2. CNN.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import re\n",
    "import numpy as np\n",
    "\n",
    "# Load Pre-trained Word2vec Model\n",
    "from gensim.models import word2vec\n",
    "\n",
    "# Data visualization\n",
    "from matplotlib import pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "# CNN\n",
    "from keras.preprocessing.text import Tokenizer, text_to_word_sequence\n",
    "from keras.preprocessing.sequence import pad_sequences\n",
    "from keras import losses\n",
    "\n",
    "from keras.layers import Embedding\n",
    "from keras.layers import Dense, Input, Flatten\n",
    "from keras.layers import Conv1D, MaxPooling1D, Embedding, Merge, Dropout\n",
    "from keras.models import Model, Sequential"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Load Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "TRAIN_FILE = 'data/ratings_train.txt'\n",
    "TEST_FILE = 'data/ratings_test.txt'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Read data\n",
    "train_reviews = []\n",
    "with open(TRAIN_FILE, 'r') as f:\n",
    "    next(f) # header skip\n",
    "    for line in f.readlines():\n",
    "        _, review, label = line.strip().split('\\t')\n",
    "        train_reviews.append((review.strip(), label.strip()))\n",
    "        \n",
    "test_reviews = []        \n",
    "with open(TEST_FILE, 'r') as f:\n",
    "    next(f) # header skip\n",
    "    for line in f.readlines():\n",
    "        _, review, label = line.strip().split('\\t')\n",
    "        test_reviews.append((review.strip(), label.strip()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from konlpy.tag import Twitter\n",
    "twi_tagger = Twitter()\n",
    "\n",
    "def tokenizer(doc):\n",
    "    tokens = ['/'.join(t) for t in twi_tagger.pos(doc, norm=True, stem=True)]\n",
    "    return tokens"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 3min 38s, sys: 2.1 s, total: 3min 40s\n",
      "Wall time: 3min 36s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "train_reviews = [(tokenizer(review), int(label)) for review, label in train_reviews]\n",
    "test_reviews = [(tokenizer(review), int(label)) for review, label in test_reviews]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(['아/Exclamation',\n",
       "   '더빙/Noun',\n",
       "   '../Punctuation',\n",
       "   '진짜/Noun',\n",
       "   '짜증/Noun',\n",
       "   '나다/Verb',\n",
       "   '목소리/Noun'],\n",
       "  0),\n",
       " (['흠/Noun',\n",
       "   '.../Punctuation',\n",
       "   '포스터/Noun',\n",
       "   '보고/Noun',\n",
       "   '초딩/Noun',\n",
       "   '영화/Noun',\n",
       "   '줄/Noun',\n",
       "   '..../Punctuation',\n",
       "   '오버/Noun',\n",
       "   '연기/Noun',\n",
       "   '조차/Josa',\n",
       "   '가볍다/Adjective',\n",
       "   '않다/Verb'],\n",
       "  1),\n",
       " (['너무/Noun',\n",
       "   '재/Noun',\n",
       "   '밓었/Noun',\n",
       "   '다/Josa',\n",
       "   '그래서/Adverb',\n",
       "   '보다/Verb',\n",
       "   '추천/Noun',\n",
       "   '하다/Verb'],\n",
       "  0),\n",
       " (['교도소/Noun',\n",
       "   '이야기/Noun',\n",
       "   '구먼/Noun',\n",
       "   '../Punctuation',\n",
       "   '솔직하다/Adjective',\n",
       "   '재미/Noun',\n",
       "   '는/Josa',\n",
       "   '없다/Adjective',\n",
       "   '../Punctuation',\n",
       "   '평점/Noun',\n",
       "   '조정/Noun'],\n",
       "  0),\n",
       " (['사이몬페그/Noun',\n",
       "   '의/Josa',\n",
       "   '익살/Noun',\n",
       "   '스런/Josa',\n",
       "   '연기/Noun',\n",
       "   '가/Josa',\n",
       "   '돋보이다/Verb',\n",
       "   '영화/Noun',\n",
       "   '!/Punctuation',\n",
       "   '스파이더맨/Noun',\n",
       "   '에서/Josa',\n",
       "   '늙다/Verb',\n",
       "   '보이다/Verb',\n",
       "   '하다/Verb',\n",
       "   '크다/Verb',\n",
       "   '스틴/Noun',\n",
       "   '던스트/Noun',\n",
       "   '가/Josa',\n",
       "   '너무나/Noun',\n",
       "   '도/Josa',\n",
       "   '이쁘다/Adjective',\n",
       "   '보이다/Verb'],\n",
       "  1)]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_reviews[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train = [review for review, _ in train_reviews] # Review Sequence of Train data\n",
    "y_train = [label for _, label in train_reviews] # Label of Train data\n",
    "X_test = [review for review, _ in test_reviews] # Review Sequence of Test data\n",
    "y_test = [label for _, label in test_reviews] # Label of Test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[['아/Exclamation',\n",
       "  '더빙/Noun',\n",
       "  '../Punctuation',\n",
       "  '진짜/Noun',\n",
       "  '짜증/Noun',\n",
       "  '나다/Verb',\n",
       "  '목소리/Noun'],\n",
       " ['흠/Noun',\n",
       "  '.../Punctuation',\n",
       "  '포스터/Noun',\n",
       "  '보고/Noun',\n",
       "  '초딩/Noun',\n",
       "  '영화/Noun',\n",
       "  '줄/Noun',\n",
       "  '..../Punctuation',\n",
       "  '오버/Noun',\n",
       "  '연기/Noun',\n",
       "  '조차/Josa',\n",
       "  '가볍다/Adjective',\n",
       "  '않다/Verb'],\n",
       " ['너무/Noun',\n",
       "  '재/Noun',\n",
       "  '밓었/Noun',\n",
       "  '다/Josa',\n",
       "  '그래서/Adverb',\n",
       "  '보다/Verb',\n",
       "  '추천/Noun',\n",
       "  '하다/Verb']]"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train[:3]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load pre-trained Word2vec Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "model_file = 'w2v/300dim_1minwords_4context_twi' \n",
    "w2v = word2vec.Word2Vec.load(model_file).wv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['./Punctuation',\n",
       " '영화/Noun',\n",
       " '하다/Verb',\n",
       " '이/Josa',\n",
       " '보다/Verb',\n",
       " '의/Josa',\n",
       " '../Punctuation',\n",
       " '에/Josa',\n",
       " '가/Josa',\n",
       " '을/Josa',\n",
       " '도/Josa',\n",
       " '.../Punctuation',\n",
       " '은/Josa',\n",
       " '들/Suffix',\n",
       " ',/Punctuation',\n",
       " '는/Josa',\n",
       " '없다/Adjective',\n",
       " '를/Josa',\n",
       " '있다/Adjective',\n",
       " '좋다/Adjective']"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "w2v.index2word[:20] # word2vec 모델안에 저장된 단어 리스트 (빈도 순으로 내림차 정렬)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "55824"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 전체 단어 갯수\n",
    "len(w2v.index2word)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([  1.15485497e-01,  -9.72887222e-03,  -5.56833856e-03,\n",
       "         1.75525676e-02,  -6.77212402e-02,   1.92349672e-01,\n",
       "        -1.34603202e-01,  -1.91257656e-01,   1.58908665e-01,\n",
       "        -4.98501696e-02,   1.13174751e-01,   7.07398355e-02,\n",
       "        -1.50875881e-01,  -1.18156578e-02,  -1.69227123e-01,\n",
       "         9.30930376e-02,  -1.14399679e-01,  -1.48578174e-02,\n",
       "         8.34368765e-02,  -2.13713288e-01,  -1.65826097e-01,\n",
       "        -8.56175050e-02,   6.88526630e-02,  -4.82723908e-03,\n",
       "         5.24557121e-02,   1.27921864e-01,  -1.63795203e-01,\n",
       "         1.46113053e-01,  -2.27731720e-01,   7.80711919e-02,\n",
       "         5.02348365e-03,  -5.55018410e-02,   2.47064799e-01,\n",
       "         1.34310266e-03,  -5.46680354e-02,  -1.31333932e-01,\n",
       "        -1.93441600e-01,   2.58420229e-01,   2.56266713e-01,\n",
       "        -1.17789015e-01,   3.42658699e-01,   8.53507221e-02,\n",
       "        -8.06954503e-02,   6.00078069e-02,   1.93113104e-01,\n",
       "         1.04480915e-01,  -3.99550870e-02,   7.88475350e-02,\n",
       "        -1.88823506e-01,  -1.33192241e-01,  -6.05737083e-02,\n",
       "        -4.06450257e-02,  -9.53207538e-02,   2.14590713e-01,\n",
       "         1.07909136e-01,   4.77354564e-02,  -2.30905995e-01,\n",
       "         1.46023938e-02,  -8.78749713e-02,   9.05322209e-02,\n",
