Preprocessing_on_test.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Determining Best Combination of Preprocessing Techniques on Test Set"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Loading Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/HeddaVik/anaconda/lib/python3.6/importlib/_bootstrap.py:205: RuntimeWarning: compiletime version 3.5 of module 'tensorflow.python.framework.fast_tensor_util' does not match runtime version 3.6\n",
      "  return f(*args, **kwds)\n",
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "\n",
    "# internal imports\n",
    "import helpers as HL\n",
    "import glove_module as GV\n",
    "import neural_nets as NN\n",
    "import validation_and_prediction as VP\n",
    "import tokenizing as TO"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create the word embeddings using the created gensim-.txt file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Pick one, the higher dimension, the better result and longer computational time. \n",
    "global_vectors=HL.get_global_vectors(200)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Creating corpus:\n",
    "In addition to the acutal corpus, some additional information is needed "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "full_corpus, nr_pos_tweets, nr_neg_tweets, total_training_tweets=HL.get_corpus(test=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Picking the neural net\n",
    "At this stage, we want to use the simple neural net"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "neural_nets=[NN.basic_model_adam]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Initializing variables to apply all preprocessing techniques:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Initializing vectors:\n",
    "corpuses=[]\n",
    "corpuses.append(full_corpus)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Defining names of corpuses: \n",
    "names=['original_corpus','SH_corpus','SHM_corpus','H_corpus','HK_corpus','PS_corpus','NS__corpus','OS_corpus','N_corpus','NM_corpus','ST_corpus','SP_corpus','E_corpus','SN_corpus','RS_corpus','EX_corpus','N-2_corpus','N-3_corpus','N-4_corpus']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Defining inputs to preprocessing function: \n",
    "inputs=[{'hashtag': True, 'segmentation_hash': True},\n",
    "        {'hashtag':True,'segmentation_hash': True,'hashtag_mention':True},\n",
    "        {'hearts':True},\n",
    "        {'hugs_and_kisses':True},\n",
    "        {'pos_smilies':True},\n",
    "        {'neg_smilies':True},\n",
    "        {'other_smilies':True},\n",
    "        {'numbers':True},\n",
    "        {'numbers':True,'number_mention':True},\n",
    "        {'stemming':True},\n",
    "        {'spelling':True},#Warning: When True, it takes app  2.5 h on test set. Recomended to always set to false \n",
    "        {'elongation':True},\n",
    "        {'set_to_not':True},\n",
    "        {'remove_signs':True},\n",
    "        {'exclamation':True}]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Applying all preprocessing techniques to the original corpus: "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reading twitter - 1grams ...\n",
      "Reading twitter - 2grams ...\n",
      "Time in min before starting first for loop: 0.1436559518178304\n",
      "Time in min after 1  tweets: 0.14389031728108723\n",
      "Time in min after 25001  tweets: 0.30348013242085775\n",
      "Time in min after 50001  tweets: 0.43382955392201744\n",
      "Time in min after 75001  tweets: 0.558660888671875\n",
      "Time in min after 100001  tweets: 0.688304070631663\n",
      "Time in min after 125001  tweets: 0.7780535181363424\n",
      "Time in min after 150001  tweets: 0.8791473507881165\n",
      "Time in min after 175001  tweets: 0.9658855517705282\n",
      "Time in min after 200001  tweets: 1.0498501340548196\n",
      "Time in min total: 1.088587518533071\n",
      "Reading twitter - 1grams ...\n",
      "Reading twitter - 2grams ...\n",
      "Time in min before starting first for loop: 0.14544553756713868\n",
      "Time in min after 1  tweets: 0.14553653399149577\n",
      "Time in min after 25001  tweets: 0.36271895170211793\n",
      "Time in min after 50001  tweets: 0.48187756538391113\n",
      "Time in min after 75001  tweets: 0.6029905160268147\n",
      "Time in min after 100001  tweets: 0.7281522313753764\n",
      "Time in min after 125001  tweets: 0.824855101108551\n",
      "Time in min after 150001  tweets: 0.9295561194419861\n",
      "Time in min after 175001  tweets: 1.0864377816518147\n",
      "Time in min after 200001  tweets: 1.179166603088379\n",
      "Time in min total: 1.2214224020640054\n",
      "Time in min before starting first for loop: 1.986821492513021e-07\n",
      "Time in min after 1  tweets: 1.1066595713297525e-05\n",
      "Time in min after 25001  tweets: 0.036681199073791505\n",
      "Time in min after 50001  tweets: 0.08183350165685017\n",
      "Time in min after 75001  tweets: 0.12411334911982218\n",
      "Time in min after 100001  tweets: 0.16018033425013226\n",
      "Time in min after 125001  tweets: 0.2153873324394226\n",
      "Time in min after 150001  tweets: 0.2625072479248047\n",
      "Time in min after 175001  tweets: 0.30550899902979534\n",
      "Time in min after 200001  tweets: 0.34981273412704467\n",
      "Time in min total: 0.3665616512298584\n",
      "Time in min before starting first for loop: 2.3444493611653645e-07\n",
      "Time in min after 1  tweets: 9.985764821370443e-06\n",
      "Time in min after 25001  tweets: 0.039212135473887126\n",
      "Time in min after 50001  tweets: 0.0747536857922872\n",
      "Time in min after 75001  tweets: 0.10981480280558269\n",
      "Time in min after 100001  tweets: 0.1449432651201884\n",
      "Time in min after 125001  tweets: 0.1890914003054301\n",
      "Time in min after 150001  tweets: 0.23168223698933918\n",
      "Time in min after 175001  tweets: 0.2735080679257711\n",
      "Time in min after 200001  tweets: 0.3156483848889669\n",
      "Time in min total: 0.33231285015741985\n",
      "Time in min before starting first for loop: 5.1657358805338543e-08\n",
      "Time in min after 1  tweets: 1.0482470194498698e-05\n",
      "Time in min after 25001  tweets: 0.03705546458562215\n",
      "Time in min after 50001  tweets: 0.07720609903335571\n",
      "Time in min after 75001  tweets: 0.11368538538614908\n",
      "Time in min after 100001  tweets: 0.15132960081100463\n",
      "Time in min after 125001  tweets: 0.19682021538416544\n",
      "Time in min after 150001  tweets: 0.2404235323270162\n",
      "Time in min after 175001  tweets: 0.283683717250824\n",
      "Time in min after 200001  tweets: 0.3272354006767273\n",
      "Time in min total: 0.34498658577601116\n",
      "Time in min before starting first for loop: 9.934107462565105e-08\n",
      "Time in min after 1  tweets: 1.4917055765787761e-05\n",
      "Time in min after 25001  tweets: 0.03606611887613932\n",
      "Time in min after 50001  tweets: 0.07045566240946452\n",
      "Time in min after 75001  tweets: 0.10504428148269654\n",
      "Time in min after 100001  tweets: 0.13978913227717082\n",
      "Time in min after 125001  tweets: 0.18047536611557008\n",
      "Time in min after 150001  tweets: 0.22287368377049763\n",
      "Time in min after 175001  tweets: 0.26526919603347776\n",
      "Time in min after 200001  tweets: 0.3065086285273234\n",
      "Time in min total: 0.3234136501948039\n",
      "Time in min before starting first for loop: 5.1657358805338543e-08\n",
      "Time in min after 1  tweets: 1.2469291687011718e-05\n",
      "Time in min after 25001  tweets: 0.035004301865895586\n",
      "Time in min after 50001  tweets: 0.06929828723271687\n",
      "Time in min after 75001  tweets: 0.10434546868006388\n",
      "Time in min after 100001  tweets: 0.13944608370463055\n",
      "Time in min after 125001  tweets: 0.1821918527285258\n",
      "Time in min after 150001  tweets: 0.22387633323669434\n",
      "Time in min after 175001  tweets: 0.2664627552032471\n",
      "Time in min after 200001  tweets: 0.3079224189122518\n",
      "Time in min total: 0.32483073870340984\n",
      "Time in min before starting first for loop: 8.344650268554687e-08\n",
      "Time in min after 1  tweets: 6.735324859619141e-06\n",
      "Time in min after 25001  tweets: 0.03259623448053996\n",
      "Time in min after 50001  tweets: 0.06572908560434977\n",
      "Time in min after 75001  tweets: 0.09914958079655965\n",
      "Time in min after 100001  tweets: 0.1320775310198466\n",
      "Time in min after 125001  tweets: 0.17361761728922526\n",
      "Time in min after 150001  tweets: 0.21472488244374593\n",
      "Time in min after 175001  tweets: 0.25475576321283977\n",
      "Time in min after 200001  tweets: 0.29412871996561685\n",
      "Time in min total: 0.3098848978678385\n",
      "Time in min before starting first for loop: 2.3444493611653645e-07\n",
      "Time in min after 1  tweets: 1.441637674967448e-05\n",
      "Time in min after 25001  tweets: 0.03439921538035075\n",
      "Time in min after 50001  tweets: 0.06756931940714518\n",
      "Time in min after 75001  tweets: 0.10130543311436971\n",
      "Time in min after 100001  tweets: 0.1343284209569295\n",
      "Time in min after 125001  tweets: 0.17505690256754558\n",
      "Time in min after 150001  tweets: 0.21505904992421468\n",
      "Time in min after 175001  tweets: 0.25519610246022545\n",
      "Time in min after 200001  tweets: 0.2957972208658854\n",
      "Time in min total: 0.3111751159032186\n",
      "Time in min before starting first for loop: 3.635883331298828e-06\n",
      "Time in min after 1  tweets: 8.583863576253256e-05\n",
      "Time in min after 25001  tweets: 0.16376915375391643\n",
      "Time in min after 50001  tweets: 0.33275983333587644\n",
      "Time in min after 75001  tweets: 0.49913777112960817\n",
      "Time in min after 100001  tweets: 0.6666138569513956\n",
      "Time in min after 125001  tweets: 0.8682217359542846\n",
      "Time in min after 150001  tweets: 1.070938221613566\n",
      "Time in min after 175001  tweets: 1.2717960556348165\n",
      "Time in min after 200001  tweets: 1.4712692181269327\n",
      "Time in min total: 1.55093963543574\n",
      "Reading english - 1grams ...\n",
      "Time in min before starting first for loop: 0.01388164758682251\n",
      "Time in min after 1  tweets: 0.017032114664713542\n",
      "Time in min after 25001  tweets: 20.19349333445231\n",
      "Time in min after 50001  tweets: 40.48593374490738\n",
      "Time in min after 75001  tweets: 59.45620791514715\n",
      "Time in min after 100001  tweets: 78.07600614627202\n",
      "Time in min after 125001  tweets: 99.36792829831441\n",
      "Time in min after 150001  tweets: 119.6522124171257\n",
      "Time in min after 175001  tweets: 138.66170159578323\n",
      "Time in min after 200001  tweets: 157.25536909898122\n",
      "Time in min total: 164.43695551554362\n",
      "Time in min before starting first for loop: 0.0010034521420796713\n",
      "Time in min after 1  tweets: 0.0010468721389770507\n",
      "Time in min after 25001  tweets: 0.14110540548960368\n",
      "Time in min after 50001  tweets: 0.27995618581771853\n",
      "Time in min after 75001  tweets: 0.42334138552347816\n",
      "Time in min after 100001  tweets: 0.5639538526535034\n",
      "Time in min after 125001  tweets: 0.7303165038426717\n",
      "Time in min after 150001  tweets: 0.8945788502693176\n",
      "Time in min after 175001  tweets: 1.0580472389856974\n",
      "Time in min after 200001  tweets: 1.2217705845832825\n",
      "Time in min total: 1.2869434197743734\n",
      "Time in min before starting first for loop: 9.934107462565105e-08\n",
      "Time in min after 1  tweets: 7.80026117960612e-06\n",
      "Time in min after 25001  tweets: 0.03499529759089152\n",
      "Time in min after 50001  tweets: 0.0696573813756307\n",
      "Time in min after 75001  tweets: 0.1042669177055359\n",
      "Time in min after 100001  tweets: 0.13850518465042114\n",
      "Time in min after 125001  tweets: 0.18059913317362467\n",
      "Time in min after 150001  tweets: 0.22157548268636068\n",
      "Time in min after 175001  tweets: 0.2637664993604024\n",
      "Time in min after 200001  tweets: 0.30549081563949587\n",
      "Time in min total: 0.3221431533495585\n",
      "Time in min before starting first for loop: 1.1523564656575521e-07\n",
      "Time in min after 1  tweets: 1.1797746022542318e-05\n",
      "Time in min after 25001  tweets: 0.03666576941808065\n",
      "Time in min after 50001  tweets: 0.07264928023020427\n",
      "Time in min after 75001  tweets: 0.10947771469751993\n",
      "Time in min after 100001  tweets: 0.14807653427124023\n",
      "Time in min after 125001  tweets: 0.19276991685231526\n",
      "Time in min after 150001  tweets: 0.23616008361180624\n",
      "Time in min after 175001  tweets: 0.27958924770355226\n",
      "Time in min after 200001  tweets: 0.3219693342844645\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time in min total: 0.3391306519508362\n",
      "Time in min before starting first for loop: 1.0331471761067709e-07\n",
      "Time in min after 1  tweets: 2.296765645345052e-05\n",
      "Time in min after 25001  tweets: 0.03366755247116089\n",
      "Time in min after 50001  tweets: 0.06769651571909586\n",
      "Time in min after 75001  tweets: 0.10239670276641846\n",
      "Time in min after 100001  tweets: 0.1364606499671936\n",
      "Time in min after 125001  tweets: 0.1760465661684672\n",
      "Time in min after 150001  tweets: 0.21609342098236084\n",
      "Time in min after 175001  tweets: 0.2563955863316854\n",
      "Time in min after 200001  tweets: 0.2959631681442261\n",
      "Time in min total: 0.3117439309755961\n"
     ]
    }
   ],
   "source": [
    "\n",
    "for input_ in inputs: \n",
    "        corpus=TO.preprocess_corpus(full_corpus, **input_)\n",
    "        corpuses.append(corpus)\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "ns=[2,3,4]\n",
    "for n in ns: \n",
    "    corpus=TO.creating_n_grams_corpus(n,full_corpus)\n",
    "    corpuses.append(corpus)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Testing all preprocessing techniques: "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4701 - acc: 0.7652Epoch 00001: val_loss improved from inf to 0.40993, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 4s 30us/step - loss: 0.4699 - acc: 0.7653 - val_loss: 0.4099 - val_acc: 0.8052\n",
      "Epoch 2/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.4091 - acc: 0.8073Epoch 00002: val_loss improved from 0.40993 to 0.39679, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.4090 - acc: 0.8072 - val_loss: 0.3968 - val_acc: 0.8118\n",
      "Epoch 3/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3957 - acc: 0.8156Epoch 00003: val_loss improved from 0.39679 to 0.39259, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3957 - acc: 0.8156 - val_loss: 0.3926 - val_acc: 0.8148\n",
      "Epoch 4/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3877 - acc: 0.8201Epoch 00004: val_loss improved from 0.39259 to 0.38873, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3876 - acc: 0.8201 - val_loss: 0.3887 - val_acc: 0.8187\n",
      "Epoch 5/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3804 - acc: 0.8243Epoch 00005: val_loss improved from 0.38873 to 0.38618, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3803 - acc: 0.8241 - val_loss: 0.3862 - val_acc: 0.8182\n",
      "Epoch 6/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3746 - acc: 0.8267Epoch 00006: val_loss improved from 0.38618 to 0.38460, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3745 - acc: 0.8266 - val_loss: 0.3846 - val_acc: 0.8197\n",
      "Epoch 7/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3691 - acc: 0.8296Epoch 00007: val_loss improved from 0.38460 to 0.38364, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3690 - acc: 0.8297 - val_loss: 0.3836 - val_acc: 0.8189\n",
      "Epoch 8/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3642 - acc: 0.8328Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3645 - acc: 0.8326 - val_loss: 0.3842 - val_acc: 0.8195\n",
      "Epoch 9/100\n",
      "128000/133332 [===========================>..] - ETA: 0s - loss: 0.3603 - acc: 0.8345Epoch 00009: val_loss improved from 0.38364 to 0.38302, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3604 - acc: 0.8344 - val_loss: 0.3830 - val_acc: 0.8199\n",
      "Epoch 10/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3566 - acc: 0.8370Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3568 - acc: 0.8369 - val_loss: 0.3878 - val_acc: 0.8199\n",
      "Epoch 11/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3528 - acc: 0.8397Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3534 - acc: 0.8393 - val_loss: 0.3854 - val_acc: 0.8198\n",
      "Epoch 12/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3496 - acc: 0.8410Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3496 - acc: 0.8411 - val_loss: 0.3865 - val_acc: 0.8203\n",
      "Epoch 00012: early stopping\n",
      "66668/66668 [==============================] - 2s 30us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.4656 - acc: 0.7664Epoch 00001: val_loss improved from inf to 0.42971, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 3s 19us/step - loss: 0.4645 - acc: 0.7672 - val_loss: 0.4297 - val_acc: 0.7939\n",
      "Epoch 2/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4028 - acc: 0.8105Epoch 00002: val_loss improved from 0.42971 to 0.41527, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.4027 - acc: 0.8105 - val_loss: 0.4153 - val_acc: 0.8035\n",
