\n",
      "Epoch:44 Batch:40 Log-loss0.03381931409239769\n",
      "Epoch:44 Batch:80 Log-loss0.045408520847558975\n",
      "Epoch:44 Batch:120 Log-loss0.04696900025010109\n",
      "Epoch:44 Batch:160 Log-loss0.035124000161886215\n",
      "Epoch:44 Batch:200 Log-loss0.05058664083480835\n",
      "Epoch:44 Batch:240 Log-loss0.061946768313646317\n",
      "Epoch:44 Batch:280 Log-loss0.04035746306180954\n",
      "Epoch:44 Batch:320 Log-loss0.03693334385752678\n",
      "Epoch:44 Batch:360 Log-loss0.04026741161942482\n",
      "Epoch:44 Batch:400 Log-loss0.033640678972005844\n",
      "Epoch:44 Batch:440 Log-loss0.034867141395807266\n",
      "Epoch:44 Batch:480 Log-loss0.02929726056754589\n",
      "Epoch:44 Batch:520 Log-loss0.035537656396627426\n",
      "Epoch:44 Batch:560 Log-loss0.024604352191090584\n",
      "Epoch average log loss:0.03668963035036411\n",
      "In Epoch44, val_loss:0.03746491128825227, best_val_loss:0.037187047823038145, best_auc:0.9889335917214264\n",
      "Epoch:45 Batch:0 Log-loss0.03961610421538353\n",
      "Epoch:45 Batch:40 Log-loss0.04841339588165283\n",
      "Epoch:45 Batch:80 Log-loss0.022393397986888885\n",
      "Epoch:45 Batch:120 Log-loss0.034102603793144226\n",
      "Epoch:45 Batch:160 Log-loss0.02802799455821514\n",
      "Epoch:45 Batch:200 Log-loss0.0468687005341053\n",
      "Epoch:45 Batch:240 Log-loss0.04173804819583893\n",
      "Epoch:45 Batch:280 Log-loss0.04868317022919655\n",
      "Epoch:45 Batch:320 Log-loss0.04020778834819794\n",
      "Epoch:45 Batch:360 Log-loss0.03284776583313942\n",
      "Epoch:45 Batch:400 Log-loss0.043810080736875534\n",
      "Epoch:45 Batch:440 Log-loss0.03616361320018768\n",
      "Epoch:45 Batch:480 Log-loss0.030474631115794182\n",
      "Epoch:45 Batch:520 Log-loss0.025500783696770668\n",
      "Epoch:45 Batch:560 Log-loss0.025793680921196938\n",
      "Epoch average log loss:0.03672576240995633\n",
      "In Epoch45, val_loss:0.03736899266467802, best_val_loss:0.037187047823038145, best_auc:0.9889335917214264\n",
      "Epoch:46 Batch:0 Log-loss0.029167594388127327\n",
      "Epoch:46 Batch:40 Log-loss0.03613545373082161\n",
      "Epoch:46 Batch:80 Log-loss0.027204664424061775\n",
      "Epoch:46 Batch:120 Log-loss0.03930847346782684\n",
      "Epoch:46 Batch:160 Log-loss0.03739314153790474\n",
      "Epoch:46 Batch:200 Log-loss0.0395079143345356\n",
      "Epoch:46 Batch:240 Log-loss0.05430882051587105\n",
      "Epoch:46 Batch:280 Log-loss0.04689684510231018\n",
      "Epoch:46 Batch:320 Log-loss0.039858344942331314\n",
      "Epoch:46 Batch:360 Log-loss0.035380978137254715\n",
      "Epoch:46 Batch:400 Log-loss0.03116876445710659\n",
      "Epoch:46 Batch:440 Log-loss0.03970806673169136\n",
      "Epoch:46 Batch:480 Log-loss0.03346617892384529\n",
      "Epoch:46 Batch:520 Log-loss0.03172599896788597\n",
      "Epoch:46 Batch:560 Log-loss0.048927102237939835\n",
      "Epoch average log loss:0.03674454153182783\n",
      "In Epoch46, val_loss:0.037268073350224924, best_val_loss:0.037187047823038145, best_auc:0.9889335917214264\n",
      "Epoch:47 Batch:0 Log-loss0.028306225314736366\n",
      "Epoch:47 Batch:40 Log-loss0.04166557639837265\n",
      "Epoch:47 Batch:80 Log-loss0.02781062014400959\n",