       "         1.32880166e-01,  -4.20286022e-02,   2.20020026e-01,\n",
       "        -1.85120657e-01,  -4.22908179e-02,  -2.35657319e-01,\n",
       "         1.17383949e-01,  -6.39364645e-02,   1.91948816e-01,\n",
       "         2.55807400e-01,   3.54893133e-03,   5.54829575e-02,\n",
       "         3.06932605e-04,  -9.99246314e-02,   1.42559931e-01,\n",
       "         1.28023028e-02,  -2.00300813e-01,  -2.11234793e-01,\n",
       "         1.14299074e-01,  -1.06211126e-01,  -1.40312210e-01,\n",
       "         7.84162953e-02,   3.37163508e-01,   1.65302947e-01,\n",
       "         4.47856337e-02,  -1.75992489e-01,  -7.91939273e-02,\n",
       "        -4.99228528e-03,  -7.32031018e-02,   9.54040363e-02,\n",
       "        -4.97552902e-02,   2.06284106e-01,   8.81180689e-02,\n",
       "         1.83105713e-03,   1.11310249e-02,  -3.52989770e-02,\n",
       "        -3.67492214e-02,  -1.07369512e-01,  -7.96680450e-02,\n",
       "        -1.79662213e-01,   9.93814245e-02,   1.26140296e-01,\n",
       "         1.44092906e-02,   5.12933880e-02,   4.07808535e-02,\n",
       "         7.89111406e-02,   2.86042035e-01,   5.98724224e-02,\n",
       "         1.12200826e-01,  -3.20644528e-02,   2.37214684e-01,\n",
       "         6.33760020e-02,  -8.80013332e-02,  -1.98463462e-02,\n",
       "         9.44928378e-02,  -3.49671058e-02,  -7.63970390e-02,\n",
       "         1.71033338e-01,  -8.68522078e-02,  -2.08877418e-02,\n",
       "        -2.01017573e-01,   2.78508216e-01,  -1.81243554e-01,\n",
       "        -1.05797641e-01,  -5.18440036e-04,  -2.79276390e-02,\n",
       "        -2.53143460e-02,   4.74799760e-02,   2.32009619e-01,\n",
       "        -7.97004625e-02,  -4.03310172e-03,  -1.15693748e-01,\n",
       "         8.13721046e-02,  -2.78729983e-02,   2.99412012e-02,\n",
       "        -1.30992055e-01,  -1.26432166e-01,   6.48387074e-02,\n",
       "        -2.27928177e-01,   3.07616442e-02,  -1.06125548e-01,\n",
       "        -5.23327217e-02,  -5.96265979e-02,   1.30652338e-01,\n",
       "         8.94280300e-02,  -1.34362206e-01,  -6.56797215e-02,\n",
       "        -4.62407917e-02,   1.71121776e-01,  -1.20300904e-01,\n",
       "        -8.42313915e-02,   1.59327731e-01,  -8.43717828e-02,\n",
       "        -3.90265509e-02,   1.42683819e-01,   1.04979046e-01,\n",
       "         1.40504181e-01,  -1.10552415e-01,  -1.07614473e-02,\n",
       "         2.86752641e-01,  -1.09475307e-01,   1.65257588e-01,\n",
       "         1.89211871e-02,  -1.18457526e-01,  -8.04612041e-02,\n",
       "         2.37820558e-02,   2.13586301e-01,   3.46252583e-02,\n",
       "        -8.72238800e-02,  -1.92899331e-01,   3.25451680e-02,\n",
       "         2.45296545e-02,   9.21931714e-02,   1.20234750e-01,\n",
       "         6.79589733e-02,   4.31217216e-02,  -1.06252044e-01,\n",
       "         4.81806844e-02,  -1.02006853e-01,   1.29484698e-01,\n",
       "        -1.04605876e-01,   1.13596939e-01,   1.55478537e-01,\n",
       "        -5.61273694e-02,   1.82307586e-01,   1.92528352e-01,\n",
       "        -1.11568630e-01,   1.06347144e-01,  -6.30847886e-02,\n",
       "        -3.72889899e-02,  -1.49249703e-01,  -1.97416648e-01,\n",
       "         1.47466455e-02,  -4.41414900e-02,  -5.76090440e-02,\n",
       "         2.17976779e-01,  -2.14700893e-01,   3.60526443e-02,\n",
       "         1.70584634e-01,   2.89417207e-02,   1.91728368e-01,\n",
       "         9.61994082e-02,  -1.36581421e-01,   1.05987107e-02,\n",
       "         1.61323085e-01,   4.69110049e-02,   3.28736395e-01,\n",
       "         1.97201222e-01,  -3.31028014e-01,  -2.09483877e-03,\n",
       "         1.01380900e-01,   7.77546987e-02,  -1.51182815e-01,\n",
       "         6.96311444e-02,  -9.31567848e-02,  -9.54377949e-02,\n",
       "         7.82368779e-02,   1.31242663e-01,   2.24449202e-01,\n",
       "         1.54698566e-01,  -3.87205295e-02,  -4.74480633e-03,\n",
       "         1.29841849e-01,   2.23213553e-01,  -1.66708514e-01,\n",
       "        -1.08406208e-01,   1.56748071e-01,  -1.14397690e-01,\n",
       "        -1.47864386e-01,  -1.20540924e-01,  -1.34064183e-01,\n",
       "         7.99354166e-02,   1.45975441e-01,  -9.52561200e-02,\n",
       "        -3.05800419e-02,  -2.15499982e-01,   1.03048772e-01,\n",
       "        -8.31153095e-02,  -2.72422194e-01,  -1.39074892e-01,\n",
       "         4.22542207e-02,  -1.23891316e-01,  -7.37906396e-02,\n",
       "        -1.68351337e-01,   1.22408181e-01,   7.11716861e-02,\n",
       "        -1.96064979e-01,   1.49034902e-01,  -1.76816374e-01,\n",
       "        -1.67572185e-01,  -1.42139226e-01,  -4.24626982e-03,\n",
       "         8.11210573e-02,   2.43810579e-01,  -2.42919460e-01,\n",
       "        -1.63861103e-02,   1.26451358e-01,   1.76874340e-01,\n",
       "         4.16510478e-02,   1.91798553e-01,  -3.13572623e-02,\n",
       "         3.26219425e-02,   1.83291689e-01,  -3.18218581e-02,\n",
       "         7.18268082e-02,   2.00794905e-01,  -6.18604105e-03,\n",
       "         2.50558648e-02,  -9.41286013e-02,   1.71833597e-02,\n",
       "         7.59971663e-02,   1.03493966e-01,   2.57131875e-01,\n",
       "         2.04962324e-02,  -1.04192771e-01,  -6.81965873e-02,\n",
       "         2.25957725e-02,   9.48282424e-03,  -2.91927122e-02,\n",
       "        -6.08425122e-03,   7.43896049e-03,   3.46372873e-02,\n",
       "         6.54274449e-02,  -2.58177482e-02,  -4.10043716e-01,\n",
       "         2.48455405e-02,   6.14158474e-02,   2.25749686e-01,\n",
       "         2.26292878e-01,   3.79625298e-02,   1.71133876e-01,\n",
       "         7.16445372e-02,  -1.30598405e-02,   2.12686062e-01,\n",
       "         6.40102010e-03,   8.31959248e-02,  -1.69522703e-01,\n",
       "        -2.06538841e-01,  -3.79136205e-02,   6.15510270e-02], dtype=float32)"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "w2v['영화/Noun']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Word - Index Mapping"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'./Punctuation': 1,\n",
       " '영화/Noun': 2,\n",
       " '하다/Verb': 3,\n",
       " '이/Josa': 4,\n",
       " '보다/Verb': 5,\n",
       " '의/Josa': 6,\n",
       " '../Punctuation': 7,\n",
       " '에/Josa': 8,\n",
       " '가/Josa': 9,\n",
       " '을/Josa': 10,\n",
       " '도/Josa': 11,\n",
       " '.../Punctuation': 12,\n",
       " '은/Josa': 13,\n",
       " '들/Suffix': 14,\n",
       " ',/Punctuation': 15,\n",
       " '는/Josa': 16,\n",
       " '없다/Adjective': 17,\n",
       " '를/Josa': 18,\n",
       " '있다/Adjective': 19,\n",
       " '좋다/Adjective': 20,\n",
       " '너무/Noun': 21,\n",
       " '되다/Verb': 22,\n",
       " '이/Determiner': 23,\n",
       " '?/Punctuation': 24,\n",
       " '재밌다/Adjective': 25,\n",
       " '정말/Noun': 26,\n",
       " 'ㅋㅋ/KoreanParticle': 27,\n",
       " '다/Josa': 28,\n",
       " '것/Noun': 29,\n",
       " '이/Noun': 30,\n",
       " '진짜/Noun': 31,\n",
       " '같다/Adjective': 32,\n",
       " '한/Josa': 33,\n",
       " '!/Punctuation': 34,\n",
       " '적/Suffix': 35,\n",
       " '으로/Josa': 36,\n",
       " '아니다/Adjective': 37,\n",
       " '않다/Verb': 38,\n",
       " '로/Josa': 39,\n",
       " '점/Noun': 40,\n",
       " '연기/Noun': 41,\n",
       " '만/Josa': 42,\n",
       " '나오다/Verb': 43,\n",
       " '만들다/Verb': 44,\n",
       " '이렇다/Adjective': 45,\n",
       " '에서/Josa': 46,\n",
       " '평점/Noun': 47,\n",
       " '~/Punctuation': 48,\n",
       " '과/Josa': 49,\n",
       " '최고/Noun': 50,\n",
       " '이다/Verb': 51,\n",
       " '1/Number': 52,\n",
       " '왜/Noun': 53,\n",
       " '생각/Noun': 54,\n",
       " '스토리/Noun': 55,\n",
       " '인/Josa': 56,\n",
       " '..../Punctuation': 57,\n",
       " '드라마/Noun': 58,\n",
       " '이다/Josa': 59,\n",
       " '감동/Noun': 60,\n",