      "Epoch 3/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3901 - acc: 0.8166Epoch 00003: val_loss improved from 0.41527 to 0.41077, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 11us/step - loss: 0.3901 - acc: 0.8165 - val_loss: 0.4108 - val_acc: 0.8040\n",
      "Epoch 4/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3818 - acc: 0.8215Epoch 00004: val_loss improved from 0.41077 to 0.40619, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 1s 11us/step - loss: 0.3817 - acc: 0.8214 - val_loss: 0.4062 - val_acc: 0.8084\n",
      "Epoch 5/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3750 - acc: 0.8266Epoch 00005: val_loss improved from 0.40619 to 0.40319, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3749 - acc: 0.8266 - val_loss: 0.4032 - val_acc: 0.8081\n",
      "Epoch 6/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3685 - acc: 0.8297Epoch 00006: val_loss improved from 0.40319 to 0.40289, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3690 - acc: 0.8293 - val_loss: 0.4029 - val_acc: 0.8082\n",
      "Epoch 7/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3638 - acc: 0.8333Epoch 00007: val_loss improved from 0.40289 to 0.40077, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3639 - acc: 0.8331 - val_loss: 0.4008 - val_acc: 0.8061\n",
      "Epoch 8/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3591 - acc: 0.8351Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 1s 11us/step - loss: 0.3589 - acc: 0.8353 - val_loss: 0.4015 - val_acc: 0.8065\n",
      "Epoch 9/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3551 - acc: 0.8371Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3547 - acc: 0.8374 - val_loss: 0.4017 - val_acc: 0.8135\n",
      "Epoch 10/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3510 - acc: 0.8399Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3511 - acc: 0.8397 - val_loss: 0.4014 - val_acc: 0.8085\n",
      "Epoch 00010: early stopping\n",
      "66666/66666 [==============================] - 2s 35us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.4673 - acc: 0.7655Epoch 00001: val_loss improved from inf to 0.42746, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 3s 20us/step - loss: 0.4651 - acc: 0.7670 - val_loss: 0.4275 - val_acc: 0.7951\n",
      "Epoch 2/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.4014 - acc: 0.8114Epoch 00002: val_loss improved from 0.42746 to 0.41446, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 11us/step - loss: 0.4011 - acc: 0.8116 - val_loss: 0.4145 - val_acc: 0.8030\n",
      "Epoch 3/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3876 - acc: 0.8188Epoch 00003: val_loss improved from 0.41446 to 0.40798, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 11us/step - loss: 0.3876 - acc: 0.8187 - val_loss: 0.4080 - val_acc: 0.8089\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 4/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3794 - acc: 0.8237Epoch 00004: val_loss improved from 0.40798 to 0.40560, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 1s 11us/step - loss: 0.3791 - acc: 0.8237 - val_loss: 0.4056 - val_acc: 0.8102\n",
      "Epoch 5/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3726 - acc: 0.8271Epoch 00005: val_loss improved from 0.40560 to 0.40470, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 1s 11us/step - loss: 0.3728 - acc: 0.8270 - val_loss: 0.4047 - val_acc: 0.8100\n",
      "Epoch 6/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3673 - acc: 0.8299Epoch 00006: val_loss improved from 0.40470 to 0.40184, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 1s 11us/step - loss: 0.3672 - acc: 0.8300 - val_loss: 0.4018 - val_acc: 0.8121\n",
      "Epoch 7/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3622 - acc: 0.8333Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 1s 11us/step - loss: 0.3619 - acc: 0.8335 - val_loss: 0.4029 - val_acc: 0.8131\n",
      "Epoch 8/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3575 - acc: 0.8352Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 11us/step - loss: 0.3575 - acc: 0.8353 - val_loss: 0.4019 - val_acc: 0.8116\n",
      "Epoch 9/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3532 - acc: 0.8378Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 1s 11us/step - loss: 0.3530 - acc: 0.8380 - val_loss: 0.4024 - val_acc: 0.8111\n",
      "Epoch 00009: early stopping\n",
      "66666/66666 [==============================] - 2s 30us/step\n",
      "Model:  basic_model_adam\n",
      "0.81% (+/- 0.01%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.4686 - acc: 0.7665Epoch 00001: val_loss improved from inf to 0.40450, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 5s 41us/step - loss: 0.4665 - acc: 0.7681 - val_loss: 0.4045 - val_acc: 0.8096\n",
      "Epoch 2/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4046 - acc: 0.8095Epoch 00002: val_loss improved from 0.40450 to 0.39147, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.4046 - acc: 0.8095 - val_loss: 0.3915 - val_acc: 0.8146\n",
      "Epoch 3/100\n",
      "128000/133332 [===========================>..] - ETA: 0s - loss: 0.3913 - acc: 0.8177Epoch 00003: val_loss improved from 0.39147 to 0.38730, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3913 - acc: 0.8176 - val_loss: 0.3873 - val_acc: 0.8191\n",
      "Epoch 4/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3829 - acc: 0.8229Epoch 00004: val_loss improved from 0.38730 to 0.38305, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3829 - acc: 0.8230 - val_loss: 0.3831 - val_acc: 0.8196\n",
      "Epoch 5/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3756 - acc: 0.8266Epoch 00005: val_loss improved from 0.38305 to 0.38029, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3755 - acc: 0.8263 - val_loss: 0.3803 - val_acc: 0.8211\n",
      "Epoch 6/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3695 - acc: 0.8302Epoch 00006: val_loss improved from 0.38029 to 0.37894, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3694 - acc: 0.8302 - val_loss: 0.3789 - val_acc: 0.8223\n",
      "Epoch 7/100\n",
      "128000/133332 [===========================>..] - ETA: 0s - loss: 0.3635 - acc: 0.8331Epoch 00007: val_loss improved from 0.37894 to 0.37823, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3642 - acc: 0.8327 - val_loss: 0.3782 - val_acc: 0.8225\n",
      "Epoch 8/100\n",
      "128000/133332 [===========================>..] - ETA: 0s - loss: 0.3595 - acc: 0.8355Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3597 - acc: 0.8355 - val_loss: 0.3786 - val_acc: 0.8222\n",
      "Epoch 9/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3553 - acc: 0.8377Epoch 00009: val_loss improved from 0.37823 to 0.37794, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3555 - acc: 0.8377 - val_loss: 0.3779 - val_acc: 0.8232\n",
      "Epoch 10/100\n",
      "128000/133332 [===========================>..] - ETA: 0s - loss: 0.3512 - acc: 0.8398Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3518 - acc: 0.8393 - val_loss: 0.3814 - val_acc: 0.8212\n",
      "Epoch 11/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3482 - acc: 0.8420Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3482 - acc: 0.8419 - val_loss: 0.3795 - val_acc: 0.8226\n",
      "Epoch 12/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3448 - acc: 0.8440Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3446 - acc: 0.8441 - val_loss: 0.3807 - val_acc: 0.8222\n",
      "Epoch 00012: early stopping\n",
      "66668/66668 [==============================] - 2s 34us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.4618 - acc: 0.7686Epoch 00001: val_loss improved from inf to 0.42457, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 3s 22us/step - loss: 0.4612 - acc: 0.7692 - val_loss: 0.4246 - val_acc: 0.7979\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3981 - acc: 0.8133Epoch 00002: val_loss improved from 0.42457 to 0.41039, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3981 - acc: 0.8133 - val_loss: 0.4104 - val_acc: 0.8061\n",
      "Epoch 3/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3853 - acc: 0.8197Epoch 00003: val_loss improved from 0.41039 to 0.40518, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3854 - acc: 0.8195 - val_loss: 0.4052 - val_acc: 0.8072\n",
      "Epoch 4/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3769 - acc: 0.8248Epoch 00004: val_loss improved from 0.40518 to 0.40148, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3768 - acc: 0.8249 - val_loss: 0.4015 - val_acc: 0.8113\n",
      "Epoch 5/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3700 - acc: 0.8288Epoch 00005: val_loss improved from 0.40148 to 0.39805, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 18us/step - loss: 0.3698 - acc: 0.8289 - val_loss: 0.3980 - val_acc: 0.8106\n",
      "Epoch 6/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3636 - acc: 0.8323Epoch 00006: val_loss improved from 0.39805 to 0.39777, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3640 - acc: 0.8322 - val_loss: 0.3978 - val_acc: 0.8117\n",
      "Epoch 7/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3587 - acc: 0.8355Epoch 00007: val_loss improved from 0.39777 to 0.39563, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3589 - acc: 0.8353 - val_loss: 0.3956 - val_acc: 0.8101\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 8/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3537 - acc: 0.8382Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3538 - acc: 0.8381 - val_loss: 0.3958 - val_acc: 0.8107\n",
      "Epoch 9/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3501 - acc: 0.8398Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3497 - acc: 0.8401 - val_loss: 0.3963 - val_acc: 0.8168\n",
      "Epoch 10/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3463 - acc: 0.8421Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3463 - acc: 0.8422 - val_loss: 0.3974 - val_acc: 0.8109\n",
      "Epoch 00010: early stopping\n",
      "66666/66666 [==============================] - 2s 35us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4631 - acc: 0.7680Epoch 00001: val_loss improved from inf to 0.42310, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 3s 19us/step - loss: 0.4612 - acc: 0.7694 - val_loss: 0.4231 - val_acc: 0.7981\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3959 - acc: 0.8146Epoch 00002: val_loss improved from 0.42310 to 0.41015, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3960 - acc: 0.8146 - val_loss: 0.4102 - val_acc: 0.8065\n",
      "Epoch 3/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3822 - acc: 0.8219Epoch 00003: val_loss improved from 0.41015 to 0.40368, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 11us/step - loss: 0.3823 - acc: 0.8218 - val_loss: 0.4037 - val_acc: 0.8108\n",
      "Epoch 4/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3741 - acc: 0.8264Epoch 00004: val_loss improved from 0.40368 to 0.40114, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 11us/step - loss: 0.3737 - acc: 0.8266 - val_loss: 0.4011 - val_acc: 0.8131\n",
      "Epoch 5/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3671 - acc: 0.8301Epoch 00005: val_loss improved from 0.40114 to 0.40053, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 11us/step - loss: 0.3672 - acc: 0.8300 - val_loss: 0.4005 - val_acc: 0.8132\n",
      "Epoch 6/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3617 - acc: 0.8337Epoch 00006: val_loss improved from 0.40053 to 0.39731, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3616 - acc: 0.8337 - val_loss: 0.3973 - val_acc: 0.8149\n",
      "Epoch 7/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3566 - acc: 0.8366Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 11us/step - loss: 0.3562 - acc: 0.8367 - val_loss: 0.3983 - val_acc: 0.8146\n",
      "Epoch 8/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3517 - acc: 0.8382Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 1s 11us/step - loss: 0.3517 - acc: 0.8383 - val_loss: 0.3979 - val_acc: 0.8143\n",
      "Epoch 9/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3472 - acc: 0.8413Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3470 - acc: 0.8414 - val_loss: 0.3976 - val_acc: 0.8139\n",
      "Epoch 00009: early stopping\n",
      "66666/66666 [==============================] - 2s 34us/step\n",
      "Model:  basic_model_adam\n",
      "0.82% (+/- 0.01%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4662 - acc: 0.7683Epoch 00001: val_loss improved from inf to 0.40440, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 6s 48us/step - loss: 0.4660 - acc: 0.7684 - val_loss: 0.4044 - val_acc: 0.8086\n",
      "Epoch 2/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.4043 - acc: 0.8101Epoch 00002: val_loss improved from 0.40440 to 0.39153, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.4042 - acc: 0.8101 - val_loss: 0.3915 - val_acc: 0.8145\n",
      "Epoch 3/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3908 - acc: 0.8173Epoch 00003: val_loss improved from 0.39153 to 0.38714, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3909 - acc: 0.8172 - val_loss: 0.3871 - val_acc: 0.8185\n",
      "Epoch 4/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3826 - acc: 0.8231Epoch 00004: val_loss improved from 0.38714 to 0.38316, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3825 - acc: 0.8231 - val_loss: 0.3832 - val_acc: 0.8195\n",
      "Epoch 5/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3749 - acc: 0.8269Epoch 00005: val_loss improved from 0.38316 to 0.38041, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3752 - acc: 0.8266 - val_loss: 0.3804 - val_acc: 0.8214\n",
      "Epoch 6/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3690 - acc: 0.8304Epoch 00006: val_loss improved from 0.38041 to 0.37899, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3691 - acc: 0.8303 - val_loss: 0.3790 - val_acc: 0.8224\n",
      "Epoch 7/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3635 - acc: 0.8332Epoch 00007: val_loss improved from 0.37899 to 0.37784, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3638 - acc: 0.8330 - val_loss: 0.3778 - val_acc: 0.8224\n",
      "Epoch 8/100\n",
      "128000/133332 [===========================>..] - ETA: 0s - loss: 0.3591 - acc: 0.8356Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3592 - acc: 0.8355 - val_loss: 0.3780 - val_acc: 0.8240\n",
      "Epoch 9/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3550 - acc: 0.8377Epoch 00009: val_loss improved from 0.37784 to 0.37771, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3551 - acc: 0.8376 - val_loss: 0.3777 - val_acc: 0.8232\n",
      "Epoch 10/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3516 - acc: 0.8399Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3515 - acc: 0.8400 - val_loss: 0.3818 - val_acc: 0.8212\n",
      "Epoch 11/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3479 - acc: 0.8419Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3480 - acc: 0.8419 - val_loss: 0.3795 - val_acc: 0.8227\n",
      "Epoch 12/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3441 - acc: 0.8446Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3441 - acc: 0.8447 - val_loss: 0.3804 - val_acc: 0.8224\n",
      "Epoch 00012: early stopping\n",
      "66668/66668 [==============================] - 3s 49us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4610 - acc: 0.7697Epoch 00001: val_loss improved from inf to 0.42445, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 4s 29us/step - loss: 0.4608 - acc: 0.7699 - val_loss: 0.4245 - val_acc: 0.7975\n",
      "Epoch 2/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3979 - acc: 0.8133Epoch 00002: val_loss improved from 0.42445 to 0.41017, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3977 - acc: 0.8133 - val_loss: 0.4102 - val_acc: 0.8059\n",
      "Epoch 3/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3852 - acc: 0.8198Epoch 00003: val_loss improved from 0.41017 to 0.40519, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3850 - acc: 0.8197 - val_loss: 0.4052 - val_acc: 0.8077\n",
      "Epoch 4/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3765 - acc: 0.8253Epoch 00004: val_loss improved from 0.40519 to 0.40162, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3765 - acc: 0.8252 - val_loss: 0.4016 - val_acc: 0.8112\n",
      "Epoch 5/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3698 - acc: 0.8291Epoch 00005: val_loss improved from 0.40162 to 0.39798, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3696 - acc: 0.8292 - val_loss: 0.3980 - val_acc: 0.8112\n",
      "Epoch 6/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3636 - acc: 0.8326Epoch 00006: val_loss improved from 0.39798 to 0.39759, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3638 - acc: 0.8325 - val_loss: 0.3976 - val_acc: 0.8113\n",
      "Epoch 7/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3585 - acc: 0.8357Epoch 00007: val_loss improved from 0.39759 to 0.39567, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3588 - acc: 0.8355 - val_loss: 0.3957 - val_acc: 0.8104\n",
      "Epoch 8/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3537 - acc: 0.8376Epoch 00008: val_loss improved from 0.39567 to 0.39540, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3536 - acc: 0.8375 - val_loss: 0.3954 - val_acc: 0.8119\n",
      "Epoch 9/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3496 - acc: 0.8396Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3495 - acc: 0.8397 - val_loss: 0.3960 - val_acc: 0.8167\n",
      "Epoch 10/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3462 - acc: 0.8419Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 11us/step - loss: 0.3460 - acc: 0.8421 - val_loss: 0.3962 - val_acc: 0.8118\n",