      "Epoch:47 Batch:120 Log-loss0.042322877794504166\n",
      "Epoch:47 Batch:160 Log-loss0.03688504919409752\n",
      "Epoch:47 Batch:200 Log-loss0.02853623777627945\n",
      "Epoch:47 Batch:240 Log-loss0.03474757447838783\n",
      "Epoch:47 Batch:280 Log-loss0.05562729761004448\n",
      "Epoch:47 Batch:320 Log-loss0.03748201206326485\n",
      "Epoch:47 Batch:360 Log-loss0.031562112271785736\n",
      "Epoch:47 Batch:400 Log-loss0.04009081795811653\n"
     ]
    }
   ],
   "source": [
    "model, best_logloss, best_auc, best_val_pred = trainer.train_folds(X=train_sequences, y=training_labels,\n",
    "                    fold_count=10, batch_size=256, get_model_func=get_bgru_network, skip_fold=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Make predections"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from sklearn.metrics import roc_auc_score\n",
    "\n",
    "train_fold_preditcions = np.concatenate(best_val_pred, axis=0)\n",
    "training_auc = roc_auc_score(training_labels, train_fold_preditcions)\n",
    "print(\"Training AUC\", training_auc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "path = 'Dataset/'\n",
    "TRAIN_DATA_FILE=path + 'train.csv'\n",
    "TEST_DATA_FILE=path + 'test.csv'\n",
    "test_df = pd.read_csv(TEST_DATA_FILE)\n",
    "train_df = pd.read_csv(TRAIN_DATA_FILE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "CLASSES = [\"toxic\", \"severe_toxic\", \"obscene\", \"threat\", \"insult\", \"identity_hate\"]\n",
    "submit_path_prefix = \"results/bgru/Fasttext-BGRU-\" + str(MAX_SEQUENCE_LENGTH) \n",
    "\n",
    "print(\"Predicting testing results...\")\n",
    "test_predicts_list = []\n",
    "for fold_id, model in enumerate(new_models):\n",
    "    test_predicts = model.predict(test_sequences, batch_size=256, verbose=1)\n",
    "    test_predicts_list.append(test_predicts)\n",
    "\n",
    "test_predicts = np.zeros(test_predicts_list[0].shape)\n",
    "for fold_predict in test_predicts_list:\n",
    "    test_predicts += fold_predict\n",
    "test_predicts /= len(test_predicts_list)\n",
    "\n",
    "test_ids = test_df[\"id\"].values\n",
    "test_ids = test_ids.reshape((len(test_ids), 1))\n",
    "\n",
    "test_predicts = pd.DataFrame(data=test_predicts, columns=CLASSES)\n",
    "test_predicts[\"id\"] = test_ids\n",
    "test_predicts = test_predicts[[\"id\"] + CLASSES]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "submit_path = submit_path_prefix + \"-L{:4f}-A{:4f}.csv\".format(best_logloss, best_auc)\n",
    "test_predicts.to_csv(submit_path, index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "print(\"Predicting training results...\")\n",
    "\n",
    "train_ids = train_df[\"id\"].values\n",
    "train_ids = train_ids.reshape((len(train_ids), 1))\n",
    "\n",
    "train_predicts = pd.DataFrame(data=train_fold_preditcions, columns=CLASSES)\n",
    "train_predicts[\"id\"] = train_ids\n",
    "train_predicts = train_predicts[[\"id\"] + CLASSES]\n",
    "submit_path = submit_path_prefix + \"-Train-L{:4f}-A{:4f}.csv\".format(best_logloss, best_auc)\n",
    "train_predicts.to_csv(submit_path, index=False)"
   ]
  }
 ],
 "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": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}