       " '사람/Noun': 61,\n",
       " '보고/Noun': 62,\n",
       " '말/Noun': 63,\n",
       " '더/Noun': 64,\n",
       " '아깝다/Adjective': 65,\n",
       " '내/Noun': 66,\n",
       " '다/Adverb': 67,\n",
       " '때/Noun': 68,\n",
       " '배우/Noun': 69,\n",
       " '고/Josa': 70,\n",
       " '안/Noun': 71,\n",
       " '감독/Noun': 72,\n",
       " '재미있다/Adjective': 73,\n",
       " '그냥/Noun': 74,\n",
       " '한/Determiner': 75,\n",
       " '와/Josa': 76,\n",
       " '게/Josa': 77,\n",
       " '거/Noun': 78,\n",
       " '뭐/Noun': 79,\n",
       " '내용/Noun': 80,\n",
       " '하고/Josa': 81,\n",
       " '봐/Noun': 82,\n",
       " '재미/Noun': 83,\n",
       " '안/VerbPrefix': 84,\n",
       " '시간/Noun': 85,\n",
       " '아/Exclamation': 86,\n",
       " '그/Noun': 87,\n",
       " '!!/Punctuation': 88,\n",
       " '들다/Verb': 89,\n",
       " '재미없다/Adjective': 90,\n",
       " '자다/Verb': 91,\n",
       " '가다/Verb': 92,\n",
       " '좀/Noun': 93,\n",
       " '그렇다/Adjective': 94,\n",
       " '10/Number': 95,\n",
       " '못/VerbPrefix': 96,\n",
       " '지루하다/Adjective': 97,\n",
       " '쓰레기/Noun': 98,\n",
       " '까지/Josa': 99,\n",
       " '주다/Verb': 100,\n",
       " '나/Noun': 101,\n",
       " '2/Number': 102,\n",
       " '수/Noun': 103,\n",
       " '싶다/Verb': 104,\n",
       " '사랑/Noun': 105,\n",
       " '알다/Verb': 106,\n",
       " '작품/Noun': 107,\n",
       " '하나/Noun': 108,\n",
       " '다시/Noun': 109,\n",
       " 'ㅠㅠ/KoreanParticle': 110,\n",
       " '나/Josa': 111,\n",
       " '보다/Josa': 112,\n",
       " '라/Josa': 113,\n",
       " '마지막/Noun': 114,\n",
       " '요/Josa': 115,\n",
       " '볼/Noun': 116,\n",
       " '지/Josa': 117,\n",
       " 'ㅋ/KoreanParticle': 118,\n",
       " '이다/Adjective': 119,\n",
       " '3/Number': 120,\n",
       " '이렇게/Adverb': 121,\n",
       " '모르다/Verb': 122,\n",
       " '처음/Noun': 123,\n",
       " '정도/Noun': 124,\n",
       " '완전/Noun': 125,\n",
       " '많다/Adjective': 126,\n",
       " '나다/Verb': 127,\n",
       " '장면/Noun': 128,\n",
       " '차다/Verb': 129,\n",
       " '액션/Noun': 130,\n",
       " '인데/Josa': 131,\n",
       " '지금/Noun': 132,\n",
       " '보이다/Verb': 133,\n",
       " '최악/Noun': 134,\n",
       " \"'/Punctuation\": 135,\n",
       " '이야기/Noun': 136,\n",
       " '별로/Noun': 137,\n",
       " '중/Suffix': 138,\n",
       " '돈/Noun': 139,\n",
       " '임/Noun': 140,\n",
       " '연출/Noun': 141,\n",
       " '난/Noun': 142,\n",
       " '끝/Noun': 143,\n",
       " '느낌/Noun': 144,\n",
       " '좋아하다/Adjective': 145,\n",
       " '받다/Verb': 146,\n",
       " '쓰다/Verb': 147,\n",
       " '넘다/Verb': 148,\n",
       " '듯/Noun': 149,\n",
       " '걸/Noun': 150,\n",
       " '개/Noun': 151,\n",
       " '명작/Noun': 152,\n",
       " 'ㅎㅎ/KoreanParticle': 153,\n",
       " '역시/Noun': 154,\n",
       " '그리고/Conjunction': 155,\n",
       " '년/Noun': 156,\n",
       " '주인공/Noun': 157,\n",
       " '함/Noun': 158,\n",
       " '서/Josa': 159,\n",
       " '이해/Noun': 160,\n",
       " '남다/Verb': 161,\n",
       " '^^/Punctuation': 162,\n",
       " 'ㅡㅡ/KoreanParticle': 163,\n",
       " '많이/Adverb': 164,\n",
       " '먹다/Verb': 165,\n",
       " '괜찮다/Adjective': 166,\n",
       " '어떻다/Adjective': 167,\n",
       " '!!!/Punctuation': 168,\n",
       " '이영화/Noun': 169,\n",
       " '버리다/Verb': 170,\n",
       " '그/Determiner': 171,\n",
       " '라고/Josa': 172,\n",
       " '느끼다/Verb': 173,\n",
       " '또/Noun': 174,\n",
       " '여자/Noun': 175,\n",
       " '때문/Noun': 176,\n",
       " '잘/VerbPrefix': 177,\n",
       " '분/Noun': 178,\n",
       " '아름답다/Adjective': 179,\n",
       " '못/Noun': 180,\n",
       " '중/Noun': 181,\n",
       " '기억/Noun': 182,\n",
       " '내/Determiner': 183,\n",
       " '야/Josa': 184,\n",
       " '이고/Josa': 185,\n",
       " '에게/Josa': 186,\n",
       " '보기/Noun': 187,\n",
       " '와/Noun': 188,\n",
       " '편/Noun': 189,\n",
       " '부터/Josa': 190,\n",
       " '무슨/Noun': 191,\n",
       " '이나/Josa': 192,\n",
       " '엔/Josa': 193,\n",
       " '인생/Noun': 194,\n",
       " '결말/Noun': 195,\n",
       " '마음/Noun': 196,\n",
       " '꼭/Noun': 197,\n",
       " '랑/Josa': 198,\n",
       " '현실/Noun': 199,\n",
       " '애/Noun': 200,\n",
       " '유치/Noun': 201,\n",
       " '짜증/Noun': 202,\n",
       " '뻔하다/Adjective': 203,\n",
       " '소재/Noun': 204,\n",
       " '아쉽다/Adjective': 205,\n",
       " '수준/Noun': 206,\n",
       " '맞다/Verb': 207,\n",
       " 'ㅠ/KoreanParticle': 208,\n",
       " '하지만/Conjunction': 209,\n",
       " '반전/Noun': 210,\n",
       " '~~/Punctuation': 211,\n",
       " '어리다/Verb': 212,\n",
       " '무섭다/Adjective': 213,\n",
       " '허다/Verb': 214,\n",
       " '없이/Adverb': 215,\n",
       " '면/Josa': 216,\n",
       " '재다/Verb': 217,\n",
       " '냐/Josa': 218,\n",
       " '네/Josa': 219,\n",
       " '님/Suffix': 220,\n",
       " '한번/Noun': 221,\n",
       " '눈물/Noun': 222,\n",
       " '인지/Josa': 223,\n",
       " ';;/Punctuation': 224,\n",
       " '끄다/Verb': 225,\n",
       " '웃기다/Verb': 226,\n",
       " '매력/Noun': 227,\n",
       " '전개/Noun': 228,\n",
       " '끝나다/Verb': 229,\n",
       " '??/Punctuation': 230,\n",
       " '아/Josa': 231,\n",
       " '한국/Noun': 232,\n",
       " '가장/Noun': 233,\n",
       " '남자/Noun': 234,\n",
       " '일본/Noun': 235,\n",
       " '성/Suffix': 236,\n",
       " '가슴/Noun': 237,\n",
       " '낮다/Adjective': 238,\n",
       " '슬프다/Adjective': 239,\n",
       " '알/Noun': 240,\n",
       " '따다/Verb': 241,\n",
       " '라는/Josa': 242,\n",
       " '전/Noun': 243,\n",
       " '지다/Verb': 244,\n",
       " '이/Suffix': 245,\n",
       " '높다/Adjective': 246,\n",
       " '죽다/Verb': 247,\n",
       " '저/Noun': 248,\n",
       " '원작/Noun': 249,\n",
       " '다른/Noun': 250,\n",
       " '멋지다/Adjective': 251,\n",
       " '아이/Noun': 252,\n",
       " '말다/Verb': 253,\n",
       " '크다/Verb': 254,\n",
       " '짱/Noun': 255,\n",
       " '짜다/Verb': 256,\n",
       " '기대/Noun': 257,\n",
       " '음악/Noun': 258,\n",
       " '오다/Verb': 259,\n",
       " '되어다/Verb': 260,\n",
       " '해주다/Verb': 261,\n",
       " '별/Noun': 262,\n",
       " '인간/Noun': 263,\n",
       " '인가/Josa': 264,\n",
       " '잘하다/Verb': 265,\n",
       " '줄/Noun': 266,\n",
       " '내다/Verb': 267,\n",
       " '우리/Noun': 268,\n",
       " '야하다/Adjective': 269,\n",
       " '솔직하다/Adjective': 270,\n",
       " '추천/Noun': 271,\n",
       " '속/Noun': 272,\n",
       " '저/Determiner': 273,\n",
       " '실망/Noun': 274,\n",
       " '4/Number': 275,\n",
       " '5/Number': 276,\n",
       " '모든/Noun': 277,\n",
       " '게/Noun': 278,\n",
       " '자체/Noun': 279,\n",
       " '씨/Suffix': 280,\n",
       " '영상/Noun': 281,\n",
       " '요/Noun': 282,\n",
       " '화/Suffix': 283,\n",
       " '빠지다/Verb': 284,\n",
       " '네/Suffix': 285,\n",
       " '도/Noun': 286,\n",
       " '치다/Verb': 287,\n",
       " '대박/Noun': 288,\n",
       " '해/Josa': 289,\n",
       " '찍다/Verb': 290,\n",
       " '일/Noun': 291,\n",
       " '대한/Noun': 292,\n",
       " '눈/Noun': 293,\n",
       " '번/Noun': 294,\n",
       " '코미디/Noun': 295,\n",
       " '캐릭터/Noun': 296,\n",
       " '보지/Noun': 297,\n",
       " '이네/Josa': 298,\n",
       " '모두/Noun': 299,\n",
       " '해서/Josa': 300,\n",
       " '지만/Josa': 301,\n",
       " '그래도/Adverb': 302,\n",
       " '돼다/Verb': 303,\n",
       " '뭔가/Noun': 304,\n",
       " '연기력/Noun': 305,\n",
       " '이상/Noun': 306,\n",
       " '이쁘다/Adjective': 307,\n",
       " '아직/Noun': 308,\n",
       " '모습/Noun': 309,\n",
       " '여운/Noun': 310,\n",
       " '전혀/Noun': 311,\n",
       " '미치다/Adjective': 312,\n",
       " '움/Noun': 313,\n",
       " '귀엽다/Adjective': 314,\n",
       " '개봉/Noun': 315,\n",
       " '굿/Noun': 316,\n",
       " '근데/Adverb': 317,\n",
       " '공감/Noun': 318,\n",
       " '중간/Noun': 319,\n",
       " '기대하다/Adjective': 320,\n",
       " '표현/Noun': 321,\n",