      "Epoch 11/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3421 - acc: 0.8446Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 11us/step - loss: 0.3423 - acc: 0.8443 - val_loss: 0.3981 - val_acc: 0.8106\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 2s 36us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4612 - acc: 0.7692Epoch 00001: val_loss improved from inf to 0.42272, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 3s 25us/step - loss: 0.4609 - acc: 0.7695 - val_loss: 0.4227 - val_acc: 0.7988\n",
      "Epoch 2/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3961 - acc: 0.8140Epoch 00002: val_loss improved from 0.42272 to 0.40980, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3958 - acc: 0.8142 - val_loss: 0.4098 - val_acc: 0.8061\n",
      "Epoch 3/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3820 - acc: 0.8220Epoch 00003: val_loss improved from 0.40980 to 0.40311, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3820 - acc: 0.8219 - val_loss: 0.4031 - val_acc: 0.8108\n",
      "Epoch 4/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3737 - acc: 0.8270Epoch 00004: val_loss improved from 0.40311 to 0.40071, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3734 - acc: 0.8271 - val_loss: 0.4007 - val_acc: 0.8132\n",
      "Epoch 5/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3668 - acc: 0.8301Epoch 00005: val_loss improved from 0.40071 to 0.40001, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3669 - acc: 0.8301 - val_loss: 0.4000 - val_acc: 0.8127\n",
      "Epoch 6/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3616 - acc: 0.8334Epoch 00006: val_loss improved from 0.40001 to 0.39697, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3615 - acc: 0.8334 - val_loss: 0.3970 - val_acc: 0.8143\n",
      "Epoch 7/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3564 - acc: 0.8362Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3561 - acc: 0.8363 - val_loss: 0.3980 - val_acc: 0.8155\n",
      "Epoch 8/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3516 - acc: 0.8384Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3517 - acc: 0.8385 - val_loss: 0.3974 - val_acc: 0.8137\n",
      "Epoch 9/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3470 - acc: 0.8414Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3471 - acc: 0.8413 - val_loss: 0.3974 - val_acc: 0.8140\n",
      "Epoch 00009: early stopping\n",
      "66666/66666 [==============================] - 2s 37us/step\n",
      "Model:  basic_model_adam\n",
      "0.82% (+/- 0.00%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "129024/133332 [============================>.] - ETA: 1s - loss: 0.4723 - acc: 0.7637Epoch 00001: val_loss improved from inf to 0.40985, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 41s 310us/step - loss: 0.4703 - acc: 0.7654 - val_loss: 0.4099 - val_acc: 0.8055\n",
      "Epoch 2/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.4092 - acc: 0.8076Epoch 00002: val_loss improved from 0.40985 to 0.39695, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.4092 - acc: 0.8075 - val_loss: 0.3970 - val_acc: 0.8113\n",
      "Epoch 3/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3956 - acc: 0.8153Epoch 00003: val_loss improved from 0.39695 to 0.39242, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3959 - acc: 0.8150 - val_loss: 0.3924 - val_acc: 0.8151\n",
      "Epoch 4/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3879 - acc: 0.8201Epoch 00004: val_loss improved from 0.39242 to 0.38886, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3879 - acc: 0.8202 - val_loss: 0.3889 - val_acc: 0.8175\n",
      "Epoch 5/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3805 - acc: 0.8243Epoch 00005: val_loss improved from 0.38886 to 0.38652, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3805 - acc: 0.8240 - val_loss: 0.3865 - val_acc: 0.8179\n",
      "Epoch 6/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3747 - acc: 0.8267- ETA: 1s - loss: Epoch 00006: val_loss improved from 0.38652 to 0.38459, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3747 - acc: 0.8266 - val_loss: 0.3846 - val_acc: 0.8188\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 7/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3689 - acc: 0.8299Epoch 00007: val_loss improved from 0.38459 to 0.38388, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3692 - acc: 0.8297 - val_loss: 0.3839 - val_acc: 0.8186\n",
      "Epoch 8/100\n",
      "128000/133332 [===========================>..] - ETA: 0s - loss: 0.3645 - acc: 0.8331Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3647 - acc: 0.8328 - val_loss: 0.3844 - val_acc: 0.8183\n",
      "Epoch 9/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3605 - acc: 0.8348Epoch 00009: val_loss improved from 0.38388 to 0.38288, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3606 - acc: 0.8346 - val_loss: 0.3829 - val_acc: 0.8199\n",
      "Epoch 10/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3567 - acc: 0.8369Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3569 - acc: 0.8368 - val_loss: 0.3880 - val_acc: 0.8181\n",
      "Epoch 11/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3530 - acc: 0.8395Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3535 - acc: 0.8391 - val_loss: 0.3854 - val_acc: 0.8196\n",
      "Epoch 12/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3499 - acc: 0.8412Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3499 - acc: 0.8413 - val_loss: 0.3866 - val_acc: 0.8197\n",
      "Epoch 00012: early stopping\n",
      "66668/66668 [==============================] - 3s 39us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.4655 - acc: 0.7665Epoch 00001: val_loss improved from inf to 0.42994, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 4s 27us/step - loss: 0.4649 - acc: 0.7671 - val_loss: 0.4299 - val_acc: 0.7940\n",
      "Epoch 2/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4031 - acc: 0.8103Epoch 00002: val_loss improved from 0.42994 to 0.41518, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.4030 - acc: 0.8103 - val_loss: 0.4152 - val_acc: 0.8032\n",
      "Epoch 3/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3904 - acc: 0.8165Epoch 00003: val_loss improved from 0.41518 to 0.41040, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3904 - acc: 0.8165 - val_loss: 0.4104 - val_acc: 0.8035\n",
      "Epoch 4/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3820 - acc: 0.8216Epoch 00004: val_loss improved from 0.41040 to 0.40569, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3820 - acc: 0.8216 - val_loss: 0.4057 - val_acc: 0.8077\n",
      "Epoch 5/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3753 - acc: 0.8265Epoch 00005: val_loss improved from 0.40569 to 0.40294, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3751 - acc: 0.8266 - val_loss: 0.4029 - val_acc: 0.8081\n",
      "Epoch 6/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3691 - acc: 0.8291Epoch 00006: val_loss improved from 0.40294 to 0.40260, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3693 - acc: 0.8291 - val_loss: 0.4026 - val_acc: 0.8079\n",
      "Epoch 7/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3643 - acc: 0.8330Epoch 00007: val_loss improved from 0.40260 to 0.40067, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3642 - acc: 0.8331 - val_loss: 0.4007 - val_acc: 0.8124\n",
      "Epoch 8/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3592 - acc: 0.8354Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3592 - acc: 0.8354 - val_loss: 0.4009 - val_acc: 0.8072\n",
      "Epoch 9/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3552 - acc: 0.8376Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3550 - acc: 0.8377 - val_loss: 0.4008 - val_acc: 0.8117\n",
      "Epoch 10/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3514 - acc: 0.8397Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3513 - acc: 0.8398 - val_loss: 0.4009 - val_acc: 0.8095\n",
      "Epoch 00010: early stopping\n",
      "66666/66666 [==============================] - 3s 42us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.4669 - acc: 0.7657Epoch 00001: val_loss improved from inf to 0.42773, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 4s 28us/step - loss: 0.4656 - acc: 0.7667 - val_loss: 0.4277 - val_acc: 0.7949\n",
      "Epoch 2/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.4017 - acc: 0.8112Epoch 00002: val_loss improved from 0.42773 to 0.41465, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.4013 - acc: 0.8115 - val_loss: 0.4146 - val_acc: 0.8029\n",
      "Epoch 3/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3878 - acc: 0.8185Epoch 00003: val_loss improved from 0.41465 to 0.40820, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3878 - acc: 0.8185 - val_loss: 0.4082 - val_acc: 0.8084\n",
      "Epoch 4/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3797 - acc: 0.8232Epoch 00004: val_loss improved from 0.40820 to 0.40587, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3794 - acc: 0.8233 - val_loss: 0.4059 - val_acc: 0.8097\n",
      "Epoch 5/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3729 - acc: 0.8269Epoch 00005: val_loss improved from 0.40587 to 0.40484, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3730 - acc: 0.8267 - val_loss: 0.4048 - val_acc: 0.8099\n",
      "Epoch 6/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3676 - acc: 0.8297Epoch 00006: val_loss improved from 0.40484 to 0.40199, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3675 - acc: 0.8298 - val_loss: 0.4020 - val_acc: 0.8122\n",
      "Epoch 7/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3623 - acc: 0.8328Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3622 - acc: 0.8329 - val_loss: 0.4030 - val_acc: 0.8133\n",
      "Epoch 8/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3576 - acc: 0.8354Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 3s 20us/step - loss: 0.3577 - acc: 0.8353 - val_loss: 0.4020 - val_acc: 0.8107\n",
      "Epoch 9/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3534 - acc: 0.8378Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3532 - acc: 0.8380 - val_loss: 0.4027 - val_acc: 0.8111\n",
      "Epoch 00009: early stopping\n",
      "66666/66666 [==============================] - 4s 54us/step\n",
      "Model:  basic_model_adam\n",
      "0.81% (+/- 0.00%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.4706 - acc: 0.7653Epoch 00001: val_loss improved from inf to 0.40996, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 9s 64us/step - loss: 0.4699 - acc: 0.7657 - val_loss: 0.4100 - val_acc: 0.8045\n",
      "Epoch 2/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.4093 - acc: 0.8073Epoch 00002: val_loss improved from 0.40996 to 0.39698, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.4092 - acc: 0.8072 - val_loss: 0.3970 - val_acc: 0.8106\n",
      "Epoch 3/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3959 - acc: 0.8156Epoch 00003: val_loss improved from 0.39698 to 0.39283, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3959 - acc: 0.8154 - val_loss: 0.3928 - val_acc: 0.8143\n",
      "Epoch 4/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3878 - acc: 0.8203Epoch 00004: val_loss improved from 0.39283 to 0.38892, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3878 - acc: 0.8203 - val_loss: 0.3889 - val_acc: 0.8178\n",
      "Epoch 5/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3805 - acc: 0.8239Epoch 00005: val_loss improved from 0.38892 to 0.38645, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3805 - acc: 0.8238 - val_loss: 0.3865 - val_acc: 0.8181\n",
      "Epoch 6/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3747 - acc: 0.8269Epoch 00006: val_loss improved from 0.38645 to 0.38485, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3747 - acc: 0.8268 - val_loss: 0.3848 - val_acc: 0.8193\n",
      "Epoch 7/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3693 - acc: 0.8298Epoch 00007: val_loss improved from 0.38485 to 0.38431, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3692 - acc: 0.8298 - val_loss: 0.3843 - val_acc: 0.8193\n",
      "Epoch 8/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3645 - acc: 0.8330Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3647 - acc: 0.8329 - val_loss: 0.3848 - val_acc: 0.8193\n",
      "Epoch 9/100\n",
      "128000/133332 [===========================>..] - ETA: 0s - loss: 0.3604 - acc: 0.8353Epoch 00009: val_loss improved from 0.38431 to 0.38341, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3606 - acc: 0.8351 - val_loss: 0.3834 - val_acc: 0.8210\n",
      "Epoch 10/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3567 - acc: 0.8373Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3571 - acc: 0.8369 - val_loss: 0.3880 - val_acc: 0.8182\n",
      "Epoch 11/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3536 - acc: 0.8391Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3537 - acc: 0.8390 - val_loss: 0.3857 - val_acc: 0.8193\n",
      "Epoch 12/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3500 - acc: 0.8408Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3498 - acc: 0.8410 - val_loss: 0.3869 - val_acc: 0.8193\n",
      "Epoch 00012: early stopping\n",
      "66668/66668 [==============================] - 3s 47us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.4663 - acc: 0.7659Epoch 00001: val_loss improved from inf to 0.42974, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 4s 28us/step - loss: 0.4645 - acc: 0.7671 - val_loss: 0.4297 - val_acc: 0.7938\n",
      "Epoch 2/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.4029 - acc: 0.8105Epoch 00002: val_loss improved from 0.42974 to 0.41539, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.4028 - acc: 0.8105 - val_loss: 0.4154 - val_acc: 0.8045\n",
      "Epoch 3/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3902 - acc: 0.8166Epoch 00003: val_loss improved from 0.41539 to 0.41058, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3902 - acc: 0.8165 - val_loss: 0.4106 - val_acc: 0.8035\n",
      "Epoch 4/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3818 - acc: 0.8218Epoch 00004: val_loss improved from 0.41058 to 0.40621, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3817 - acc: 0.8217 - val_loss: 0.4062 - val_acc: 0.8079\n",
      "Epoch 5/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3751 - acc: 0.8269Epoch 00005: val_loss improved from 0.40621 to 0.40305, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3750 - acc: 0.8269 - val_loss: 0.4031 - val_acc: 0.8081\n",
      "Epoch 6/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3689 - acc: 0.8295Epoch 00006: val_loss improved from 0.40305 to 0.40284, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 17us/step - loss: 0.3691 - acc: 0.8295 - val_loss: 0.4028 - val_acc: 0.8084\n",
      "Epoch 7/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3643 - acc: 0.8330Epoch 00007: val_loss improved from 0.40284 to 0.40047, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3641 - acc: 0.8328 - val_loss: 0.4005 - val_acc: 0.8127\n",
      "Epoch 8/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3593 - acc: 0.8350Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3592 - acc: 0.8351 - val_loss: 0.4012 - val_acc: 0.8075\n",
      "Epoch 9/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3554 - acc: 0.8373Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3550 - acc: 0.8375 - val_loss: 0.4011 - val_acc: 0.8134\n",
      "Epoch 10/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3516 - acc: 0.8401Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3514 - acc: 0.8400 - val_loss: 0.4009 - val_acc: 0.8136\n",
      "Epoch 00010: early stopping\n",
      "66666/66666 [==============================] - 4s 59us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4653 - acc: 0.7670Epoch 00001: val_loss improved from inf to 0.42747, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 5s 36us/step - loss: 0.4652 - acc: 0.7670 - val_loss: 0.4275 - val_acc: 0.7952\n",
      "Epoch 2/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4014 - acc: 0.8118Epoch 00002: val_loss improved from 0.42747 to 0.41459, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.4013 - acc: 0.8118 - val_loss: 0.4146 - val_acc: 0.8032\n",
      "Epoch 3/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3879 - acc: 0.8190Epoch 00003: val_loss improved from 0.41459 to 0.40814, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 3s 20us/step - loss: 0.3879 - acc: 0.8189 - val_loss: 0.4081 - val_acc: 0.8083\n",
      "Epoch 4/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3797 - acc: 0.8233Epoch 00004: val_loss improved from 0.40814 to 0.40566, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3794 - acc: 0.8234 - val_loss: 0.4057 - val_acc: 0.8100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 5/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3729 - acc: 0.8271Epoch 00005: val_loss improved from 0.40566 to 0.40463, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3730 - acc: 0.8271 - val_loss: 0.4046 - val_acc: 0.8097\n",
      "Epoch 6/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3676 - acc: 0.8304Epoch 00006: val_loss improved from 0.40463 to 0.40175, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3675 - acc: 0.8305 - val_loss: 0.4018 - val_acc: 0.8125\n",
      "Epoch 7/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3626 - acc: 0.8334Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3623 - acc: 0.8335 - val_loss: 0.4029 - val_acc: 0.8127\n",
      "Epoch 8/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3580 - acc: 0.8347Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3579 - acc: 0.8348 - val_loss: 0.4018 - val_acc: 0.8121\n",
      "Epoch 9/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3534 - acc: 0.8379Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 11us/step - loss: 0.3534 - acc: 0.8381 - val_loss: 0.4026 - val_acc: 0.8118\n",
      "Epoch 00009: early stopping\n",
      "66666/66666 [==============================] - 4s 56us/step\n",
      "Model:  basic_model_adam\n",
      "0.82% (+/- 0.00%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4700 - acc: 0.7652Epoch 00001: val_loss improved from inf to 0.40979, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 13s 94us/step - loss: 0.4699 - acc: 0.7653 - val_loss: 0.4098 - val_acc: 0.8053\n",
      "Epoch 2/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.4093 - acc: 0.8071Epoch 00002: val_loss improved from 0.40979 to 0.39667, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 17us/step - loss: 0.4091 - acc: 0.8070 - val_loss: 0.3967 - val_acc: 0.8116\n",
      "Epoch 3/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3959 - acc: 0.8154Epoch 00003: val_loss improved from 0.39667 to 0.39258, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3959 - acc: 0.8153 - val_loss: 0.3926 - val_acc: 0.8149\n",
      "Epoch 4/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3879 - acc: 0.8199Epoch 00004: val_loss improved from 0.39258 to 0.38875, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3879 - acc: 0.8200 - val_loss: 0.3887 - val_acc: 0.8179\n",
      "Epoch 5/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3806 - acc: 0.8241Epoch 00005: val_loss improved from 0.38875 to 0.38613, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3806 - acc: 0.8238 - val_loss: 0.3861 - val_acc: 0.8176\n",
      "Epoch 6/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3749 - acc: 0.8266Epoch 00006: val_loss improved from 0.38613 to 0.38446, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3747 - acc: 0.8266 - val_loss: 0.3845 - val_acc: 0.8197\n",
      "Epoch 7/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3694 - acc: 0.8294Epoch 00007: val_loss improved from 0.38446 to 0.38361, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3694 - acc: 0.8294 - val_loss: 0.3836 - val_acc: 0.8185\n",
      "Epoch 8/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3645 - acc: 0.8332Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3648 - acc: 0.8329 - val_loss: 0.3842 - val_acc: 0.8189\n",
      "Epoch 9/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3604 - acc: 0.8349Epoch 00009: val_loss improved from 0.38361 to 0.38289, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3606 - acc: 0.8347 - val_loss: 0.3829 - val_acc: 0.8194\n",
      "Epoch 10/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3567 - acc: 0.8370Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3569 - acc: 0.8369 - val_loss: 0.3880 - val_acc: 0.8177\n",
      "Epoch 11/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3532 - acc: 0.8394Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3534 - acc: 0.8392 - val_loss: 0.3853 - val_acc: 0.8196\n",
      "Epoch 12/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3499 - acc: 0.8407Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 16us/step - loss: 0.3497 - acc: 0.8408 - val_loss: 0.3864 - val_acc: 0.8196\n",
      "Epoch 00012: early stopping\n",
      "66668/66668 [==============================] - 4s 56us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4645 - acc: 0.7669Epoch 00001: val_loss improved from inf to 0.42967, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 4s 32us/step - loss: 0.4643 - acc: 0.7671 - val_loss: 0.4297 - val_acc: 0.7940\n",
      "Epoch 2/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.4027 - acc: 0.8107Epoch 00002: val_loss improved from 0.42967 to 0.41527, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.4027 - acc: 0.8107 - val_loss: 0.4153 - val_acc: 0.8032\n",
      "Epoch 3/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3903 - acc: 0.8166Epoch 00003: val_loss improved from 0.41527 to 0.41083, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3901 - acc: 0.8165 - val_loss: 0.4108 - val_acc: 0.8040\n",
      "Epoch 4/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3817 - acc: 0.8216Epoch 00004: val_loss improved from 0.41083 to 0.40647, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3817 - acc: 0.8216 - val_loss: 0.4065 - val_acc: 0.8079\n",
      "Epoch 5/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3750 - acc: 0.8268Epoch 00005: val_loss improved from 0.40647 to 0.40372, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3749 - acc: 0.8269 - val_loss: 0.4037 - val_acc: 0.8078\n",
      "Epoch 6/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3687 - acc: 0.8300Epoch 00006: val_loss improved from 0.40372 to 0.40331, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3691 - acc: 0.8297 - val_loss: 0.4033 - val_acc: 0.8071\n",
      "Epoch 7/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3640 - acc: 0.8333Epoch 00007: val_loss improved from 0.40331 to 0.40130, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3639 - acc: 0.8333 - val_loss: 0.4013 - val_acc: 0.8062\n",
      "Epoch 8/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3590 - acc: 0.8355Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3590 - acc: 0.8355 - val_loss: 0.4017 - val_acc: 0.8066\n",
      "Epoch 9/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3552 - acc: 0.8373Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3548 - acc: 0.8376 - val_loss: 0.4021 - val_acc: 0.8128\n",
      "Epoch 10/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3510 - acc: 0.8401Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3511 - acc: 0.8400 - val_loss: 0.4021 - val_acc: 0.8077\n",
      "Epoch 00010: early stopping\n",
      "66666/66666 [==============================] - 3s 46us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4668 - acc: 0.7662Epoch 00001: val_loss improved from inf to 0.42742, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 4s 33us/step - loss: 0.4650 - acc: 0.7676 - val_loss: 0.4274 - val_acc: 0.7946\n",
      "Epoch 2/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.4013 - acc: 0.8118Epoch 00002: val_loss improved from 0.42742 to 0.41444, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.4011 - acc: 0.8119 - val_loss: 0.4144 - val_acc: 0.8033\n",
      "Epoch 3/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3875 - acc: 0.8190Epoch 00003: val_loss improved from 0.41444 to 0.40814, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3876 - acc: 0.8190 - val_loss: 0.4081 - val_acc: 0.8084\n",
      "Epoch 4/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3795 - acc: 0.8236Epoch 00004: val_loss improved from 0.40814 to 0.40571, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3791 - acc: 0.8237 - val_loss: 0.4057 - val_acc: 0.8103\n",
      "Epoch 5/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3729 - acc: 0.8272Epoch 00005: val_loss improved from 0.40571 to 0.40484, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3728 - acc: 0.8271 - val_loss: 0.4048 - val_acc: 0.8102\n",
      "Epoch 6/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3673 - acc: 0.8300Epoch 00006: val_loss improved from 0.40484 to 0.40185, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3672 - acc: 0.8301 - val_loss: 0.4019 - val_acc: 0.8123\n",
      "Epoch 7/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3624 - acc: 0.8332Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3620 - acc: 0.8333 - val_loss: 0.4030 - val_acc: 0.8132\n",
      "Epoch 8/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3574 - acc: 0.8351Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3576 - acc: 0.8351 - val_loss: 0.4020 - val_acc: 0.8120\n",
      "Epoch 9/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3532 - acc: 0.8380Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3530 - acc: 0.8380 - val_loss: 0.4026 - val_acc: 0.8119\n",
      "Epoch 00009: early stopping\n",
      "66666/66666 [==============================] - 3s 47us/step\n",
      "Model:  basic_model_adam\n",
      "0.81% (+/- 0.01%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.4706 - acc: 0.7649Epoch 00001: val_loss improved from inf to 0.40995, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 12s 92us/step - loss: 0.4699 - acc: 0.7653 - val_loss: 0.4100 - val_acc: 0.8051\n",
      "Epoch 2/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.4092 - acc: 0.8071Epoch 00002: val_loss improved from 0.40995 to 0.39694, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.4090 - acc: 0.8071 - val_loss: 0.3969 - val_acc: 0.8117\n",
      "Epoch 3/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3957 - acc: 0.8156Epoch 00003: val_loss improved from 0.39694 to 0.39274, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3957 - acc: 0.8155 - val_loss: 0.3927 - val_acc: 0.8150\n",
      "Epoch 4/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3877 - acc: 0.8199Epoch 00004: val_loss improved from 0.39274 to 0.38904, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3877 - acc: 0.8200 - val_loss: 0.3890 - val_acc: 0.8180\n",
      "Epoch 5/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3804 - acc: 0.8241Epoch 00005: val_loss improved from 0.38904 to 0.38662, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3804 - acc: 0.8239 - val_loss: 0.3866 - val_acc: 0.8173\n",
      "Epoch 6/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3747 - acc: 0.8266Epoch 00006: val_loss improved from 0.38662 to 0.38499, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3745 - acc: 0.8266 - val_loss: 0.3850 - val_acc: 0.8193\n",
      "Epoch 7/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3686 - acc: 0.8303Epoch 00007: val_loss improved from 0.38499 to 0.38406, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3691 - acc: 0.8299 - val_loss: 0.3841 - val_acc: 0.8188\n",
      "Epoch 8/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3644 - acc: 0.8330Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3646 - acc: 0.8327 - val_loss: 0.3846 - val_acc: 0.8182\n",
      "Epoch 9/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3603 - acc: 0.8344Epoch 00009: val_loss improved from 0.38406 to 0.38338, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3604 - acc: 0.8343 - val_loss: 0.3834 - val_acc: 0.8196\n",
      "Epoch 10/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3565 - acc: 0.8369Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3568 - acc: 0.8367 - val_loss: 0.3880 - val_acc: 0.8179\n",
      "Epoch 11/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3533 - acc: 0.8397Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3535 - acc: 0.8396 - val_loss: 0.3857 - val_acc: 0.8201\n",
      "Epoch 12/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3500 - acc: 0.8406Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 18us/step - loss: 0.3498 - acc: 0.8408 - val_loss: 0.3870 - val_acc: 0.8201\n",
      "Epoch 00012: early stopping\n",
      "66668/66668 [==============================] - 3s 48us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.4656 - acc: 0.7664Epoch 00001: val_loss improved from inf to 0.42974, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 5s 35us/step - loss: 0.4645 - acc: 0.7673 - val_loss: 0.4297 - val_acc: 0.7940\n",
      "Epoch 2/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4029 - acc: 0.8106Epoch 00002: val_loss improved from 0.42974 to 0.41536, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.4027 - acc: 0.8106 - val_loss: 0.4154 - val_acc: 0.8035\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 3/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3903 - acc: 0.8165Epoch 00003: val_loss improved from 0.41536 to 0.41075, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3902 - acc: 0.8165 - val_loss: 0.4108 - val_acc: 0.8036\n",
      "Epoch 4/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3818 - acc: 0.8216Epoch 00004: val_loss improved from 0.41075 to 0.40649, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3817 - acc: 0.8215 - val_loss: 0.4065 - val_acc: 0.8081\n",
      "Epoch 5/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3750 - acc: 0.8264Epoch 00005: val_loss improved from 0.40649 to 0.40343, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3749 - acc: 0.8264 - val_loss: 0.4034 - val_acc: 0.8081\n",
      "Epoch 6/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3687 - acc: 0.8296Epoch 00006: val_loss improved from 0.40343 to 0.40338, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3690 - acc: 0.8295 - val_loss: 0.4034 - val_acc: 0.8081\n",
      "Epoch 7/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3641 - acc: 0.8335Epoch 00007: val_loss improved from 0.40338 to 0.40117, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3639 - acc: 0.8334 - val_loss: 0.4012 - val_acc: 0.8068\n",
      "Epoch 8/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3590 - acc: 0.8350Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3590 - acc: 0.8351 - val_loss: 0.4020 - val_acc: 0.8064\n",
      "Epoch 9/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3549 - acc: 0.8375Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3549 - acc: 0.8376 - val_loss: 0.4018 - val_acc: 0.8086\n",
      "Epoch 10/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3512 - acc: 0.8398Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3512 - acc: 0.8397 - val_loss: 0.4015 - val_acc: 0.8091\n",
      "Epoch 00010: early stopping\n",
      "66666/66666 [==============================] - 3s 52us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4670 - acc: 0.7657Epoch 00001: val_loss improved from inf to 0.42748, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 5s 34us/step - loss: 0.4652 - acc: 0.7670 - val_loss: 0.4275 - val_acc: 0.7950\n",
      "Epoch 2/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.4014 - acc: 0.8115Epoch 00002: val_loss improved from 0.42748 to 0.41458, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.4011 - acc: 0.8117 - val_loss: 0.4146 - val_acc: 0.8028\n",
      "Epoch 3/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3876 - acc: 0.8187Epoch 00003: val_loss improved from 0.41458 to 0.40800, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3876 - acc: 0.8185 - val_loss: 0.4080 - val_acc: 0.8083\n",
      "Epoch 4/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3794 - acc: 0.8235Epoch 00004: val_loss improved from 0.40800 to 0.40565, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3792 - acc: 0.8235 - val_loss: 0.4056 - val_acc: 0.8102\n",
      "Epoch 5/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3728 - acc: 0.8270Epoch 00005: val_loss improved from 0.40565 to 0.40459, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3728 - acc: 0.8270 - val_loss: 0.4046 - val_acc: 0.8101\n",
      "Epoch 6/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3674 - acc: 0.8300Epoch 00006: val_loss improved from 0.40459 to 0.40174, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3674 - acc: 0.8300 - val_loss: 0.4017 - val_acc: 0.8122\n",
      "Epoch 7/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3622 - acc: 0.8334Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3621 - acc: 0.8335 - val_loss: 0.4028 - val_acc: 0.8131\n",
      "Epoch 8/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3576 - acc: 0.8350Epoch 00008: val_loss improved from 0.40174 to 0.40172, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3576 - acc: 0.8351 - val_loss: 0.4017 - val_acc: 0.8118\n",
      "Epoch 9/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3533 - acc: 0.8379Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3531 - acc: 0.8380 - val_loss: 0.4021 - val_acc: 0.8116\n",
      "Epoch 10/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3493 - acc: 0.8398Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3495 - acc: 0.8397 - val_loss: 0.4029 - val_acc: 0.8132\n",
      "Epoch 11/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3458 - acc: 0.8419Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3460 - acc: 0.8418 - val_loss: 0.4039 - val_acc: 0.8117\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 3s 49us/step\n",
      "Model:  basic_model_adam\n",
      "0.81% (+/- 0.01%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4700 - acc: 0.7653Epoch 00001: val_loss improved from inf to 0.40981, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 56s 416us/step - loss: 0.4699 - acc: 0.7654 - val_loss: 0.4098 - val_acc: 0.8054\n",
      "Epoch 2/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4089 - acc: 0.8073Epoch 00002: val_loss improved from 0.40981 to 0.39675, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.4090 - acc: 0.8073 - val_loss: 0.3968 - val_acc: 0.8118\n",
      "Epoch 3/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3957 - acc: 0.8155Epoch 00003: val_loss improved from 0.39675 to 0.39265, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3957 - acc: 0.8154 - val_loss: 0.3927 - val_acc: 0.8146\n",
      "Epoch 4/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3876 - acc: 0.8204Epoch 00004: val_loss improved from 0.39265 to 0.38871, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3876 - acc: 0.8204 - val_loss: 0.3887 - val_acc: 0.8179\n",
      "Epoch 5/100\n",
      "128000/133332 [===========================>..] - ETA: 0s - loss: 0.3803 - acc: 0.8244Epoch 00005: val_loss improved from 0.38871 to 0.38619, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3802 - acc: 0.8241 - val_loss: 0.3862 - val_acc: 0.8177\n",
      "Epoch 6/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3744 - acc: 0.8268Epoch 00006: val_loss improved from 0.38619 to 0.38465, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3744 - acc: 0.8269 - val_loss: 0.3846 - val_acc: 0.8196\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 7/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3687 - acc: 0.8302Epoch 00007: val_loss improved from 0.38465 to 0.38377, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3689 - acc: 0.8300 - val_loss: 0.3838 - val_acc: 0.8188\n",