       " '시리즈/Noun': 322,\n",
       " 'ㅜㅜ/KoreanParticle': 323,\n",
       " '시키다/Verb': 324,\n",
       " '계속/Noun': 325,\n",
       " '준/Noun': 326,\n",
       " '에는/Josa': 327,\n",
       " '아주/Noun': 328,\n",
       " '-/Punctuation': 329,\n",
       " '이제/Noun': 330,\n",
       " '살다/Verb': 331,\n",
       " '(/Punctuation': 332,\n",
       " '늘다/Verb': 333,\n",
       " '부분/Noun': 334,\n",
       " '위/Noun': 335,\n",
       " '기/Noun': 336,\n",
       " '감/Noun': 337,\n",
       " ';/Punctuation': 338,\n",
       " ')/Punctuation': 339,\n",
       " '7/Number': 340,\n",
       " '이랑/Josa': 341,\n",
       " '제목/Noun': 342,\n",
       " '힘들다/Adjective': 343,\n",
       " '가족/Noun': 344,\n",
       " 'OO/Alpha': 345,\n",
       " '진심/Noun': 346,\n",
       " 'ㅎ/KoreanParticle': 347,\n",
       " '이라/Josa': 348,\n",
       " '내내/Noun': 349,\n",
       " '기분/Noun': 350,\n",
       " '0/Number': 351,\n",
       " '대단하다/Adjective': 352,\n",
       " '용/Noun': 353,\n",
       " '건지다/Verb': 354,\n",
       " '믿다/Verb': 355,\n",
       " '밖에/Josa': 356,\n",
       " '처럼/Josa': 357,\n",
       " '두/Determiner': 358,\n",
       " '\"\"\"\"/Punctuation': 359,\n",
       " '몰입/Noun': 360,\n",
       " '대사/Noun': 361,\n",
       " '억지/Noun': 362,\n",
       " ',,/Punctuation': 363,\n",
       " '대다/Verb': 364,\n",
       " '후/Noun': 365,\n",
       " '급/Noun': 366,\n",
       " '여주/Noun': 367,\n",
       " '작가/Noun': 368,\n",
       " '제/Noun': 369,\n",
       " '이라는/Josa': 370,\n",
       " '욕/Noun': 371,\n",
       " '어이/Noun': 372,\n",
       " '공포/Noun': 373,\n",
       " '떨어지다/Verb': 374,\n",
       " '스럽다/Adjective': 375,\n",
       " '8/Number': 376,\n",
       " '요즘/Noun': 377,\n",
       " '점도/Noun': 378,\n",
       " '웃다/Verb': 379,\n",
       " '울다/Verb': 380,\n",
       " '긴장감/Noun': 381,\n",
       " '비/Noun': 382,\n",
       " '살/Noun': 383,\n",
       " '조금/Noun': 384,\n",
       " '애니/Noun': 385,\n",
       " '스릴러/Noun': 386,\n",
       " '잔잔/Noun': 387,\n",
       " '개인/Noun': 388,\n",
       " '무엇/Noun': 389,\n",
       " '극장/Noun': 390,\n",
       " '시작/Noun': 391,\n",
       " '특히/Adverb': 392,\n",
       " '잼/Noun': 393,\n",
       " '데/Noun': 394,\n",
       " '한테/Josa': 395,\n",
       " '삶/Noun': 396,\n",
       " '노래/Noun': 397,\n",
       " '이유/Noun': 398,\n",
       " '이상하다/Adjective': 399,\n",
       " 'ㅡ/KoreanParticle': 400,\n",
       " '알바/Noun': 401,\n",
       " '깊다/Adjective': 402,\n",
       " '나름/Noun': 403,\n",
       " '의/Noun': 404,\n",
       " '제일/Noun': 405,\n",
       " '제대로/Noun': 406,\n",
       " '뿐/Noun': 407,\n",
       " '예쁘다/Adjective': 408,\n",
       " '물/Noun': 409,\n",
       " '당시/Noun': 410,\n",
       " '놈/Noun': 411,\n",
       " '차라리/Noun': 412,\n",
       " '점수/Noun': 413,\n",
       " '에도/Josa': 414,\n",
       " '...../Punctuation': 415,\n",
       " '쓸다/Verb': 416,\n",
       " '남/Noun': 417,\n",
       " '보/Noun': 418,\n",
       " '느껴지다/Verb': 419,\n",
       " '노잼/Noun': 420,\n",
       " '오랜만/Noun': 421,\n",
       " '친구/Noun': 422,\n",
       " '막장/Noun': 423,\n",
       " '접/Noun': 424,\n",
       " '따뜻하다/Adjective': 425,\n",
       " '빼다/Verb': 426,\n",
       " '의미/Noun': 427,\n",
       " '건/Noun': 428,\n",
       " '영화로/Noun': 429,\n",
       " '세상/Noun': 430,\n",
       " '어/Noun': 431,\n",
       " '\"\"\"/Punctuation': 432,\n",
       " '력/Suffix': 433,\n",
       " '같이/Josa': 434,\n",
       " '절대/Noun': 435,\n",
       " '훌륭하다/Adjective': 436,\n",
       " '공포영화/Noun': 437,\n",
       " '이라고/Josa': 438,\n",
       " '잇다/Verb': 439,\n",
       " '시대/Noun': 440,\n",
       " '너무나/Noun': 441,\n",
       " '다르다/Adjective': 442,\n",
       " '류/Noun': 443,\n",
       " '년대/Noun': 444,\n",
       " '찾다/Verb': 445,\n",
       " '미국/Noun': 446,\n",
       " '♥/Foreign': 447,\n",
       " '웃음/Noun': 448,\n",
       " '9/Number': 449,\n",
       " '대/Suffix': 450,\n",
       " '뭘/Noun': 451,\n",
       " '도대체/Noun': 452,\n",
       " '생각나다/Verb': 453,\n",
       " '우리나라/Noun': 454,\n",
       " '앞/Noun': 455,\n",
       " '기다/Verb': 456,\n",
       " '그리다/Verb': 457,\n",
       " '설정/Noun': 458,\n",
       " '이란/Josa': 459,\n",
       " '6/Number': 460,\n",
       " '걍/Adverb': 461,\n",
       " '수작/Noun': 462,\n",
       " '명/Noun': 463,\n",
       " '초반/Noun': 464,\n",
       " '그저/Adverb': 465,\n",
       " '망/Noun': 466,\n",
       " '엄청/Adverb': 467,\n",
       " '출연/Noun': 468,\n",
       " '딱/Adverb': 469,\n",
       " '죽이다/Verb': 470,\n",
       " '해보다/Verb': 471,\n",
       " '사실/Noun': 472,\n",
       " '밉다/Adjective': 473,\n",
       " '싫다/Adjective': 474,\n",
       " '마르다/Adjective': 475,\n",
       " '아프다/Adjective': 476,\n",
       " '이야/Josa': 477,\n",
       " '시나리오/Noun': 478,\n",
       " '필요없다/Adjective': 479,\n",
       " '장난/Noun': 480,\n",
       " '소름/Noun': 481,\n",
       " '어쩌다/Verb': 482,\n",
       " '소리/Noun': 483,\n",
       " '한국영/Noun': 484,\n",
       " '관객/Noun': 485,\n",
       " '신선하다/Adjective': 486,\n",
       " '서다/Verb': 487,\n",
       " '답답하다/Adjective': 488,\n",
       " 'OOO/Alpha': 489,\n",
       " '분위기/Noun': 490,\n",
       " '봄/Noun': 491,\n",
       " '추억/Noun': 492,\n",
       " '강추/Noun': 493,\n",
       " '더럽다/Adjective': 494,\n",
       " '멋있다/Adjective': 495,\n",
       " '문제/Noun': 496,\n",
       " '나쁘다/Adjective': 497,\n",
       " '가지/Noun': 498,\n",
       " '뭔/NounPrefix': 499,\n",
       " '몰르다/Verb': 500,\n",
       " '최고다/Noun': 501,\n",
       " '나가다/Verb': 502,\n",
       " '졸작/Noun': 503,\n",
       " '지루함/Noun': 504,\n",
       " '결국/Adverb': 505,\n",
       " '니/Josa': 506,\n",
       " '마다/Josa': 507,\n",
       " '정신/Noun': 508,\n",
       " '자신/Noun': 509,\n",
       " '인상/Noun': 510,\n",
       " '제발/Noun': 511,\n",
       " '팬/Noun': 512,\n",
       " '놓다/Verb': 513,\n",
       " '감정/Noun': 514,\n",
       " '20/Number': 515,\n",
       " '부족하다/Adjective': 516,\n",
       " '무/Noun': 517,\n",
       " '영화관/Noun': 518,\n",
       " '에서/Noun': 519,\n",
       " '안타깝다/Adjective': 520,\n",
       " '엄마/Noun': 521,\n",
       " '만/Suffix': 522,\n",
       " '만/Noun': 523,\n",
       " '위해/Noun': 524,\n",
       " '해/Noun': 525,\n",
       " '적다/Verb': 526,\n",
       " '~~~/Punctuation': 527,\n",
       " '머/Noun': 528,\n",
       " '오늘/Noun': 529,\n",
       " '이지만/Josa': 530,\n",
       " '주연/Noun': 531,\n",
       " '더빙/Noun': 532,\n",
       " '잊다/Verb': 533,\n",
       " '아무리/Adverb': 534,\n",
       " '라면/Noun': 535,\n",
       " '후회/Noun': 536,\n",
       " '니까/Josa': 537,\n",
       " '했지만/Josa': 538,\n",
       " '엔딩/Noun': 539,\n",
       " '포스터/Noun': 540,\n",
       " '휴/Exclamation': 541,\n",
       " '짱/Suffix': 542,\n",
       " '코믹/Noun': 543,\n",
       " '평가/Noun': 544,\n",
       " '로/Noun': 545,\n",
       " '엉/Exclamation': 546,\n",
       " '캐스팅/Noun': 547,\n",
       " '어색하다/Adjective': 548,\n",
       " '그렇게/Adverb': 549,\n",
       " '자/Suffix': 550,\n",
       " '든/Josa': 551,\n",
       " '배우다/Verb': 552,\n",
       " '진부/Noun': 553,\n",
       " '얼마나/Noun': 554,\n",
       " '애니메이션/Noun': 555,\n",
       " '뒤/Noun': 556,\n",
       " '킬링타임/Noun': 557,\n",
       " '진정/Noun': 558,\n",
       " '등/Noun': 559,\n",
       " '!!!!/Punctuation': 560,\n",
       " '구성/Noun': 561,\n",
       " '찌다/Verb': 562,\n",
       " '들이다/Verb': 563,\n",
       " '살리다/Verb': 564,\n",
       " '배경/Noun': 565,\n",
       " '개연/Noun': 566,\n",
       " '날/Noun': 567,\n",
       " '를/Noun': 568,\n",
       " '음/Noun': 569,\n",
       " '스릴/Noun': 570,\n",
       " '티비/Noun': 571,\n",
       " '충격/Noun': 572,\n",
       " '갈수록/Noun': 573,\n",
       " '얘기/Noun': 574,\n",
       " 'ㅉㅉ/KoreanParticle': 575,\n",
       " '뿐/Suffix': 576,\n",
       " '전쟁/Noun': 577,\n",
       " '만화/Noun': 578,\n",
       " '여/Josa': 579,\n",