      "Epoch 8/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3640 - acc: 0.8325Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3644 - acc: 0.8322 - val_loss: 0.3843 - val_acc: 0.8192\n",
      "Epoch 9/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3601 - acc: 0.8344Epoch 00009: val_loss improved from 0.38377 to 0.38299, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3603 - acc: 0.8344 - val_loss: 0.3830 - val_acc: 0.8198\n",
      "Epoch 10/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3565 - acc: 0.8369Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3566 - acc: 0.8369 - val_loss: 0.3877 - val_acc: 0.8178\n",
      "Epoch 11/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3532 - acc: 0.8394Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3532 - acc: 0.8394 - val_loss: 0.3856 - val_acc: 0.8197\n",
      "Epoch 12/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3494 - acc: 0.8413Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3494 - acc: 0.8414 - val_loss: 0.3866 - val_acc: 0.8202\n",
      "Epoch 00012: early stopping\n",
      "66668/66668 [==============================] - 4s 55us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.4662 - acc: 0.7661Epoch 00001: val_loss improved from inf to 0.42965, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 5s 39us/step - loss: 0.4644 - acc: 0.7674 - val_loss: 0.4296 - val_acc: 0.7939\n",
      "Epoch 2/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.4028 - acc: 0.8106Epoch 00002: val_loss improved from 0.42965 to 0.41526, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.4026 - acc: 0.8105 - val_loss: 0.4153 - val_acc: 0.8036\n",
      "Epoch 3/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3901 - acc: 0.8167Epoch 00003: val_loss improved from 0.41526 to 0.41080, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3901 - acc: 0.8167 - val_loss: 0.4108 - val_acc: 0.8039\n",
      "Epoch 4/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3817 - acc: 0.8216Epoch 00004: val_loss improved from 0.41080 to 0.40633, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3817 - acc: 0.8216 - val_loss: 0.4063 - val_acc: 0.8078\n",
      "Epoch 5/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3749 - acc: 0.8265Epoch 00005: val_loss improved from 0.40633 to 0.40345, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3748 - acc: 0.8266 - val_loss: 0.4034 - val_acc: 0.8083\n",
      "Epoch 6/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3688 - acc: 0.8298Epoch 00006: val_loss improved from 0.40345 to 0.40270, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3689 - acc: 0.8297 - val_loss: 0.4027 - val_acc: 0.8085\n",
      "Epoch 7/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3636 - acc: 0.8336Epoch 00007: val_loss improved from 0.40270 to 0.40065, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3638 - acc: 0.8334 - val_loss: 0.4006 - val_acc: 0.8073\n",
      "Epoch 8/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3588 - acc: 0.8354Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3588 - acc: 0.8354 - val_loss: 0.4017 - val_acc: 0.8069\n",
      "Epoch 9/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3551 - acc: 0.8376Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3547 - acc: 0.8379 - val_loss: 0.4016 - val_acc: 0.8139\n",
      "Epoch 10/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3509 - acc: 0.8399Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3510 - acc: 0.8398 - val_loss: 0.4013 - val_acc: 0.8091\n",
      "Epoch 00010: early stopping\n",
      "66666/66666 [==============================] - 4s 53us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.4657 - acc: 0.7666Epoch 00001: val_loss improved from inf to 0.42731, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 5s 40us/step - loss: 0.4651 - acc: 0.7672 - val_loss: 0.4273 - val_acc: 0.7952\n",
      "Epoch 2/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.4013 - acc: 0.8114Epoch 00002: val_loss improved from 0.42731 to 0.41437, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.4010 - acc: 0.8116 - val_loss: 0.4144 - val_acc: 0.8032\n",
      "Epoch 3/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3876 - acc: 0.8186Epoch 00003: val_loss improved from 0.41437 to 0.40785, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3875 - acc: 0.8184 - val_loss: 0.4078 - val_acc: 0.8088\n",
      "Epoch 4/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3793 - acc: 0.8240Epoch 00004: val_loss improved from 0.40785 to 0.40542, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3791 - acc: 0.8240 - val_loss: 0.4054 - val_acc: 0.8097\n",
      "Epoch 5/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3728 - acc: 0.8271Epoch 00005: val_loss improved from 0.40542 to 0.40431, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3727 - acc: 0.8271 - val_loss: 0.4043 - val_acc: 0.8102\n",
      "Epoch 6/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3674 - acc: 0.8301Epoch 00006: val_loss improved from 0.40431 to 0.40152, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3672 - acc: 0.8302 - val_loss: 0.4015 - val_acc: 0.8120\n",
      "Epoch 7/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3622 - acc: 0.8332Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3619 - acc: 0.8334 - val_loss: 0.4026 - val_acc: 0.8134\n",
      "Epoch 8/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3573 - acc: 0.8352Epoch 00008: val_loss improved from 0.40152 to 0.40141, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3574 - acc: 0.8352 - val_loss: 0.4014 - val_acc: 0.8117\n",
      "Epoch 9/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3531 - acc: 0.8375Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3529 - acc: 0.8376 - val_loss: 0.4020 - val_acc: 0.8115\n",
      "Epoch 10/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3491 - acc: 0.8398Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3494 - acc: 0.8396 - val_loss: 0.4029 - val_acc: 0.8128\n",
      "Epoch 11/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3460 - acc: 0.8418Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3460 - acc: 0.8418 - val_loss: 0.4037 - val_acc: 0.8115\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 4s 57us/step\n",
      "Model:  basic_model_adam\n",
      "0.81% (+/- 0.01%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.4715 - acc: 0.7640Epoch 00001: val_loss improved from inf to 0.40993, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 15s 111us/step - loss: 0.4699 - acc: 0.7653 - val_loss: 0.4099 - val_acc: 0.8052\n",
      "Epoch 2/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.4091 - acc: 0.8073Epoch 00002: val_loss improved from 0.40993 to 0.39679, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.4090 - acc: 0.8072 - val_loss: 0.3968 - val_acc: 0.8118\n",
      "Epoch 3/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3957 - acc: 0.8156Epoch 00003: val_loss improved from 0.39679 to 0.39259, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3957 - acc: 0.8156 - val_loss: 0.3926 - val_acc: 0.8148\n",
      "Epoch 4/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3877 - acc: 0.8202Epoch 00004: val_loss improved from 0.39259 to 0.38873, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3876 - acc: 0.8201 - val_loss: 0.3887 - val_acc: 0.8187\n",
      "Epoch 5/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3803 - acc: 0.8242Epoch 00005: val_loss improved from 0.38873 to 0.38618, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3803 - acc: 0.8241 - val_loss: 0.3862 - val_acc: 0.8182\n",
      "Epoch 6/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3746 - acc: 0.8267Epoch 00006: val_loss improved from 0.38618 to 0.38460, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3745 - acc: 0.8266 - val_loss: 0.3846 - val_acc: 0.8197\n",
      "Epoch 7/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3691 - acc: 0.8296Epoch 00007: val_loss improved from 0.38460 to 0.38364, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3690 - acc: 0.8297 - val_loss: 0.3836 - val_acc: 0.8189\n",
      "Epoch 8/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3643 - acc: 0.8329Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3645 - acc: 0.8326 - val_loss: 0.3842 - val_acc: 0.8195\n",
      "Epoch 9/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3602 - acc: 0.8345Epoch 00009: val_loss improved from 0.38364 to 0.38302, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3604 - acc: 0.8344 - val_loss: 0.3830 - val_acc: 0.8199\n",
      "Epoch 10/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3565 - acc: 0.8371Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3568 - acc: 0.8369 - val_loss: 0.3878 - val_acc: 0.8199\n",
      "Epoch 11/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3534 - acc: 0.8393Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3534 - acc: 0.8393 - val_loss: 0.3854 - val_acc: 0.8198\n",
      "Epoch 12/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3496 - acc: 0.8411Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3496 - acc: 0.8411 - val_loss: 0.3865 - val_acc: 0.8203\n",
      "Epoch 00012: early stopping\n",
      "66668/66668 [==============================] - 4s 54us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4647 - acc: 0.7670Epoch 00001: val_loss improved from inf to 0.42971, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 5s 36us/step - loss: 0.4645 - acc: 0.7672 - val_loss: 0.4297 - val_acc: 0.7939\n",
      "Epoch 2/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.4029 - acc: 0.8105Epoch 00002: val_loss improved from 0.42971 to 0.41527, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.4027 - acc: 0.8105 - val_loss: 0.4153 - val_acc: 0.8035\n",
      "Epoch 3/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3903 - acc: 0.8165Epoch 00003: val_loss improved from 0.41527 to 0.41077, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3901 - acc: 0.8165 - val_loss: 0.4108 - val_acc: 0.8040\n",
      "Epoch 4/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3818 - acc: 0.8215Epoch 00004: val_loss improved from 0.41077 to 0.40619, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3817 - acc: 0.8214 - val_loss: 0.4062 - val_acc: 0.8084\n",
      "Epoch 5/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3750 - acc: 0.8265Epoch 00005: val_loss improved from 0.40619 to 0.40319, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3749 - acc: 0.8266 - val_loss: 0.4032 - val_acc: 0.8081\n",
      "Epoch 6/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3689 - acc: 0.8294Epoch 00006: val_loss improved from 0.40319 to 0.40289, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3690 - acc: 0.8293 - val_loss: 0.4029 - val_acc: 0.8082\n",
      "Epoch 7/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3641 - acc: 0.8331Epoch 00007: val_loss improved from 0.40289 to 0.40077, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3639 - acc: 0.8331 - val_loss: 0.4008 - val_acc: 0.8061\n",
      "Epoch 8/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3590 - acc: 0.8352Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3589 - acc: 0.8353 - val_loss: 0.4015 - val_acc: 0.8065\n",
      "Epoch 9/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3551 - acc: 0.8371Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3547 - acc: 0.8374 - val_loss: 0.4017 - val_acc: 0.8135\n",
      "Epoch 10/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3511 - acc: 0.8397Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3511 - acc: 0.8397 - val_loss: 0.4014 - val_acc: 0.8085\n",
      "Epoch 00010: early stopping\n",
      "66666/66666 [==============================] - 4s 54us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.4673 - acc: 0.7655Epoch 00001: val_loss improved from inf to 0.42746, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 5s 38us/step - loss: 0.4651 - acc: 0.7670 - val_loss: 0.4275 - val_acc: 0.7951\n",
      "Epoch 2/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4013 - acc: 0.8115Epoch 00002: val_loss improved from 0.42746 to 0.41446, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.4011 - acc: 0.8116 - val_loss: 0.4145 - val_acc: 0.8030\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 3/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3876 - acc: 0.8188Epoch 00003: val_loss improved from 0.41446 to 0.40798, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3876 - acc: 0.8187 - val_loss: 0.4080 - val_acc: 0.8089\n",
      "Epoch 4/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3796 - acc: 0.8236Epoch 00004: val_loss improved from 0.40798 to 0.40560, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3791 - acc: 0.8237 - val_loss: 0.4056 - val_acc: 0.8102\n",
      "Epoch 5/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3728 - acc: 0.8270Epoch 00005: val_loss improved from 0.40560 to 0.40470, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3728 - acc: 0.8270 - val_loss: 0.4047 - val_acc: 0.8100\n",
      "Epoch 6/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3673 - acc: 0.8300Epoch 00006: val_loss improved from 0.40470 to 0.40184, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3672 - acc: 0.8300 - val_loss: 0.4018 - val_acc: 0.8121\n",
      "Epoch 7/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3620 - acc: 0.8335Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3619 - acc: 0.8335 - val_loss: 0.4029 - val_acc: 0.8131\n",
      "Epoch 8/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3575 - acc: 0.8352Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3575 - acc: 0.8353 - val_loss: 0.4019 - val_acc: 0.8116\n",
      "Epoch 9/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3531 - acc: 0.8379Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3530 - acc: 0.8380 - val_loss: 0.4024 - val_acc: 0.8111\n",
      "Epoch 00009: early stopping\n",
      "66666/66666 [==============================] - 4s 55us/step\n",
      "Model:  basic_model_adam\n",
      "0.81% (+/- 0.01%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.4706 - acc: 0.7650Epoch 00001: val_loss improved from inf to 0.40980, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 54s 407us/step - loss: 0.4699 - acc: 0.7654 - val_loss: 0.4098 - val_acc: 0.8060\n",
      "Epoch 2/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.4093 - acc: 0.8069Epoch 00002: val_loss improved from 0.40980 to 0.39697, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.4091 - acc: 0.8069 - val_loss: 0.3970 - val_acc: 0.8115\n",
      "Epoch 3/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3958 - acc: 0.8151Epoch 00003: val_loss improved from 0.39697 to 0.39286, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3959 - acc: 0.8151 - val_loss: 0.3929 - val_acc: 0.8136\n",
      "Epoch 4/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3876 - acc: 0.8201Epoch 00004: val_loss improved from 0.39286 to 0.38876, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3877 - acc: 0.8201 - val_loss: 0.3888 - val_acc: 0.8165\n",
      "Epoch 5/100\n",
      "128000/133332 [===========================>..] - ETA: 0s - loss: 0.3804 - acc: 0.8242Epoch 00005: val_loss improved from 0.38876 to 0.38637, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3803 - acc: 0.8240 - val_loss: 0.3864 - val_acc: 0.8167\n",
      "Epoch 6/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3745 - acc: 0.8275Epoch 00006: val_loss improved from 0.38637 to 0.38451, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3745 - acc: 0.8275 - val_loss: 0.3845 - val_acc: 0.8177\n",
      "Epoch 7/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3689 - acc: 0.8302Epoch 00007: val_loss improved from 0.38451 to 0.38364, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3691 - acc: 0.8300 - val_loss: 0.3836 - val_acc: 0.8189\n",
      "Epoch 8/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3643 - acc: 0.8333Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3644 - acc: 0.8331 - val_loss: 0.3844 - val_acc: 0.8186\n",
      "Epoch 9/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3602 - acc: 0.8349Epoch 00009: val_loss improved from 0.38364 to 0.38304, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3604 - acc: 0.8348 - val_loss: 0.3830 - val_acc: 0.8205\n",
      "Epoch 10/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3563 - acc: 0.8369Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3566 - acc: 0.8367 - val_loss: 0.3877 - val_acc: 0.8181\n",
      "Epoch 11/100\n",
      "128000/133332 [===========================>..] - ETA: 0s - loss: 0.3526 - acc: 0.8397Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3531 - acc: 0.8393 - val_loss: 0.3854 - val_acc: 0.8191\n",
      "Epoch 12/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3492 - acc: 0.8408Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3492 - acc: 0.8409 - val_loss: 0.3866 - val_acc: 0.8191\n",
      "Epoch 00012: early stopping\n",
      "66668/66668 [==============================] - 4s 60us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.4651 - acc: 0.7667Epoch 00001: val_loss improved from inf to 0.43005, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 6s 45us/step - loss: 0.4645 - acc: 0.7673 - val_loss: 0.4301 - val_acc: 0.7937\n",
      "Epoch 2/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4030 - acc: 0.8106Epoch 00002: val_loss improved from 0.43005 to 0.41536, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.4029 - acc: 0.8106 - val_loss: 0.4154 - val_acc: 0.8032\n",
      "Epoch 3/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3905 - acc: 0.8162Epoch 00003: val_loss improved from 0.41536 to 0.41084, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3902 - acc: 0.8164 - val_loss: 0.4108 - val_acc: 0.8038\n",
      "Epoch 4/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3817 - acc: 0.8219Epoch 00004: val_loss improved from 0.41084 to 0.40692, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3817 - acc: 0.8218 - val_loss: 0.4069 - val_acc: 0.8089\n",
      "Epoch 5/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3747 - acc: 0.8268Epoch 00005: val_loss improved from 0.40692 to 0.40381, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3748 - acc: 0.8268 - val_loss: 0.4038 - val_acc: 0.8084\n",
      "Epoch 6/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3685 - acc: 0.8300Epoch 00006: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3689 - acc: 0.8296 - val_loss: 0.4042 - val_acc: 0.8077\n",
      "Epoch 7/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3636 - acc: 0.8333Epoch 00007: val_loss improved from 0.40381 to 0.40226, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3637 - acc: 0.8331 - val_loss: 0.4023 - val_acc: 0.8063\n",
      "Epoch 8/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3587 - acc: 0.8358Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3586 - acc: 0.8358 - val_loss: 0.4028 - val_acc: 0.8067\n",
      "Epoch 9/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3545 - acc: 0.8379Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3545 - acc: 0.8380 - val_loss: 0.4030 - val_acc: 0.8071\n",
      "Epoch 10/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3508 - acc: 0.8400Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3508 - acc: 0.8399 - val_loss: 0.4031 - val_acc: 0.8084\n",
      "Epoch 00010: early stopping\n",
      "66666/66666 [==============================] - 4s 56us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.4656 - acc: 0.7670Epoch 00001: val_loss improved from inf to 0.42784, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 6s 43us/step - loss: 0.4650 - acc: 0.7675 - val_loss: 0.4278 - val_acc: 0.7952\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4010 - acc: 0.8113Epoch 00002: val_loss improved from 0.42784 to 0.41481, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.4011 - acc: 0.8113 - val_loss: 0.4148 - val_acc: 0.8038\n",
      "Epoch 3/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3875 - acc: 0.8184Epoch 00003: val_loss improved from 0.41481 to 0.40840, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3876 - acc: 0.8183 - val_loss: 0.4084 - val_acc: 0.8083\n",
      "Epoch 4/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3795 - acc: 0.8233Epoch 00004: val_loss improved from 0.40840 to 0.40605, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3791 - acc: 0.8234 - val_loss: 0.4061 - val_acc: 0.8102\n",
      "Epoch 5/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3727 - acc: 0.8272Epoch 00005: val_loss improved from 0.40605 to 0.40488, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3727 - acc: 0.8272 - val_loss: 0.4049 - val_acc: 0.8099\n",
      "Epoch 6/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3672 - acc: 0.8303Epoch 00006: val_loss improved from 0.40488 to 0.40205, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3671 - acc: 0.8303 - val_loss: 0.4020 - val_acc: 0.8118\n",
      "Epoch 7/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3621 - acc: 0.8330Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3618 - acc: 0.8331 - val_loss: 0.4033 - val_acc: 0.8125\n",
      "Epoch 8/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3574 - acc: 0.8353Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3573 - acc: 0.8354 - val_loss: 0.4021 - val_acc: 0.8116\n",
      "Epoch 9/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3528 - acc: 0.8378Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3529 - acc: 0.8377 - val_loss: 0.4030 - val_acc: 0.8112\n",
      "Epoch 00009: early stopping\n",
      "66666/66666 [==============================] - 4s 59us/step\n",
      "Model:  basic_model_adam\n",
      "0.81% (+/- 0.01%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "128000/133332 [===========================>..] - ETA: 0s - loss: 0.4919 - acc: 0.7492Epoch 00001: val_loss improved from inf to 0.43217, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 17s 130us/step - loss: 0.4896 - acc: 0.7511 - val_loss: 0.4322 - val_acc: 0.7870\n",
      "Epoch 2/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.4290 - acc: 0.7946Epoch 00002: val_loss improved from 0.43217 to 0.41715, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.4288 - acc: 0.7947 - val_loss: 0.4172 - val_acc: 0.7947\n",
      "Epoch 3/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.4140 - acc: 0.8041Epoch 00003: val_loss improved from 0.41715 to 0.41319, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.4137 - acc: 0.8043 - val_loss: 0.4132 - val_acc: 0.7987\n",
      "Epoch 4/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4048 - acc: 0.8094Epoch 00004: val_loss improved from 0.41319 to 0.40683, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.4048 - acc: 0.8094 - val_loss: 0.4068 - val_acc: 0.8048\n",
      "Epoch 5/100\n",
      "128000/133332 [===========================>..] - ETA: 0s - loss: 0.3967 - acc: 0.8149Epoch 00005: val_loss improved from 0.40683 to 0.40416, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3967 - acc: 0.8146 - val_loss: 0.4042 - val_acc: 0.8069\n",
      "Epoch 6/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3905 - acc: 0.8181Epoch 00006: val_loss improved from 0.40416 to 0.40126, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3904 - acc: 0.8181 - val_loss: 0.4013 - val_acc: 0.8093\n",
      "Epoch 7/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3845 - acc: 0.8211Epoch 00007: val_loss improved from 0.40126 to 0.40052, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3844 - acc: 0.8212 - val_loss: 0.4005 - val_acc: 0.8096\n",
      "Epoch 8/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3794 - acc: 0.8237Epoch 00008: val_loss improved from 0.40052 to 0.40051, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3796 - acc: 0.8236 - val_loss: 0.4005 - val_acc: 0.8088\n",
      "Epoch 9/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3748 - acc: 0.8272Epoch 00009: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3750 - acc: 0.8270 - val_loss: 0.4012 - val_acc: 0.8086\n",
      "Epoch 10/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3706 - acc: 0.8292Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3711 - acc: 0.8289 - val_loss: 0.4037 - val_acc: 0.8083\n",
      "Epoch 11/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3675 - acc: 0.8317Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3674 - acc: 0.8317 - val_loss: 0.4022 - val_acc: 0.8095\n",
      "Epoch 00011: early stopping\n",
      "66668/66668 [==============================] - 4s 57us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.4869 - acc: 0.7502Epoch 00001: val_loss improved from inf to 0.44865, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 5s 41us/step - loss: 0.4851 - acc: 0.7516 - val_loss: 0.4487 - val_acc: 0.7798\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 2/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.4227 - acc: 0.7964Epoch 00002: val_loss improved from 0.44865 to 0.43320, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.4225 - acc: 0.7965 - val_loss: 0.4332 - val_acc: 0.7891\n",
      "Epoch 3/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4078 - acc: 0.8057Epoch 00003: val_loss improved from 0.43320 to 0.42713, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.4078 - acc: 0.8058 - val_loss: 0.4271 - val_acc: 0.7927\n",
      "Epoch 4/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3979 - acc: 0.8117Epoch 00004: val_loss improved from 0.42713 to 0.42207, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3979 - acc: 0.8115 - val_loss: 0.4221 - val_acc: 0.8009\n",
      "Epoch 5/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3906 - acc: 0.8157Epoch 00005: val_loss improved from 0.42207 to 0.41838, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3905 - acc: 0.8159 - val_loss: 0.4184 - val_acc: 0.8010\n",
      "Epoch 6/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3830 - acc: 0.8206Epoch 00006: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3832 - acc: 0.8204 - val_loss: 0.4194 - val_acc: 0.8018\n",
      "Epoch 7/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3780 - acc: 0.8232Epoch 00007: val_loss improved from 0.41838 to 0.41652, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3779 - acc: 0.8232 - val_loss: 0.4165 - val_acc: 0.8025\n",
      "Epoch 8/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3726 - acc: 0.8263Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3725 - acc: 0.8263 - val_loss: 0.4171 - val_acc: 0.8018\n",
      "Epoch 9/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3679 - acc: 0.8291Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3677 - acc: 0.8291 - val_loss: 0.4171 - val_acc: 0.8031\n",
      "Epoch 10/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3639 - acc: 0.8315Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3639 - acc: 0.8316 - val_loss: 0.4175 - val_acc: 0.8048\n",
      "Epoch 00010: early stopping\n",
      "66666/66666 [==============================] - 5s 68us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.4863 - acc: 0.7500Epoch 00001: val_loss improved from inf to 0.44904, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 7s 50us/step - loss: 0.4850 - acc: 0.7509 - val_loss: 0.4490 - val_acc: 0.7806\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4211 - acc: 0.7982Epoch 00002: val_loss improved from 0.44904 to 0.43359, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.4212 - acc: 0.7981 - val_loss: 0.4336 - val_acc: 0.7903\n",
      "Epoch 3/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4054 - acc: 0.8075Epoch 00003: val_loss improved from 0.43359 to 0.42554, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.4054 - acc: 0.8076 - val_loss: 0.4255 - val_acc: 0.7968\n",
      "Epoch 4/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3956 - acc: 0.8133Epoch 00004: val_loss improved from 0.42554 to 0.42357, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 18us/step - loss: 0.3954 - acc: 0.8134 - val_loss: 0.4236 - val_acc: 0.8004\n",
      "Epoch 5/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3880 - acc: 0.8171Epoch 00005: val_loss improved from 0.42357 to 0.42104, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 18us/step - loss: 0.3883 - acc: 0.8170 - val_loss: 0.4210 - val_acc: 0.8001\n",
      "Epoch 6/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3816 - acc: 0.8210Epoch 00006: val_loss improved from 0.42104 to 0.41829, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3816 - acc: 0.8209 - val_loss: 0.4183 - val_acc: 0.8025\n",
      "Epoch 7/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3761 - acc: 0.8236Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3760 - acc: 0.8237 - val_loss: 0.4198 - val_acc: 0.8003\n",
      "Epoch 8/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3706 - acc: 0.8269Epoch 00008: val_loss improved from 0.41829 to 0.41768, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3707 - acc: 0.8268 - val_loss: 0.4177 - val_acc: 0.8032\n",
      "Epoch 9/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3661 - acc: 0.8290Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3660 - acc: 0.8291 - val_loss: 0.4190 - val_acc: 0.8027\n",
      "Epoch 10/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3622 - acc: 0.8314Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3624 - acc: 0.8313 - val_loss: 0.4204 - val_acc: 0.8012\n",
      "Epoch 11/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3584 - acc: 0.8337Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3584 - acc: 0.8336 - val_loss: 0.4201 - val_acc: 0.8027\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 4s 59us/step\n",
      "Model:  basic_model_adam\n",
      "0.80% (+/- 0.00%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4684 - acc: 0.7671Epoch 00001: val_loss improved from inf to 0.41113, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 19s 143us/step - loss: 0.4683 - acc: 0.7671 - val_loss: 0.4111 - val_acc: 0.8021\n",
      "Epoch 2/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.4092 - acc: 0.8073Epoch 00002: val_loss improved from 0.41113 to 0.39805, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.4090 - acc: 0.8073 - val_loss: 0.3980 - val_acc: 0.8114\n",
      "Epoch 3/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3952 - acc: 0.8151Epoch 00003: val_loss improved from 0.39805 to 0.39366, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3954 - acc: 0.8149 - val_loss: 0.3937 - val_acc: 0.8155\n",
      "Epoch 4/100\n",
      "128000/133332 [===========================>..] - ETA: 0s - loss: 0.3874 - acc: 0.8205Epoch 00004: val_loss improved from 0.39366 to 0.38899, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3874 - acc: 0.8205 - val_loss: 0.3890 - val_acc: 0.8169\n",
      "Epoch 5/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3799 - acc: 0.8248Epoch 00005: val_loss improved from 0.38899 to 0.38723, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3797 - acc: 0.8247 - val_loss: 0.3872 - val_acc: 0.8186\n",
      "Epoch 6/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3738 - acc: 0.8275Epoch 00006: val_loss improved from 0.38723 to 0.38639, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3736 - acc: 0.8276 - val_loss: 0.3864 - val_acc: 0.8181\n",
      "Epoch 7/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3683 - acc: 0.8305Epoch 00007: val_loss improved from 0.38639 to 0.38487, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3685 - acc: 0.8303 - val_loss: 0.3849 - val_acc: 0.8196\n",
      "Epoch 8/100\n",
      "128000/133332 [===========================>..] - ETA: 0s - loss: 0.3635 - acc: 0.8333Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3636 - acc: 0.8333 - val_loss: 0.3861 - val_acc: 0.8187\n",
      "Epoch 9/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3596 - acc: 0.8356Epoch 00009: val_loss improved from 0.38487 to 0.38472, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3597 - acc: 0.8356 - val_loss: 0.3847 - val_acc: 0.8215\n",
      "Epoch 10/100\n",
      "128000/133332 [===========================>..] - ETA: 0s - loss: 0.3554 - acc: 0.8371Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3563 - acc: 0.8365 - val_loss: 0.3891 - val_acc: 0.8166\n",
      "Epoch 11/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3523 - acc: 0.8397Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3523 - acc: 0.8398 - val_loss: 0.3864 - val_acc: 0.8187\n",
      "Epoch 12/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3488 - acc: 0.8414Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3489 - acc: 0.8414 - val_loss: 0.3874 - val_acc: 0.8202\n",
      "Epoch 00012: early stopping\n",
      "66668/66668 [==============================] - 4s 58us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4632 - acc: 0.7684Epoch 00001: val_loss improved from inf to 0.42858, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 6s 43us/step - loss: 0.4630 - acc: 0.7685 - val_loss: 0.4286 - val_acc: 0.7948\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4027 - acc: 0.8098Epoch 00002: val_loss improved from 0.42858 to 0.41505, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.4027 - acc: 0.8098 - val_loss: 0.4151 - val_acc: 0.8033\n",
      "Epoch 3/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3900 - acc: 0.8170Epoch 00003: val_loss improved from 0.41505 to 0.41031, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3900 - acc: 0.8170 - val_loss: 0.4103 - val_acc: 0.8045\n",
      "Epoch 4/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3814 - acc: 0.8220Epoch 00004: val_loss improved from 0.41031 to 0.40565, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3813 - acc: 0.8220 - val_loss: 0.4056 - val_acc: 0.8080\n",
      "Epoch 5/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3744 - acc: 0.8262Epoch 00005: val_loss improved from 0.40565 to 0.40395, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3745 - acc: 0.8262 - val_loss: 0.4040 - val_acc: 0.8085\n",
      "Epoch 6/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3685 - acc: 0.8297Epoch 00006: val_loss improved from 0.40395 to 0.40163, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3687 - acc: 0.8295 - val_loss: 0.4016 - val_acc: 0.8079\n",
      "Epoch 7/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3638 - acc: 0.8322Epoch 00007: val_loss improved from 0.40163 to 0.40050, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3637 - acc: 0.8323 - val_loss: 0.4005 - val_acc: 0.8088\n",
      "Epoch 8/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3588 - acc: 0.8354Epoch 00008: val_loss improved from 0.40050 to 0.40044, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3588 - acc: 0.8354 - val_loss: 0.4004 - val_acc: 0.8084\n",
      "Epoch 9/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3547 - acc: 0.8378Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3544 - acc: 0.8379 - val_loss: 0.4009 - val_acc: 0.8138\n",
      "Epoch 10/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3512 - acc: 0.8388Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3512 - acc: 0.8388 - val_loss: 0.4016 - val_acc: 0.8095\n",
      "Epoch 11/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3475 - acc: 0.8413Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3475 - acc: 0.8414 - val_loss: 0.4007 - val_acc: 0.8134\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 4s 62us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4642 - acc: 0.7675Epoch 00001: val_loss improved from inf to 0.42688, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 6s 45us/step - loss: 0.4639 - acc: 0.7678 - val_loss: 0.4269 - val_acc: 0.7948\n",