       " '몰입도/Noun': 580,\n",
       " '옛날/Noun': 581,\n",
       " '이름/Noun': 582,\n",
       " '완벽하다/Adjective': 583,\n",
       " '달다/Verb': 584,\n",
       " '여기/Noun': 585,\n",
       " '얼굴/Noun': 586,\n",
       " '예술/Noun': 587,\n",
       " '머리/Noun': 588,\n",
       " '보단/Josa': 589,\n",
       " 'ㅜ/KoreanParticle': 590,\n",
       " '놀라다/Verb': 591,\n",
       " '맘/Noun': 592,\n",
       " '언제/Noun': 593,\n",
       " '웃기/Noun': 594,\n",
       " '함께/Adverb': 595,\n",
       " '두/Noun': 596,\n",
       " '행복하다/Adjective': 597,\n",
       " '니/Noun': 598,\n",
       " '영화인/Noun': 599,\n",
       " '낭비/Noun': 600,\n",
       " '제작/Noun': 601,\n",
       " '진/Noun': 602,\n",
       " '장르/Noun': 603,\n",
       " '너/Noun': 604,\n",
       " '어울리다/Verb': 605,\n",
       " '질/Noun': 606,\n",
       " '자/Noun': 607,\n",
       " '읽다/Verb': 608,\n",
       " '어디/Noun': 609,\n",
       " '책/Noun': 610,\n",
       " '그래서/Adverb': 611,\n",
       " '이후/Noun': 612,\n",
       " '집중/Noun': 613,\n",
       " '후반/Noun': 614,\n",
       " '즐겁다/Adjective': 615,\n",
       " '그것/Noun': 616,\n",
       " '다큐/Noun': 617,\n",
       " '매우/Noun': 618,\n",
       " '집/Noun': 619,\n",
       " '상황/Noun': 620,\n",
       " '역사/Noun': 621,\n",
       " '인공/Noun': 622,\n",
       " '판/Noun': 623,\n",
       " '+/Punctuation': 624,\n",
       " '둘/Noun': 625,\n",
       " '꽤/Noun': 626,\n",
       " ',,,/Punctuation': 627,\n",
       " '누구/Noun': 628,\n",
       " '작/Noun': 629,\n",
       " '시즌/Noun': 630,\n",
       " '급/Suffix': 631,\n",
       " '터지다/Verb': 632,\n",
       " '새롭다/Adjective': 633,\n",
       " '???/Punctuation': 634,\n",
       " '잡다/Verb': 635,\n",
       " '식/Suffix': 636,\n",
       " '심하다/Adjective': 637,\n",
       " '충분하다/Adjective': 638,\n",
       " '건가/Noun': 639,\n",
       " '다운/Noun': 640,\n",
       " '걸작/Noun': 641,\n",
       " '영/Noun': 642,\n",
       " '적/Noun': 643,\n",
       " '멀다/Adjective': 644,\n",
       " '존나/Noun': 645,\n",
       " '부/Noun': 646,\n",
       " '대체/Noun': 647,\n",
       " '이리/Adverb': 648,\n",
       " '예전/Noun': 649,\n",
       " '분들/Suffix': 650,\n",
       " '몇/NounPrefix': 651,\n",
       " '약간/Noun': 652,\n",
       " '별점/Noun': 653,\n",
       " '불쌍하다/Adjective': 654,\n",
       " '감사/Noun': 655,\n",
       " 'B/Alpha': 656,\n",
       " '실화/Noun': 657,\n",
       " '궁금하다/Adjective': 658,\n",
       " '신/Noun': 659,\n",
       " '어느/Adverb': 660,\n",
       " '씬/Noun': 661,\n",
       " '한마디/Noun': 662,\n",
       " '그나마/Adverb': 663,\n",
       " '자기/Noun': 664,\n",
       " '인물/Noun': 665,\n",
       " '훨씬/Adverb': 666,\n",
       " '방송/Noun': 667,\n",
       " '생기다/Verb': 668,\n",
       " '누가/Noun': 669,\n",
       " '만점/Noun': 670,\n",
       " '비슷하다/Adjective': 671,\n",
       " '뜨다/Verb': 672,\n",
       " '시절/Noun': 673,\n",
       " '남기다/Verb': 674,\n",
       " '쉬다/Verb': 675,\n",
       " '~!/Punctuation': 676,\n",
       " '드리다/Verb': 677,\n",
       " '끼다/Verb': 678,\n",
       " '어렵다/Adjective': 679,\n",
       " '네/Determiner': 680,\n",
       " '그대로/Noun': 681,\n",
       " '유쾌/Noun': 682,\n",
       " '비교/Noun': 683,\n",
       " '히/Adverb': 684,\n",
       " '대해/Noun': 685,\n",
       " '주제/Noun': 686,\n",
       " '비디오/Noun': 687,\n",
       " '초딩/Noun': 688,\n",
       " '발연기/Noun': 689,\n",
       " '극/NounPrefix': 690,\n",
       " '사회/Noun': 691,\n",
       " '아들/Noun': 692,\n",
       " '담다/Verb': 693,\n",
       " '30/Number': 694,\n",
       " '필요하다/Adjective': 695,\n",
       " '인/Noun': 696,\n",
       " '만나다/Verb': 697,\n",
       " '로맨스/Noun': 698,\n",
       " '순수하다/Adjective': 699,\n",
       " '네이버/Noun': 700,\n",
       " '상당하다/Adjective': 701,\n",
       " '원/Noun': 702,\n",
       " '전체/Noun': 703,\n",
       " '타다/Verb': 704,\n",
       " '감성/Noun': 705,\n",
       " '소설/Noun': 706,\n",
       " '다음/Noun': 707,\n",
       " 'CG/Alpha': 708,\n",
       " '굉장하다/Adjective': 709,\n",
       " '꿈/Noun': 710,\n",
       " '간만/Noun': 711,\n",
       " '란/Noun': 712,\n",
       " '혼자/Noun': 713,\n",
       " '떠나다/Verb': 714,\n",
       " '시/Noun': 715,\n",
       " '//Punctuation': 716,\n",
       " 'ㄷㄷ/KoreanParticle': 717,\n",
       " '글/Noun': 718,\n",
       " '은/Noun': 719,\n",
       " '두다/Verb': 720,\n",
       " '회/Noun': 721,\n",
       " '코/Noun': 722,\n",
       " '여배우/Noun': 723,\n",
       " '라도/Josa': 724,\n",
       " '총/Noun': 725,\n",
       " '100/Number': 726,\n",
       " '힘/Noun': 727,\n",
       " '나서다/Verb': 728,\n",
       " '편이/Noun': 729,\n",
       " '반/Noun': 730,\n",
       " '나이/Noun': 731,\n",
       " '돌리다/Verb': 732,\n",
       " '굳다/Adjective': 733,\n",
       " '바로/Noun': 734,\n",
       " '스/Noun': 735,\n",
       " '세/Noun': 736,\n",
       " '해도/Noun': 737,\n",
       " ';;;/Punctuation': 738,\n",
       " '이르다/Verb': 739,\n",
       " '넘치다/Verb': 740,\n",
       " '성룡/Noun': 741,\n",
       " '구/Noun': 742,\n",
       " '당신/Noun': 743,\n",
       " '보다는/Josa': 744,\n",
       " '오래/Noun': 745,\n",
       " '삼류/Noun': 746,\n",
       " '돋다/Adjective': 747,\n",
       " '목소리/Noun': 748,\n",
       " '화려하다/Adjective': 749,\n",
       " '빨리/Adverb': 750,\n",
       " '사랑스럽다/Adjective': 751,\n",
       " '에요/Josa': 752,\n",
       " '지나다/Verb': 753,\n",
       " '당하다/Adjective': 754,\n",
       " '엄청나다/Adjective': 755,\n",
       " '이라도/Josa': 756,\n",
       " '교훈/Noun': 757,\n",
       " '죠/Josa': 758,\n",
       " '풀다/Verb': 759,\n",
       " '년도/Noun': 760,\n",
       " '울/Noun': 761,\n",
       " '만하/Noun': 762,\n",
       " '순간/Noun': 763,\n",
       " '안보/Noun': 764,\n",
       " '답/Noun': 765,\n",
       " '어설픈/Noun': 766,\n",
       " '동안/Noun': 767,\n",
       " '선/Noun': 768,\n",
       " '존재/Noun': 769,\n",
       " '거의/Noun': 770,\n",
       " '년전/Noun': 771,\n",
       " '올리다/Verb': 772,\n",
       " '어른/Noun': 773,\n",
       " '그만/Adverb': 774,\n",
       " '전부/Noun': 775,\n",
       " '기도/Noun': 776,\n",
       " '가볍다/Adjective': 777,\n",
       " '질질/Noun': 778,\n",
       " '오/Noun': 779,\n",
       " '몇번/Noun': 780,\n",
       " '90/Number': 781,\n",
       " '스러운/Josa': 782,\n",
       " '....../Punctuation': 783,\n",
       " '라니/Josa': 784,\n",
       " '기다리다/Verb': 785,\n",
       " '독특하다/Adjective': 786,\n",
       " '재/Noun': 787,\n",
       " '말고/Josa': 788,\n",
       " '아버지/Noun': 789,\n",
       " '좀비/Noun': 790,\n",
       " '초/Noun': 791,\n",
       " '줄거리/Noun': 792,\n",
       " '실제/Noun': 793,\n",
       " '점주/Noun': 794,\n",
       " '각본/Noun': 795,\n",
       " '상/Suffix': 796,\n",
       " '죽/Noun': 797,\n",
       " '판타지/Noun': 798,\n",
       " '맞추다/Verb': 799,\n",
       " '성우/Noun': 800,\n",
       " '망치다/Verb': 801,\n",
       " '평론가/Noun': 802,\n",
       " '성/Noun': 803,\n",
       " '저런/Noun': 804,\n",
       " '시청률/Noun': 805,\n",
       " '잔인/Noun': 806,\n",
       " '에선/Josa': 807,\n",
       " '흥미롭다/Adjective': 808,\n",
       " '상영/Noun': 809,\n",
       " '평/Noun': 810,\n",
       " '이지/Noun': 811,\n",
       " '쯤/Suffix': 812,\n",
       " '그러나/Conjunction': 813,\n",
       " '까지/Noun': 814,\n",
       " '이번/Noun': 815,\n",
       " '으로도/Josa': 816,\n",
       " '기대다/Verb': 817,\n",
       " '치고/Josa': 818,\n",
       " '갑자기/Noun': 819,\n",
       " '세계/Noun': 820,\n",
       " '허무/Noun': 821,\n",
       " '가치/Noun': 822,\n",
       " '딸/Noun': 823,\n",
       " '바꾸다/Verb': 824,\n",
       " '곳/Noun': 825,\n",
       " '극장판/Noun': 826,\n",
       " '발/Noun': 827,\n",
       " '의도/Noun': 828,\n",
       " '넣다/Verb': 829,\n",
       " '중국/Noun': 830,\n",
       " 'TV/Alpha': 831,\n",
       " '필요/Noun': 832,\n",
       " '바라다/Verb': 833,\n",
       " '간/Suffix': 834,\n",
       " '요소/Noun': 835,\n",
       " '보내다/Verb': 836,\n",
       " '이하/Noun': 837,\n",
       " '이기다/Verb': 838,\n",
       " '맛/Noun': 839,\n",