      "Epoch 2/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4019 - acc: 0.8101Epoch 00002: val_loss improved from 0.42688 to 0.41284, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.4017 - acc: 0.8102 - val_loss: 0.4128 - val_acc: 0.8044\n",
      "Epoch 3/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3879 - acc: 0.8188Epoch 00003: val_loss improved from 0.41284 to 0.40725, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3881 - acc: 0.8186 - val_loss: 0.4072 - val_acc: 0.8091\n",
      "Epoch 4/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3794 - acc: 0.8233Epoch 00004: val_loss improved from 0.40725 to 0.40387, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3791 - acc: 0.8234 - val_loss: 0.4039 - val_acc: 0.8100\n",
      "Epoch 5/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3724 - acc: 0.8272Epoch 00005: val_loss improved from 0.40387 to 0.40182, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3725 - acc: 0.8270 - val_loss: 0.4018 - val_acc: 0.8110\n",
      "Epoch 6/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3668 - acc: 0.8310Epoch 00006: val_loss improved from 0.40182 to 0.39909, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3668 - acc: 0.8310 - val_loss: 0.3991 - val_acc: 0.8120\n",
      "Epoch 7/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3616 - acc: 0.8332Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3613 - acc: 0.8336 - val_loss: 0.4007 - val_acc: 0.8129\n",
      "Epoch 8/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3564 - acc: 0.8359Epoch 00008: val_loss improved from 0.39909 to 0.39830, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3564 - acc: 0.8360 - val_loss: 0.3983 - val_acc: 0.8129\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 9/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3521 - acc: 0.8391Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3520 - acc: 0.8392 - val_loss: 0.3992 - val_acc: 0.8131\n",
      "Epoch 10/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3481 - acc: 0.8412Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3482 - acc: 0.8411 - val_loss: 0.3999 - val_acc: 0.8124\n",
      "Epoch 11/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3444 - acc: 0.8431Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3448 - acc: 0.8430 - val_loss: 0.4019 - val_acc: 0.8131\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 4s 66us/step\n",
      "Model:  basic_model_adam\n",
      "0.81% (+/- 0.01%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4688 - acc: 0.7662Epoch 00001: val_loss improved from inf to 0.40838, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 64s 479us/step - loss: 0.4686 - acc: 0.7662 - val_loss: 0.4084 - val_acc: 0.8066\n",
      "Epoch 2/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4072 - acc: 0.8084Epoch 00002: val_loss improved from 0.40838 to 0.39564, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.4073 - acc: 0.8084 - val_loss: 0.3956 - val_acc: 0.8129\n",
      "Epoch 3/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3940 - acc: 0.8164Epoch 00003: val_loss improved from 0.39564 to 0.39112, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3940 - acc: 0.8163 - val_loss: 0.3911 - val_acc: 0.8159\n",
      "Epoch 4/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3860 - acc: 0.8214Epoch 00004: val_loss improved from 0.39112 to 0.38757, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3860 - acc: 0.8214 - val_loss: 0.3876 - val_acc: 0.8184\n",
      "Epoch 5/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3785 - acc: 0.8254Epoch 00005: val_loss improved from 0.38757 to 0.38551, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3788 - acc: 0.8251 - val_loss: 0.3855 - val_acc: 0.8184\n",
      "Epoch 6/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3730 - acc: 0.8284Epoch 00006: val_loss improved from 0.38551 to 0.38370, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 17us/step - loss: 0.3728 - acc: 0.8283 - val_loss: 0.3837 - val_acc: 0.8202\n",
      "Epoch 7/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3675 - acc: 0.8315Epoch 00007: val_loss improved from 0.38370 to 0.38285, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3675 - acc: 0.8315 - val_loss: 0.3829 - val_acc: 0.8192\n",
      "Epoch 8/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3625 - acc: 0.8341Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3628 - acc: 0.8339 - val_loss: 0.3835 - val_acc: 0.8188\n",
      "Epoch 9/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3587 - acc: 0.8360Epoch 00009: val_loss improved from 0.38285 to 0.38211, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3588 - acc: 0.8359 - val_loss: 0.3821 - val_acc: 0.8206\n",
      "Epoch 10/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3553 - acc: 0.8376Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3553 - acc: 0.8376 - val_loss: 0.3876 - val_acc: 0.8178\n",
      "Epoch 11/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3518 - acc: 0.8400Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3520 - acc: 0.8399 - val_loss: 0.3851 - val_acc: 0.8196\n",
      "Epoch 12/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3480 - acc: 0.8422Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 12us/step - loss: 0.3480 - acc: 0.8422 - val_loss: 0.3861 - val_acc: 0.8187\n",
      "Epoch 00012: early stopping\n",
      "66668/66668 [==============================] - 5s 69us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4646 - acc: 0.7667Epoch 00001: val_loss improved from inf to 0.42842, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 7s 55us/step - loss: 0.4631 - acc: 0.7678 - val_loss: 0.4284 - val_acc: 0.7950\n",
      "Epoch 2/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4010 - acc: 0.8118Epoch 00002: val_loss improved from 0.42842 to 0.41384, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.4008 - acc: 0.8118 - val_loss: 0.4138 - val_acc: 0.8041\n",
      "Epoch 3/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3885 - acc: 0.8178Epoch 00003: val_loss improved from 0.41384 to 0.40954, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3883 - acc: 0.8178 - val_loss: 0.4095 - val_acc: 0.8047\n",
      "Epoch 4/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3799 - acc: 0.8228Epoch 00004: val_loss improved from 0.40954 to 0.40509, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3799 - acc: 0.8227 - val_loss: 0.4051 - val_acc: 0.8094\n",
      "Epoch 5/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3734 - acc: 0.8275Epoch 00005: val_loss improved from 0.40509 to 0.40214, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3730 - acc: 0.8278 - val_loss: 0.4021 - val_acc: 0.8093\n",
      "Epoch 6/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3669 - acc: 0.8306Epoch 00006: val_loss improved from 0.40214 to 0.40172, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3672 - acc: 0.8305 - val_loss: 0.4017 - val_acc: 0.8092\n",
      "Epoch 7/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3624 - acc: 0.8340Epoch 00007: val_loss improved from 0.40172 to 0.39985, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3622 - acc: 0.8340 - val_loss: 0.3998 - val_acc: 0.8084\n",
      "Epoch 8/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3573 - acc: 0.8359Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3573 - acc: 0.8360 - val_loss: 0.4001 - val_acc: 0.8070\n",
      "Epoch 9/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3533 - acc: 0.8389Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3531 - acc: 0.8390 - val_loss: 0.4008 - val_acc: 0.8092\n",
      "Epoch 10/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3495 - acc: 0.8410Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3494 - acc: 0.8410 - val_loss: 0.4005 - val_acc: 0.8093\n",
      "Epoch 00010: early stopping\n",
      "66666/66666 [==============================] - 5s 74us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4658 - acc: 0.7664Epoch 00001: val_loss improved from inf to 0.42585, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 7s 53us/step - loss: 0.4640 - acc: 0.7678 - val_loss: 0.4258 - val_acc: 0.7963\n",
      "Epoch 2/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3999 - acc: 0.8121Epoch 00002: val_loss improved from 0.42585 to 0.41236, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3996 - acc: 0.8124 - val_loss: 0.4124 - val_acc: 0.8048\n",
      "Epoch 3/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3863 - acc: 0.8197Epoch 00003: val_loss improved from 0.41236 to 0.40630, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3862 - acc: 0.8195 - val_loss: 0.4063 - val_acc: 0.8093\n",
      "Epoch 4/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3780 - acc: 0.8241Epoch 00004: val_loss improved from 0.40630 to 0.40341, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3777 - acc: 0.8242 - val_loss: 0.4034 - val_acc: 0.8118\n",
      "Epoch 5/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3714 - acc: 0.8282Epoch 00005: val_loss improved from 0.40341 to 0.40283, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3715 - acc: 0.8282 - val_loss: 0.4028 - val_acc: 0.8113\n",
      "Epoch 6/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3661 - acc: 0.8308Epoch 00006: val_loss improved from 0.40283 to 0.39935, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3660 - acc: 0.8309 - val_loss: 0.3994 - val_acc: 0.8141\n",
      "Epoch 7/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3610 - acc: 0.8339Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3607 - acc: 0.8341 - val_loss: 0.4012 - val_acc: 0.8141\n",
      "Epoch 8/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3562 - acc: 0.8363Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3564 - acc: 0.8365 - val_loss: 0.3994 - val_acc: 0.8129\n",
      "Epoch 9/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3520 - acc: 0.8386Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3518 - acc: 0.8387 - val_loss: 0.4004 - val_acc: 0.8129\n",
      "Epoch 00009: early stopping\n",
      "66666/66666 [==============================] - 5s 69us/step\n",
      "Model:  basic_model_adam\n",
      "0.81% (+/- 0.00%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4673 - acc: 0.7674Epoch 00001: val_loss improved from inf to 0.40634, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 69s 517us/step - loss: 0.4672 - acc: 0.7675 - val_loss: 0.4063 - val_acc: 0.8087\n",
      "Epoch 2/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.4055 - acc: 0.8098Epoch 00002: val_loss improved from 0.40634 to 0.39293, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.4054 - acc: 0.8097 - val_loss: 0.3929 - val_acc: 0.8143\n",
      "Epoch 3/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3918 - acc: 0.8180Epoch 00003: val_loss improved from 0.39293 to 0.38902, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3921 - acc: 0.8177 - val_loss: 0.3890 - val_acc: 0.8173\n",
      "Epoch 4/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3840 - acc: 0.8224Epoch 00004: val_loss improved from 0.38902 to 0.38500, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3840 - acc: 0.8224 - val_loss: 0.3850 - val_acc: 0.8203\n",
      "Epoch 5/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3765 - acc: 0.8266Epoch 00005: val_loss improved from 0.38500 to 0.38219, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3765 - acc: 0.8265 - val_loss: 0.3822 - val_acc: 0.8206\n",
      "Epoch 6/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3709 - acc: 0.8299Epoch 00006: val_loss improved from 0.38219 to 0.38085, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3707 - acc: 0.8300 - val_loss: 0.3808 - val_acc: 0.8221\n",
      "Epoch 7/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3650 - acc: 0.8326Epoch 00007: val_loss improved from 0.38085 to 0.37968, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3653 - acc: 0.8324 - val_loss: 0.3797 - val_acc: 0.8222\n",
      "Epoch 8/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3604 - acc: 0.8353Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3606 - acc: 0.8351 - val_loss: 0.3802 - val_acc: 0.8221\n",
      "Epoch 9/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3566 - acc: 0.8377Epoch 00009: val_loss improved from 0.37968 to 0.37901, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3567 - acc: 0.8376 - val_loss: 0.3790 - val_acc: 0.8227\n",
      "Epoch 10/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3530 - acc: 0.8398Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3531 - acc: 0.8398 - val_loss: 0.3829 - val_acc: 0.8209\n",
      "Epoch 11/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3495 - acc: 0.8410Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3497 - acc: 0.8409 - val_loss: 0.3810 - val_acc: 0.8226\n",
      "Epoch 12/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3459 - acc: 0.8436Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3459 - acc: 0.8436 - val_loss: 0.3819 - val_acc: 0.8219\n",
      "Epoch 00012: early stopping\n",
      "66668/66668 [==============================] - 5s 71us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4620 - acc: 0.7688Epoch 00001: val_loss improved from inf to 0.42656, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 8s 59us/step - loss: 0.4618 - acc: 0.7691 - val_loss: 0.4266 - val_acc: 0.7969\n",
      "Epoch 2/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3993 - acc: 0.8125Epoch 00002: val_loss improved from 0.42656 to 0.41160, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3992 - acc: 0.8124 - val_loss: 0.4116 - val_acc: 0.8054\n",
      "Epoch 3/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3866 - acc: 0.8187Epoch 00003: val_loss improved from 0.41160 to 0.40711, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3865 - acc: 0.8187 - val_loss: 0.4071 - val_acc: 0.8075\n",
      "Epoch 4/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3780 - acc: 0.8232Epoch 00004: val_loss improved from 0.40711 to 0.40269, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3780 - acc: 0.8233 - val_loss: 0.4027 - val_acc: 0.8109\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 5/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3711 - acc: 0.8285Epoch 00005: val_loss improved from 0.40269 to 0.39928, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3711 - acc: 0.8285 - val_loss: 0.3993 - val_acc: 0.8116\n",
      "Epoch 6/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3647 - acc: 0.8318Epoch 00006: val_loss improved from 0.39928 to 0.39916, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3651 - acc: 0.8315 - val_loss: 0.3992 - val_acc: 0.8112\n",
      "Epoch 7/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3599 - acc: 0.8358Epoch 00007: val_loss improved from 0.39916 to 0.39696, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3601 - acc: 0.8356 - val_loss: 0.3970 - val_acc: 0.8111\n",
      "Epoch 8/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3550 - acc: 0.8375Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3550 - acc: 0.8376 - val_loss: 0.3972 - val_acc: 0.8126\n",
      "Epoch 9/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3509 - acc: 0.8401Epoch 00009: val_loss improved from 0.39696 to 0.39667, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3508 - acc: 0.8401 - val_loss: 0.3967 - val_acc: 0.8164\n",
      "Epoch 10/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3473 - acc: 0.8423Epoch 00010: val_loss improved from 0.39667 to 0.39648, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3471 - acc: 0.8423 - val_loss: 0.3965 - val_acc: 0.8130\n",
      "Epoch 11/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3434 - acc: 0.8447Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3434 - acc: 0.8447 - val_loss: 0.3983 - val_acc: 0.8120\n",
      "Epoch 12/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3404 - acc: 0.8464Epoch 00012: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3405 - acc: 0.8463 - val_loss: 0.3968 - val_acc: 0.8131\n",
      "Epoch 13/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3368 - acc: 0.8483Epoch 00013: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3371 - acc: 0.8481 - val_loss: 0.3971 - val_acc: 0.8167\n",
      "Epoch 00013: early stopping\n",
      "66666/66666 [==============================] - 5s 79us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4625 - acc: 0.7686Epoch 00001: val_loss improved from inf to 0.42416, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 7s 54us/step - loss: 0.4624 - acc: 0.7686 - val_loss: 0.4242 - val_acc: 0.7972\n",
      "Epoch 2/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3980 - acc: 0.8139Epoch 00002: val_loss improved from 0.42416 to 0.41088, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3976 - acc: 0.8142 - val_loss: 0.4109 - val_acc: 0.8047\n",
      "Epoch 3/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3839 - acc: 0.8213Epoch 00003: val_loss improved from 0.41088 to 0.40411, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3840 - acc: 0.8212 - val_loss: 0.4041 - val_acc: 0.8104\n",
      "Epoch 4/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3759 - acc: 0.8255Epoch 00004: val_loss improved from 0.40411 to 0.40181, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3755 - acc: 0.8257 - val_loss: 0.4018 - val_acc: 0.8128\n",
      "Epoch 5/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3690 - acc: 0.8297Epoch 00005: val_loss improved from 0.40181 to 0.40050, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3691 - acc: 0.8297 - val_loss: 0.4005 - val_acc: 0.8128\n",