       " '오글거리다/Adjective': 840,\n",
       " '화면/Noun': 841,\n",
       " '한편/Noun': 842,\n",
       " '간다/Noun': 843,\n",
       " '그녀/Noun': 844,\n",
       " '까다/Verb': 845,\n",
       " '래/Josa': 846,\n",
       " '개/Suffix': 847,\n",
       " '온/Noun': 848,\n",
       " '만들어지다/Verb': 849,\n",
       " '그때/Noun': 850,\n",
       " '항상/Noun': 851,\n",
       " '에서도/Josa': 852,\n",
       " '불다/Verb': 853,\n",
       " '볼때/Noun': 854,\n",
       " '들어가다/Verb': 855,\n",
       " '따위/Noun': 856,\n",
       " '선택/Noun': 857,\n",
       " '단순하다/Adjective': 858,\n",
       " '역/Noun': 859,\n",
       " '거리/Noun': 860,\n",
       " '땜/Noun': 861,\n",
       " '복수/Noun': 862,\n",
       " '밑/Noun': 863,\n",
       " '흥행/Noun': 864,\n",
       " '잘못/Noun': 865,\n",
       " '점준/Noun': 866,\n",
       " '고/Noun': 867,\n",
       " '탄탄/Noun': 868,\n",
       " '막/Noun': 869,\n",
       " '일단/Noun': 870,\n",
       " '스타일/Noun': 871,\n",
       " '거기/Noun': 872,\n",
       " '이에요/Josa': 873,\n",
       " '갖다/Verb': 874,\n",
       " '나라/Noun': 875,\n",
       " '에서는/Josa': 876,\n",
       " '듣다/Verb': 877,\n",
       " '여러/Noun': 878,\n",
       " '가보다/Verb': 879,\n",
       " '걸리다/Verb': 880,\n",
       " '장/Suffix': 881,\n",
       " '멜로/Noun': 882,\n",
       " '피/Noun': 883,\n",
       " '부르다/Verb': 884,\n",
       " '걸다/Verb': 885,\n",
       " '첨/Adverb': 886,\n",
       " '%/Punctuation': 887,\n",
       " '인거/Josa': 888,\n",
       " '어디서/Adverb': 889,\n",
       " '메다/Verb': 890,\n",
       " '참고/Noun': 891,\n",
       " '점점/Noun': 892,\n",
       " '확실하다/Adjective': 893,\n",
       " '아빠/Noun': 894,\n",
       " '또한/Noun': 895,\n",
       " '사건/Noun': 896,\n",
       " '똥/Noun': 897,\n",
       " '역대/Noun': 898,\n",
       " '예상/Noun': 899,\n",
       " '라서/Josa': 900,\n",
       " '더욱/Noun': 901,\n",
       " '술/Noun': 902,\n",
       " '갈다/Verb': 903,\n",
       " '관람/Noun': 904,\n",
       " 'tv/Alpha': 905,\n",
       " '불편하다/Adjective': 906,\n",
       " '손/Noun': 907,\n",
       " '즐기다/Verb': 908,\n",
       " '여/Noun': 909,\n",
       " '물론/Noun': 910,\n",
       " '잃다/Verb': 911,\n",
       " '낫다/Noun': 912,\n",
       " '완성/Noun': 913,\n",
       " '개그/Noun': 914,\n",
       " '80/Number': 915,\n",
       " '진행/Noun': 916,\n",
       " '뻔/Noun': 917,\n",
       " '똑같다/Adjective': 918,\n",
       " '리/Noun': 919,\n",
       " '돋보이다/Verb': 920,\n",
       " ':/Punctuation': 921,\n",
       " '짧다/Adjective': 922,\n",
       " '편집/Noun': 923,\n",
       " '병맛/Noun': 924,\n",
       " '몇/Noun': 925,\n",
       " '에겐/Josa': 926,\n",
       " '평범하다/Adjective': 927,\n",
       " '만에/Josa': 928,\n",
       " '때리다/Verb': 929,\n",
       " '화이팅/Noun': 930,\n",
       " '엉망/Noun': 931,\n",
       " '댓글/Noun': 932,\n",
       " '중요하다/Adjective': 933,\n",
       " '전형/Noun': 934,\n",
       " '귀신/Noun': 935,\n",
       " '관/Noun': 936,\n",
       " '지겹다/Adjective': 937,\n",
       " '뛰어나다/Adjective': 938,\n",
       " '수가/Noun': 939,\n",
       " '프랑스/Noun': 940,\n",
       " '여준/Noun': 941,\n",
       " '찾아보다/Verb': 942,\n",
       " '싸우다/Verb': 943,\n",
       " '본영/Noun': 944,\n",
       " '벗다/Verb': 945,\n",
       " '꿀잼/Noun': 946,\n",
       " '미화/Noun': 947,\n",
       " '바/Noun': 948,\n",
       " '중반/Noun': 949,\n",
       " '적당하다/Adjective': 950,\n",
       " '일이/Noun': 951,\n",
       " '그게/Noun': 952,\n",
       " '봣/Noun': 953,\n",
       " '원래/Noun': 954,\n",
       " '햇/Noun': 955,\n",
       " '자극/Noun': 956,\n",
       " '~!!/Punctuation': 957,\n",
       " '열/Noun': 958,\n",
       " '이라니/Josa': 959,\n",
       " '가지다/Verb': 960,\n",
       " '감상/Noun': 961,\n",
       " '몸/Noun': 962,\n",
       " '첫/Noun': 963,\n",
       " '그래픽/Noun': 964,\n",
       " '죽음/Noun': 965,\n",
       " '식상하다/Adjective': 966,\n",
       " '타임/Noun': 967,\n",
       " '황당/Noun': 968,\n",
       " '과/Noun': 969,\n",
       " '-_-/Punctuation': 970,\n",
       " '새끼/Noun': 971,\n",
       " '씩/Suffix': 972,\n",
       " '대로/Josa': 973,\n",
       " 'good/Alpha': 974,\n",
       " '학교/Noun': 975,\n",
       " '취향/Noun': 976,\n",
       " '불륜/Noun': 977,\n",
       " '흥미/Noun': 978,\n",
       " '해도/Josa': 979,\n",
       " '노력/Noun': 980,\n",
       " '액션영화/Noun': 981,\n",
       " '척/Noun': 982,\n",
       " '연기자/Noun': 983,\n",
       " '훈훈하다/Adjective': 984,\n",
       " '예고편/Noun': 985,\n",
       " '잠/Noun': 986,\n",
       " '천재/Noun': 987,\n",
       " '대작/Noun': 988,\n",
       " '잘만/Noun': 989,\n",
       " '과거/Noun': 990,\n",
       " '자꾸/Noun': 991,\n",
       " '난/Josa': 992,\n",
       " '비추다/Verb': 993,\n",
       " '개다/Verb': 994,\n",
       " '흥미진진/Noun': 995,\n",
       " '마무리/Noun': 996,\n",
       " '상상/Noun': 997,\n",
       " '속편/Noun': 998,\n",
       " '설명/Noun': 999,\n",
       " '바보/Noun': 1000,\n",
       " ...}"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "word2id = {w:i for i, w in enumerate(w2v.index2word, 1)}\n",
    "word2id"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[['아/Exclamation',\n",
       "  '더빙/Noun',\n",
       "  '../Punctuation',\n",
       "  '진짜/Noun',\n",
       "  '짜증/Noun',\n",
       "  '나다/Verb',\n",
       "  '목소리/Noun'],\n",
       " ['흠/Noun',\n",
       "  '.../Punctuation',\n",
       "  '포스터/Noun',\n",
       "  '보고/Noun',\n",
       "  '초딩/Noun',\n",
       "  '영화/Noun',\n",
       "  '줄/Noun',\n",
       "  '..../Punctuation',\n",
       "  '오버/Noun',\n",
       "  '연기/Noun',\n",
       "  '조차/Josa',\n",
       "  '가볍다/Adjective',\n",
       "  '않다/Verb'],\n",
       " ['너무/Noun',\n",
       "  '재/Noun',\n",
       "  '밓었/Noun',\n",
       "  '다/Josa',\n",
       "  '그래서/Adverb',\n",
       "  '보다/Verb',\n",
       "  '추천/Noun',\n",
       "  '하다/Verb']]"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train[:3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train = [[word2id[w] for w in seq] for seq in X_train]\n",
    "X_test = [[word2id[w] for w in seq] for seq in X_test]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[86, 532, 7, 31, 202, 127, 748],\n",
       " [1097, 12, 540, 62, 688, 2, 266, 57, 1498, 41, 1257, 777, 38],\n",
       " [21, 787, 30182, 28, 611, 5, 271, 3]]"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train[:3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "len_review = [len(x) for x in X_train]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "len_review.sort(reverse=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x136cd2630>]"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXUAAAD9CAYAAABDaefJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAG05JREFUeJzt3Xl8VfWd//HXJztZCCEJOzEEREDBLchSwWVcEGu1rWMd\ntY86jkVta/1Vf61tsb9qf4/WrbZ25mFV1JlOp7ZUOz+nFQVaF1BQgWBlEUVAQCAsYYdAEgjf3x/3\nRkNIcm9yl3POve/n45EH9557lneAvDl8z2bOOUREJDVkeB1ARETiR6UuIpJCVOoiIilEpS4ikkJU\n6iIiKUSlLiKSQlTqIiIpRKUuIpJCVOoiIikkK9kbLCsrc5WVlcnerIhIYJWVlTF37ty5zrkpkeZN\neqlXVlZSU1OT7M2KiASamZVFM5+GX0REUohKXUQkhajURURSiEpdRCSFqNRFRFKISl1EJIWo1EVE\nUkhgSv0/39rAi8tqvY4hIuJrgSn1372zkdkrt3odQ0TE1wJT6gB6RraISOcCU+pmKnURkUiCU+qY\n1xFERHwvMKUO4NCuuohIZwJT6hp+ERGJLDClLiIikQWq1LWjLiLSucCUuplp+EVEJILglLrXAURE\nAiAwpR6iXXURkc4EptRNu+oiIhEFptRBpzSKiEQSmFI30+CLiEgkwSl1HSoVEYkoMKUO4DT+IiLS\nqcCUuoZfREQiC06pex1ARCQAAlPqoLNfREQiCU6pm2n4RUQkgsCUuoZfREQiC0ypg85+ERGJJDCl\nrtsEiIhEFphSFxGRyAJT6obOfhERiSQ4pa7xFxGRiAJT6gBOJzWKiHQqMKWu4RcRkciiKnUzqzCz\nl8xsnpktNrMvhqdPMLNFZrbEzJ4ys6xEBdXoi4hIZNHuqT8E/NQ5dz5wGfCgmWUDDwJTnXNjgZXA\nLQlJGaY9dRGRzkVb6pnA5vDrg8DHwBRgjnNuV3j648DV8Y33GcM0pi4iEkG0wyXfA35pZguAC4E7\ngUuB1S0zOOeawnvvJzCzacA0gIqKiu4l1fCLiEhE0e6pXw/MIbQ3fh9wP5DPibc4b3dX2jk3wzlX\n7ZyrLi8v725WDb+IiEQQsdTNbCRQ5Zx7yjnX4JyrAWYS2ssf3mq+HOBoooIaekiGiEgk0eyp7wdG\nm1kRgJnlAtcA84ArzKwkPN9NwHOJCBnabqLWLCKSOiKOqTvntpjZz4A5ZtZE6B+Cp51z881senj6\nUWAZcHtC02pXXUSkU1EdKHXOvQC80M70ecC4OGdql2kARkQkosBcUQq6TYCISCSBKXUznf0iIhJJ\noEpdREQ6F5hSB42oi4hEEphSN0zPKBURiSA4pa7hFxGRiAJT6qDhFxGRSIJV6mp1EZFOBabU9YxS\nEZHIAlPqoOEXEZFIAlPq2k8XEYksMKUOaFBdRCSCwJS6mYZfREQiCU6pex1ARCQAAlPqoNEXEZFI\nAlPqZqZb74qIRBCcUgd27G/0OoaIiK8FptR3HGgkK0Mj6yIinQlMqQ8tL9BVpSIiEQSm1POyMznS\nfMzrGCIivhaYUs/JymDPoSavY4iI+FpgSr3xyDGONDs9KENEpBOBKfW+xXkANB7VEIyISEcCU+oF\nOZkA7D98xOMkIiL+FZhS79kjG4D6pmaPk4iI+FdwSj0vVOo79jd4nERExL8CU+ol+aFS15i6iEjH\nglPqBTkAvF+73+MkIiL+FZhSrywtAOCYTmkUEelQYEo9LzsU9fUPd3icRETEvwJT6mZGj+xMcrIC\nE1lEJOkC1ZDjqnrz1rpdHDumIRgRkfYEqtSLw+eq72/QBUgiIu2JqtTNLMPMHjCzBWY238zuDk+f\nYGaLzGyJmT1lZlmJDHvusDIA3lq3K5GbEREJrGj31L8P7HHOneucOw94zsyygQeBqc65scBK4JYE\n5QRg3JBSAD7cdiCRmxERCayIpW5mOcCXgIdapjnn1gNTgDnOuZbd5seBqxMRskVFaT49sjN55s2P\ndbdGEZF2RLOnPgRYDnzLzF4Lf10EDANWt8zknGsCsttbgZlNM7MaM6upq6uLKXBJfjb1Tc26B4yI\nSDuiKfVC4GJgm3PuQuBa4OHwsm13l9vdfXbOzXDOVTvnqsvLy2PJyzcvHAbAqx9sj2k9IiKpKJpS\nXwd84Jx7HsA5twNYCBgwvGWm8DDN0USEbG3MwF4AvLisNtGbEhEJnIil7pzbC2wzsykAZlYInAU8\nCVxhZiXhWW8CnktU0BajBxVzTmVvXvlgBzsPNiZ6cyIigRLt2S/fBq43s3nALOAe59wBYDowx8wW\nAmOAGQlJ2cawvoUAvLxiazI2JyISGJbss0iqq6tdTU1NTOtoPNrMKffMYUhZAbPvmERedmac0omI\n+JOZLXXOVUeaL1BXlLbIzcokNyuD9TvrWbh2p9dxRER8I5ClDvDSt88F4E9LN9Ose8GIiAABLvW+\nPfMAmL1yGzUbdnucRkTEHwJb6kV52Tx78zgAXv1wh64wFREhwKUOMLxvEQAz3viYlVv0mDsRkUCX\nenlRLr+45nQA3tu0x+M0IiLeC3SpA4yt7A3Aj/78PjsONHicRkTEW4Ev9cG98/n6pCEAbNx1yOM0\nIiLeCnypA1w0si8A//jE2zpgKiJpLSVKvbqyN6cPKgbQLXlFJK2lRKlnZhhfGVsBwNRfvelxGhER\n76REqQNcPro/vQty+GT3IQ3BiEjaSplSL87PZtrkKgC+88f3PE4jIuKNlCl1gCvPGADA3zft9TiJ\niIg3UqrU+xf34LpxFWzcdYiZiz/xOo6ISNKlVKkDXFM9GIAFuiWviKShlCv1Mwb3YvTAYmYt38qq\nWt0PRkTSS8qVOsDnx/QHYNH6XR4nERFJrpQs9Rs/VwnAfS+uouGILkYSkfSRkqWem5XJkLICANbu\nOOhxGhGR5EnJUgf46VWnAfBvr63xOImISPKkbKmfHH6Axtz3t+sKUxFJGylb6uVFudw9ZQQAf1lW\n63EaEZHkSNlSBxg9MHTnxofnrvY4iYhIcqR0qZ97chnXj6tg857DrN9Z73UcEZGES+lSBxjRLzS2\n/uT8dR4nERFJvJQv9a9OqKSyNJ8X/r6F+sajXscREUmolC91CN3oq/HoMWYt1wFTEUltaVHqT9xw\nNgCzlm+l+ZhObxSR1JUWpd6zRxYAb67ZyZtr6jxOIyKSOGlR6mbG87dOAGDJht0epxERSZy0KHWA\n4eErTB97fR3b9jV4nEZEJDHSptSLe2Rz18XDAViz44DHaUREEqNLpW5mk8xsS6v3E8xskZktMbOn\nzCwr/hHjZ+KwUgC++sximo4e8ziNiEj8RV3qZlYM3A5sDb/PBh4EpjrnxgIrgVsSETJezhxcwuTh\n5QDsqm/0OI2ISPx1ZU/9YWA60HIFzxRgjnOu5fFCjwNXxzFb3GVkGF8+ayAAkx58nSPN2lsXkdQS\nVamb2XXAUudc65uTDwM+vVOWc64JyO5g+WlmVmNmNXV13p5SePGovozoV8TRY44DDbrCVERSS8RS\nN7MK4DLn3JNtPwLaXsnT7pU9zrkZzrlq51x1eXl595LGSX5OFtMmVwEw+aHXNbYuIiklmj31q4Dh\nZjbPzOYBo8K/GjC8ZSYzy+GzoRlf+4eRfanonc/BxqPsPdTkdRwRkbiJWOrOuX91zo1zzp3vnDsf\nWBX+9THgCjMrCc96E/BcwpLGUXGPbO66JPTv0WW/elMPpxaRlNHt89Sdcw2EDpzOMbOFwBhgRryC\nJdrkk8vp1zOPXfVNuhhJRFJGl0vdOTe+1et54b34zznnvuGcC8wub0lBDj+58lQAvvHsu9TuPexx\nIhGR2KXNFaXtqa7szUUj+7Bq637er93vdRwRkZildan3Lshh+uWjAHjkr6t595M9HicSEYlNWpc6\nwMBePZg6uh+rtx/g1Q+2ex1HRCQmaV/qOVkZ/Pr6s+mZl83LK7bx8oqtXkcSEem2tC/1FpeP6c+2\nfQ38x8L1XkcREek2lXrYz744mnNPLmNdXT2PvvIRx/TYOxEJIJV6K+OG9KbxSDOPvrKGDbvqvY4j\nItJlKvVWbp5UxaPXngnAMwvWs67uoMeJRES6RqXeRlV5Afk5mTy76BOeWaDxdREJFpV6G0PLC1n1\nkylUlRWwcss+neYoIoGiUu9AVXkByzfv4+bf1lDfGIibT4qIqNQ78sQNZ3PP5SNxDv78Xi11B/T4\nOxHxP5V6B7IyMzi5bxEAP3xhBT+fuzrCEiIi3lOpd+K84eW88d0LqCovYF3dQZZu3E2zzl8XER9T\nqUdQUZrPoJJ8ajbu4cuPv8281Tu8jiQi0iGVehQe/coZPHHDWQCs2LJP4+si4lsq9Sj0Lshh8vBy\nMjOMR19Zw/kPv86RZj2wWkT8R6UepfycLP506wSuHTuY+qZmNuys17NNRcR3VOpdcGZFCeOrSgG4\n+JdvcNEv5nucSETkeCr1Lrr01H48dPUYLhrZl817DtNwpFl3dBQR31Cpd1GPnEyuqR7M54aF9thH\n/GgO5/zsVQ43aShGRLyX5XWAoLryjIE0HDnG8s17mb1yG3UHGqkozfc6loikOe2pd1PvghxuO38o\nV505EICp//om5/z0FVZvO+BxMhFJZyr1GE0YWspt5w/l0lP7seNAIx9u2+91JBFJY+Zccg/yVVdX\nu5qamqRuMxl2HGjgnJ++SllhDkV52Vw/roKbJ1V5HUtEUoSZLXXOVUeaT3vqcVJemMst51UxcWgZ\n+w8f4bUPdTsBEUk+HSiNEzPjB5eNBOBr/76Ymg27+dKvF5KZYXxvygjGVvb2OKGIpAPtqSfA1WcP\n4qyTSijIzaJm4x7mr67zOpKIpAntqSfAFacP4IrTBwAw5t65/GVZLR+Gz4o5b3gZX51Q6WE6EUll\n2lNPsC+dNYjC3Cxq9x5myYbdPK2HWYtIAmlPPcHu/cKpn77+4Qsr+H/vbubuPy3/dNpVZw5kwtBS\nL6KJSApSqSfR+KpS5n24g/kfhcbYdx5sZO/hJpW6iMRNxPPUzWwi8AOgEMgF7nPOzTWzCcCjhIZw\n3gNuc84djbTBVD1PvTu+/PhbbN/fwOfHDPh0WlaGccP4k+hXnOdhMhHxm2jPU49mTz0TuM45d8DM\nSoD5ZvYa8CAw1Tm3y8zuAG4BHospdZo5q6IX//n2Rv59YXic3UFT8zEK87K49byh3oYTkUCKWOrO\nuTdbvd0LHAamAHOcc7vC0x8H5qJS75Lpl49i+uWjPn3vnGPY9NksWLOTTLPj5u3TM5crzxiY7Igi\nEjBRj6mbWQbwMPAMMAxY3fKZc67JzLI7WXYaMA2goqKi22FTnZlxcp9CFqzdyYK1O0/4fHxVKX17\nalhGRDoWVambWR/gV8BM59yfzexOoO1gfIeD8865GcAMCI2pdzNrWph1+7k0HD3++adzV27jrueX\n8d/vbqZv0Yml3r9XHhOHliUrooj4WMRSN7Mq4N+AW5xzm8OTPwZGtJonB4h4kFQiy8rMoDDz+MsH\nhvYpBOChOavbW4QMgxX3XkpBrk5mEkl30bTAvcBNzrntrabNAb5rZk865/YANwHPJSCfAGcM7sXi\n6f9AQ9OxEz57cXktD89dzdz3t1GSn9Pu8v175TGiX89ExxQRH4im1CcCf7TjD9zdDEwH5pjZUWAZ\ncHv840mLPu0MuwCM7F8EwJ3PLetw2dysDFbedynZmbqAWCTVRXP2y7AOPloLjItvHOmqC07pw+w7\nJtF49MS9eIBZy2p5esF6Fq/fTWEnwzN9eubSv7hHomKKSJJoEDbgzIyR/TseWtm4qx6A659e1Ol6\nCnOzWP7jS8jIsE7nExF/U6mnuKmj+1NakEtTc3OH8/xt1Xb+sHgTa3YcJD8ns9P1lRflkpfd+Twi\n4h2VeorLzszg3JM7P91xT/0R/rB4E5c++kbE9U0cWsrvvz4+XvFEJM5U6sLU0f3JyjSaOhiXbzFz\nySY+2X2I+sbozl7Nz8mkzQF2EUkwlbrQIyczqlsQrNyyj6Ub93Dqj+dGtd5/Omcw939pTKzxRKQL\nVOoStZsnVTGoJB/X8cXDn5q5eBMfbT+YhFQi0ppKXaI2uHc+X59cFdW8727cy5z3t3HKPbOjXv//\nvuSUqNcvIu1TqUtCfOOCoZxUlh/1/M8t2cR7m/YmMJFIelCpS0KMGdSLMYN6RT3/Ox/v5pUPtjPx\n/lejXqZXfg5/vGU8RXkd3iBUJO2o1MUXbjtvKK9+sD3yjGG1+w6zcO0uPtl9iFMHFCcwmUiwqNTF\nF6ac1o8pp/WLev6Fa3eycO0uvvX7v0e8YKq1DDPuumQ455/SpzsxRXxPpS6BNHpQMVeeMSDqc+Zb\nzP+ojnmr61TqkrJU6hJIPfOy+dW1Z3Z5uYn3v8qcldtYV9f10y1H9e/JD6aO7PJyIsmkUpe0cnX1\nYN5cU8fBLu7hb9vXwIK1O/n+ZSN0laz4mjmX3KfLVVdXu5qamqRuUyRWT8xfxwOzP+SK0weQ2Y1O\nNzO+NrGSMwZHf0aQSGtmttQ5Vx1pPu2pi0RhbGUJQ8sLWL65e+fSb9p9iMLcLJW6JJz21EWS4NwH\nX6NHdiaTTi7v9jryczK57fyhehZtmtKeuoiPjK8qZe7KbTxfs6lbyzc7x6GmZs4+qYQLRujMHemY\n9tRFAmDtjgNc9Is3+PyY/ozoVxTTuvJzsrhh/EnkZOmZtUGiPXWRFNK3Zx4l+dnMWr6VWcu3xry+\nkf17MmFoaRySid+o1EUCoCgvm6X3XExzjP+zXlW7nysfW8jLK7ayfmd9zLn6Fedy4Yi+Ma9H4kel\nLhIQGRlGBrGdIz+opAc5WRn81zsb45QKlv2fSyjO103V/EKlLpJGSgtzWXrPRRxq6vhB5NF6aflW\nfjJrFbNW1FJakBuHdDCsTyHD+hTGZV3pSqUukmaK8rLjcrvilgO2019YGfO6WpxUms/8714Qt/Wl\nI5W6iHTLhKGlvHbXeTQc6fyB5dGa8cY6/rpqO4vX745pPXnZGZw2oJiMjPS8nYNKXUS6xcyoKo/f\nUMnwfkX8z3u1XPPk2zGv6zf/PDZt78SpUhcRX7jpc0M4c3AJx2I4w2fnwUbumPke79fuZ1BJ9I9T\nBKgszScrM/jn7qvURcQX8rIzYz53/kDDETIMHp67mofnru7Ssv9y7hB+9PlRMW3fD1TqIpIyivKy\nmTltAtv2N3RpuQde/oCNu+rZU9/U6XxmoWfj+plKXURSyjlDend5md+9vZFXPtjBmf/3bxHnnT51\nJF+fXNWdaEmhUheRtPfjL4xiSRRn3Tzy14/4eGfXn5qVTCp1EUl7pw4o5tQBxRHn++07G/nD4k38\nYXHHd9u8duxgHvjymHjG6xKVuohIlO694lRqNu7p8PNZy2tZsWVfEhOdKOZSN7O+wG+A3sB+4GvO\nudpY1ysi4jeTh5czeXjHDzr5ZFc9f1lWy9ntjM1XlhXw/C0TEn5RVDz21B8B7nXOLTKzauDnwHVx\nWK+ISKDc+LkhFOadWKsrtuxn6cY9HDl2jNyMzIRmiKnUzawXUOqcWwTgnKsxs2Iz6+Wc697DHEVE\nAuqMwb3afQ7tY6+vZdmmvSTjmUSxXj41BFjbZtrH4emfMrNpZlZjZjV1dXUxblJEJFiGlhcydXQ/\nMizx96OJdfjFgPb+7TlumnNuBjADQo+zi3GbIiKBMuW0fkw5rV9SthXrnvoGYFibaVXh6SIikmQx\nlbpzbjdwyMzOAjCzMcAujaeLiHgjHme//C/gGTMrBPYB/xyHdYqISDfEXOrOuc3ApXHIIiIiMQr+\nzYNFRORTKnURkRSiUhcRSSEqdRGRFGIuGdettt6gWR2wsZuLlwE74xgnEfye0e/5wP8Z/Z4P/J/R\n7/nAXxl3AjjnpkSaMemlHgszq3HOVXudozN+z+j3fOD/jH7PB/7P6Pd8EIyM7dHwi4hIClGpi4ik\nkKCV+gyvA0TB7xn9ng/8n9Hv+cD/Gf2eD4KR8QSBGlMXEZHOBW1PXUREOqFSFxFJIYEodTPra2az\nzWyRmf3NzAYkcFsTzexFM3vdzN4ys0vD0yeEt7/EzJ4ys6xWy0wzs6Vm9q6ZfbPV9Ewz+3V4mSVm\nNime35OZTTKzLa3e+yajmWWY2QNmtsDM5pvZ3X7KaGYVZvaSmc0zs8Vm9kW/5DOzH5jZra3eJy1T\nZ9uKkPFGM3sl/Pv5VzOr8lvGVtOnm9nMNtM8yZgQzjnffwG/A8aFX1cDv0/gtiYBReHXJcByIBt4\ng9DzWAHuAL4Zfj0KeIHQU6AMeB44LfzZbcCd4de9gflAdjy+J6AYeA6oCb/3VUbgh8Ddrd4P8VNG\nYCYwMfy6FPjI63zAQOAdoA64Ndl/rp1tq7OM4emXApnh1xOB5/yWMfzZWOCPwMxW05KeMZFfnhd2\nFD98vYDZbaa9BPRKwrYNWARcAfyw1fQc4PXw60cIl0P4/XjgkfDreUBuq8++H15XzN8ToSPzJwPv\nhN/7JmN42zWED8S3mu6njM8DFeHXucAcv+QDbuSzUk9aps621VnGdj4bCPyP3zICBcB/A6dwfKl7\nljERX0EYfonq4dbxZmYZwM+BZwg9sm91y2fOuSZC/xrT9jNgDZ894i/LOdfYzmcxfU9mdh2w1Dm3\nptVkP2UcQuh/ON8ys9fCXxf5LOP3gF+a2XeAPwF3+ixfi2Rm6mxbUTGzYuBJ4Cc+zPgAcB/Q2Ga6\nnzLGLAilHtXDreO6QbM+wLPAGy700Oz2MrS87+yz9rgOlom0XEu2CuAy59yTbT/yS0agELgY2Oac\nuxC4FniY0N83v2S8ntDe+eOEftDvB/J9lK9FMv9cu7q+45jZ2cBvgO855971U0YzuxLY4Jxb3t7H\nfsgYL0Eo9Q0k8eHW4QM8/wF81zn35/Dkj4HhrebJAY6291n4dcu/4MfC87b9bAPd/56uAoaHD0jN\nA0aFfzUfZVwHfOCcex7AObcDWOiXjGY2Eqhyzj3lnGtwztUQGmPP8kO+NpL5d6+zbXUqXJq3Af/k\nnFvVUX4PM34F+EL4Z2UmcGH4ZyjTRxnjI1njPLF8Efrv8Vnh12OA3yZwW78F+raZlkeolErC728F\nbgu/PovQ+GzLQZZngVHhz74D3B5+3RN4lc8OssTle+KzMXVfZQz/Pk4Jvy4E3gKK/JCR0JjvEj47\nIJ5L6EDZeT7JdyOfjakn7c+1s21FyJgN/AXIaGc+X2RsM72S48fUPcuYiC/PirpLIWEQMDf8G/Uy\nbUo3zttaS+jgSOuvYcD5hA6aLgR+TfhIf3iZbxMqiXeAaa2mZwNPh6cvBCbE+3siXOrh177JSOiA\n0X+1+j280E8ZgS+Gl3md0BkNX/VLPk48wJe0TJ1tq6OMhM4e+YTjf2Zm+Sljm+mVtCp1LzMm4ku3\nCRARSSFBGFMXEZEoqdRFRFKISl1EJIWo1EVEUohKXUQkhajURURSiEpdRCSFqNRFRFLI/wccMULN\nYqTHiwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x137747470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(len_review) # 시퀀스의 최대길이는 60정도로 하면 될듯."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "MAX_SEQUENCE_LENGTH = 60\n",
    "X_train = pad_sequences(X_train, maxlen=MAX_SEQUENCE_LENGTH) # MAX_SEQUENCE_LENGTH만큼 잘림.\n",
    "X_test = pad_sequences(X_test, maxlen=MAX_SEQUENCE_LENGTH) # MAX_SEQUENCE_LENGTH만큼 잘림."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[    0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,    86,\n",
       "          532,     7,    31,   202,   127,   748],\n",
       "       [    0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,  1097,    12,   540,    62,   688,     2,   266,\n",
       "           57,  1498,    41,  1257,   777,    38],\n",
       "       [    0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,    21,   787,\n",
       "        30182,    28,   611,     5,   271,     3]], dtype=int32)"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train[:3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. CNN Modeling"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Embedding matrix 생성"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "EMBEDDING_DIM = 300"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.00700791,  0.35573502,  0.51139972, ...,  0.18719359,\n",
       "         0.42934556,  0.68204269],\n",
       "       [ 0.42230761,  0.9181581 ,  0.26944045, ...,  0.16361779,\n",
       "         0.38475476,  0.13843139],\n",
       "       [ 0.71813344,  0.27722144,  0.79266566, ...,  0.7891304 ,\n",
       "         0.6024807 ,  0.1299213 ],\n",
       "       ..., \n",
       "       [ 0.69003403,  0.17318001,  0.79312408, ...,  0.8637267 ,\n",
       "         0.79911267,  0.90928977],\n",
       "       [ 0.07474478,  0.15838595,  0.88857159, ...,  0.70022907,\n",
       "         0.4349205 ,  0.51810644],\n",
       "       [ 0.64345942,  0.4018112 ,  0.29217788, ...,  0.77330443,\n",
       "         0.05750117,  0.01045013]])"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "embedding_matrix = np.random.random((len(word2id)+1, EMBEDDING_DIM))\n",
    "embedding_matrix # 랜덤값으로 초기화"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(55825, 300)"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "embedding_matrix.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "for k in word2id:\n",
    "    embedding_matrix[word2id[k]] = w2v[k]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "### Modeling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_2 (Embedding)      (None, 60, 300)           16747500  \n",
      "_________________________________________________________________\n",
      "conv1d_4 (Conv1D)            (None, 60, 300)           270300    \n",
      "_________________________________________________________________\n",
      "max_pooling1d_4 (MaxPooling1 (None, 12, 300)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_5 (Conv1D)            (None, 12, 300)           270300    \n",
      "_________________________________________________________________\n",
      "max_pooling1d_5 (MaxPooling1 (None, 4, 300)            0         \n",
      "_________________________________________________________________\n",
      "conv1d_6 (Conv1D)            (None, 4, 300)            270300    \n",
      "_________________________________________________________________\n",
      "max_pooling1d_6 (MaxPooling1 (None, 2, 300)            0         \n",
      "_________________________________________________________________\n",
      "dropout_5 (Dropout)          (None, 2, 300)            0         \n",
      "_________________________________________________________________\n",
      "flatten_2 (Flatten)          (None, 600)               0         \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 300)               180300    \n",
      "_________________________________________________________________\n",
      "dropout_6 (Dropout)          (None, 300)               0         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 10)                3010      \n",
      "_________________________________________________________________\n",
      "dense_6 (Dense)              (None, 1)                 11        \n",
      "=================================================================\n",
      "Total params: 17,741,721.0\n",
      "Trainable params: 17,741,721.0\n",
      "Non-trainable params: 0.0\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "# create the model\n",
    "DIC_SIZE = len(word2id)+1\n",
    "model = Sequential()\n",
    "model.add(Embedding(DIC_SIZE, EMBEDDING_DIM, weights=[embedding_matrix], input_length=MAX_SEQUENCE_LENGTH))\n",
    "model.add(Conv1D(filters=EMBEDDING_DIM, kernel_size=3, padding='same', activation='relu'))\n",
    "model.add(MaxPooling1D(pool_size=5))\n",
    "model.add(Conv1D(filters=EMBEDDING_DIM, kernel_size=3, padding='same', activation='relu'))\n",
    "model.add(MaxPooling1D(pool_size=3))\n",
    "model.add(Conv1D(filters=EMBEDDING_DIM, kernel_size=3, padding='same', activation='relu'))\n",
    "model.add(MaxPooling1D(pool_size=2))\n",
    "model.add(Dropout(0.15))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(300, activation='relu'))\n",
    "model.add(Dropout(0.15))\n",
    "model.add(Dense(10, activation='relu'))\n",
    "model.add(Dense(1, activation='sigmoid'))\n",
    "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "print(model.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 150000 samples, validate on 50000 samples\n",
      "Epoch 1/5\n",
      "150000/150000 [==============================] - 280s - loss: 0.6103 - acc: 0.6712 - val_loss: 0.5272 - val_acc: 0.8213\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 2/5\n",
      "150000/150000 [==============================] - 281s - loss: 0.5028 - acc: 0.8473 - val_loss: 0.5004 - val_acc: 0.8476\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 3/5\n",
      "150000/150000 [==============================] - 281s - loss: 0.4712 - acc: 0.8682 - val_loss: 0.4904 - val_acc: 0.8439\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 4/5\n",
      "150000/150000 [==============================] - 286s - loss: 0.4483 - acc: 0.8809 - val_loss: 0.4834 - val_acc: 0.8544\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 5/5\n",
      "150000/150000 [==============================] - 282s - loss: 0.4239 - acc: 0.8941 - val_loss: 0.4795 - val_acc: 0.8523\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x11ca8b400>"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(X_train, y_train, validation_data=(X_test, y_test), epochs=5, batch_size=4096)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50000/50000 [==============================] - 26s     \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n"
     ]
    }
   ],
   "source": [
    "preds = model.predict_classes(X_test, batch_size=4096).reshape(-1)\n",
    "actuals = np.array(y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.85226000000000002"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Accuracy\n",
    "np.sum(np.array(actuals) == np.array(preds)) / float(len(actuals))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}