      "Epoch 6/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3638 - acc: 0.8325Epoch 00006: val_loss improved from 0.40050 to 0.39768, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3635 - acc: 0.8327 - val_loss: 0.3977 - val_acc: 0.8143\n",
      "Epoch 7/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3584 - acc: 0.8355Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3581 - acc: 0.8356 - val_loss: 0.3987 - val_acc: 0.8159\n",
      "Epoch 8/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3536 - acc: 0.8378Epoch 00008: val_loss improved from 0.39768 to 0.39722, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 19us/step - loss: 0.3537 - acc: 0.8377 - val_loss: 0.3972 - val_acc: 0.8150\n",
      "Epoch 9/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3490 - acc: 0.8406Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 3s 19us/step - loss: 0.3491 - acc: 0.8406 - val_loss: 0.3982 - val_acc: 0.8152\n",
      "Epoch 10/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3454 - acc: 0.8423- ETA: 0s - loss: 0.3451 - acc: 0.842Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 17us/step - loss: 0.3454 - acc: 0.8423 - val_loss: 0.3986 - val_acc: 0.8147\n",
      "Epoch 11/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3415 - acc: 0.8444Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 17us/step - loss: 0.3417 - acc: 0.8444 - val_loss: 0.4005 - val_acc: 0.8134\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 5s 71us/step\n",
      "Model:  basic_model_adam\n",
      "0.82% (+/- 0.00%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.4813 - acc: 0.7563Epoch 00001: val_loss improved from inf to 0.42393, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 27s 199us/step - loss: 0.4807 - acc: 0.7567 - val_loss: 0.4239 - val_acc: 0.7963\n",
      "Epoch 2/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4229 - acc: 0.7992Epoch 00002: val_loss improved from 0.42393 to 0.41233, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.4230 - acc: 0.7992 - val_loss: 0.4123 - val_acc: 0.8022\n",
      "Epoch 3/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4099 - acc: 0.8068Epoch 00003: val_loss improved from 0.41233 to 0.40818, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.4099 - acc: 0.8067 - val_loss: 0.4082 - val_acc: 0.8042\n",
      "Epoch 4/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.4016 - acc: 0.8117Epoch 00004: val_loss improved from 0.40818 to 0.40508, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.4016 - acc: 0.8116 - val_loss: 0.4051 - val_acc: 0.8088\n",
      "Epoch 5/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3945 - acc: 0.8157Epoch 00005: val_loss improved from 0.40508 to 0.40186, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3945 - acc: 0.8156 - val_loss: 0.4019 - val_acc: 0.8107\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 6/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3887 - acc: 0.8192Epoch 00006: val_loss improved from 0.40186 to 0.40151, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3887 - acc: 0.8192 - val_loss: 0.4015 - val_acc: 0.8087\n",
      "Epoch 7/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3828 - acc: 0.8222Epoch 00007: val_loss improved from 0.40151 to 0.40019, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3831 - acc: 0.8220 - val_loss: 0.4002 - val_acc: 0.8104\n",
      "Epoch 8/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3790 - acc: 0.8247Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3790 - acc: 0.8246 - val_loss: 0.4010 - val_acc: 0.8115\n",
      "Epoch 9/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3744 - acc: 0.8276Epoch 00009: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 16us/step - loss: 0.3745 - acc: 0.8275 - val_loss: 0.4005 - val_acc: 0.8109\n",
      "Epoch 10/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3713 - acc: 0.8293Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3713 - acc: 0.8293 - val_loss: 0.4085 - val_acc: 0.8060\n",
      "Epoch 00010: early stopping\n",
      "66668/66668 [==============================] - 5s 71us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4749 - acc: 0.7594Epoch 00001: val_loss improved from inf to 0.44676, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 7s 50us/step - loss: 0.4747 - acc: 0.7595 - val_loss: 0.4468 - val_acc: 0.7844\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4164 - acc: 0.8029Epoch 00002: val_loss improved from 0.44676 to 0.43363, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.4164 - acc: 0.8029 - val_loss: 0.4336 - val_acc: 0.7931\n",
      "Epoch 3/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.4039 - acc: 0.8097Epoch 00003: val_loss improved from 0.43363 to 0.43159, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.4037 - acc: 0.8097 - val_loss: 0.4316 - val_acc: 0.7937\n",
      "Epoch 4/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3956 - acc: 0.8149Epoch 00004: val_loss improved from 0.43159 to 0.42918, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 17us/step - loss: 0.3954 - acc: 0.8149 - val_loss: 0.4292 - val_acc: 0.7984\n",
      "Epoch 5/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3885 - acc: 0.8198Epoch 00005: val_loss improved from 0.42918 to 0.42663, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3883 - acc: 0.8199 - val_loss: 0.4266 - val_acc: 0.7982\n",
      "Epoch 6/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3821 - acc: 0.8228Epoch 00006: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3824 - acc: 0.8228 - val_loss: 0.4267 - val_acc: 0.7987\n",
      "Epoch 7/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3776 - acc: 0.8257Epoch 00007: val_loss improved from 0.42663 to 0.42493, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3776 - acc: 0.8255 - val_loss: 0.4249 - val_acc: 0.7995\n",
      "Epoch 8/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3730 - acc: 0.8278Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3729 - acc: 0.8280 - val_loss: 0.4268 - val_acc: 0.7979\n",
      "Epoch 9/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3692 - acc: 0.8302Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3689 - acc: 0.8304 - val_loss: 0.4251 - val_acc: 0.7986\n",
      "Epoch 10/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3652 - acc: 0.8320Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 12us/step - loss: 0.3654 - acc: 0.8320 - val_loss: 0.4276 - val_acc: 0.7972\n",
      "Epoch 00010: early stopping\n",
      "66666/66666 [==============================] - 5s 73us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4759 - acc: 0.7583Epoch 00001: val_loss improved from inf to 0.44067, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 7s 54us/step - loss: 0.4756 - acc: 0.7585 - val_loss: 0.4407 - val_acc: 0.7882\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4157 - acc: 0.8030Epoch 00002: val_loss improved from 0.44067 to 0.42841, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.4157 - acc: 0.8030 - val_loss: 0.4284 - val_acc: 0.7954\n",
      "Epoch 3/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.4024 - acc: 0.8098Epoch 00003: val_loss improved from 0.42841 to 0.42200, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.4024 - acc: 0.8096 - val_loss: 0.4220 - val_acc: 0.7989\n",
      "Epoch 4/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3941 - acc: 0.8146Epoch 00004: val_loss improved from 0.42200 to 0.41900, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3939 - acc: 0.8147 - val_loss: 0.4190 - val_acc: 0.8017\n",
      "Epoch 5/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3872 - acc: 0.8192Epoch 00005: val_loss improved from 0.41900 to 0.41827, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3872 - acc: 0.8192 - val_loss: 0.4183 - val_acc: 0.8009\n",
      "Epoch 6/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3816 - acc: 0.8219Epoch 00006: val_loss improved from 0.41827 to 0.41560, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3814 - acc: 0.8220 - val_loss: 0.4156 - val_acc: 0.8024\n",
      "Epoch 7/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3766 - acc: 0.8258Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3762 - acc: 0.8260 - val_loss: 0.4171 - val_acc: 0.8029\n",
      "Epoch 8/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3716 - acc: 0.8274Epoch 00008: val_loss improved from 0.41560 to 0.41510, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3712 - acc: 0.8275 - val_loss: 0.4151 - val_acc: 0.8034\n",
      "Epoch 9/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3666 - acc: 0.8298Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3666 - acc: 0.8299 - val_loss: 0.4154 - val_acc: 0.8042\n",
      "Epoch 10/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3632 - acc: 0.8326Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3633 - acc: 0.8325 - val_loss: 0.4170 - val_acc: 0.8029\n",
      "Epoch 11/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3594 - acc: 0.8342Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3595 - acc: 0.8342 - val_loss: 0.4173 - val_acc: 0.8031\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 5s 73us/step\n",
      "Model:  basic_model_adam\n",
      "0.80% (+/- 0.00%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.4718 - acc: 0.7642Epoch 00001: val_loss improved from inf to 0.41075, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 68s 512us/step - loss: 0.4711 - acc: 0.7646 - val_loss: 0.4107 - val_acc: 0.8044\n",
      "Epoch 2/100\n",
      "128000/133332 [===========================>..] - ETA: 0s - loss: 0.4098 - acc: 0.8068Epoch 00002: val_loss improved from 0.41075 to 0.39729, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.4094 - acc: 0.8069 - val_loss: 0.3973 - val_acc: 0.8120\n",
      "Epoch 3/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3957 - acc: 0.8155Epoch 00003: val_loss improved from 0.39729 to 0.39313, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3959 - acc: 0.8154 - val_loss: 0.3931 - val_acc: 0.8155\n",
      "Epoch 4/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3878 - acc: 0.8203Epoch 00004: val_loss improved from 0.39313 to 0.38929, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3877 - acc: 0.8203 - val_loss: 0.3893 - val_acc: 0.8178\n",
      "Epoch 5/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3802 - acc: 0.8247Epoch 00005: val_loss improved from 0.38929 to 0.38671, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3802 - acc: 0.8246 - val_loss: 0.3867 - val_acc: 0.8181\n",
      "Epoch 6/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3744 - acc: 0.8272Epoch 00006: val_loss improved from 0.38671 to 0.38515, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3743 - acc: 0.8271 - val_loss: 0.3852 - val_acc: 0.8201\n",
      "Epoch 7/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3684 - acc: 0.8310Epoch 00007: val_loss improved from 0.38515 to 0.38430, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3688 - acc: 0.8308 - val_loss: 0.3843 - val_acc: 0.8184\n",
      "Epoch 8/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3640 - acc: 0.8334Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3644 - acc: 0.8331 - val_loss: 0.3854 - val_acc: 0.8187\n",
      "Epoch 9/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3601 - acc: 0.8354Epoch 00009: val_loss improved from 0.38430 to 0.38381, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3602 - acc: 0.8351 - val_loss: 0.3838 - val_acc: 0.8191\n",
      "Epoch 10/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3568 - acc: 0.8368Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3570 - acc: 0.8367 - val_loss: 0.3890 - val_acc: 0.8165\n",
      "Epoch 11/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3532 - acc: 0.8397Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3533 - acc: 0.8397 - val_loss: 0.3870 - val_acc: 0.8188\n",
      "Epoch 12/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3495 - acc: 0.8408Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3495 - acc: 0.8409 - val_loss: 0.3871 - val_acc: 0.8195\n",
      "Epoch 00012: early stopping\n",
      "66668/66668 [==============================] - 6s 83us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4672 - acc: 0.7654- ETA: 6s - loss: 0.512Epoch 00001: val_loss improved from inf to 0.43027, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 9s 69us/step - loss: 0.4657 - acc: 0.7664 - val_loss: 0.4303 - val_acc: 0.7935\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4030 - acc: 0.8100Epoch 00002: val_loss improved from 0.43027 to 0.41566, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.4030 - acc: 0.8100 - val_loss: 0.4157 - val_acc: 0.8025\n",
      "Epoch 3/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3904 - acc: 0.8166Epoch 00003: val_loss improved from 0.41566 to 0.41165, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3903 - acc: 0.8165 - val_loss: 0.4116 - val_acc: 0.8033\n",
      "Epoch 4/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3819 - acc: 0.8216Epoch 00004: val_loss improved from 0.41165 to 0.40698, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3819 - acc: 0.8215 - val_loss: 0.4070 - val_acc: 0.8076\n",
      "Epoch 5/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3752 - acc: 0.8265Epoch 00005: val_loss improved from 0.40698 to 0.40381, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3750 - acc: 0.8266 - val_loss: 0.4038 - val_acc: 0.8083\n",
      "Epoch 6/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3689 - acc: 0.8296Epoch 00006: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3692 - acc: 0.8293 - val_loss: 0.4040 - val_acc: 0.8089\n",
      "Epoch 7/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3641 - acc: 0.8332Epoch 00007: val_loss improved from 0.40381 to 0.40218, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3641 - acc: 0.8333 - val_loss: 0.4022 - val_acc: 0.8070\n",
      "Epoch 8/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3591 - acc: 0.8352Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3591 - acc: 0.8353 - val_loss: 0.4032 - val_acc: 0.8066\n",
      "Epoch 9/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3554 - acc: 0.8375Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3550 - acc: 0.8376 - val_loss: 0.4038 - val_acc: 0.8071\n",
      "Epoch 10/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3514 - acc: 0.8390Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3514 - acc: 0.8390 - val_loss: 0.4038 - val_acc: 0.8062\n",
      "Epoch 00010: early stopping\n",
      "66666/66666 [==============================] - 6s 92us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4666 - acc: 0.7653Epoch 00001: val_loss improved from inf to 0.42784, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 9s 65us/step - loss: 0.4666 - acc: 0.7653 - val_loss: 0.4278 - val_acc: 0.7950\n",
      "Epoch 2/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4017 - acc: 0.8115Epoch 00002: val_loss improved from 0.42784 to 0.41478, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.4015 - acc: 0.8115 - val_loss: 0.4148 - val_acc: 0.8034\n",
      "Epoch 3/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3877 - acc: 0.8195Epoch 00003: val_loss improved from 0.41478 to 0.40826, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3878 - acc: 0.8193 - val_loss: 0.4083 - val_acc: 0.8084\n",
      "Epoch 4/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3798 - acc: 0.8235Epoch 00004: val_loss improved from 0.40826 to 0.40574, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3794 - acc: 0.8237 - val_loss: 0.4057 - val_acc: 0.8099\n",
      "Epoch 5/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3730 - acc: 0.8271Epoch 00005: val_loss improved from 0.40574 to 0.40571, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3730 - acc: 0.8271 - val_loss: 0.4057 - val_acc: 0.8094\n",
      "Epoch 6/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3677 - acc: 0.8299Epoch 00006: val_loss improved from 0.40571 to 0.40221, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3676 - acc: 0.8301 - val_loss: 0.4022 - val_acc: 0.8121\n",
      "Epoch 7/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3624 - acc: 0.8332Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3622 - acc: 0.8333 - val_loss: 0.4031 - val_acc: 0.8132\n",
      "Epoch 8/100\n",
      "128000/133334 [===========================>..] - ETA: 0s - loss: 0.3578 - acc: 0.8351Epoch 00008: val_loss improved from 0.40221 to 0.40175, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3578 - acc: 0.8352 - val_loss: 0.4018 - val_acc: 0.8107\n",
      "Epoch 9/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3535 - acc: 0.8380Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3533 - acc: 0.8381 - val_loss: 0.4028 - val_acc: 0.8124\n",
      "Epoch 10/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3496 - acc: 0.8402Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3496 - acc: 0.8401 - val_loss: 0.4043 - val_acc: 0.8112\n",
      "Epoch 11/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3462 - acc: 0.8425Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3462 - acc: 0.8424 - val_loss: 0.4046 - val_acc: 0.8111\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 5s 79us/step\n",
      "Model:  basic_model_adam\n",
      "0.81% (+/- 0.01%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4707 - acc: 0.7644Epoch 00001: val_loss improved from inf to 0.41144, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 94s 706us/step - loss: 0.4706 - acc: 0.7645 - val_loss: 0.4114 - val_acc: 0.8024\n",
      "Epoch 2/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.4087 - acc: 0.8067Epoch 00002: val_loss improved from 0.41144 to 0.39714, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 17us/step - loss: 0.4085 - acc: 0.8067 - val_loss: 0.3971 - val_acc: 0.8098\n",
      "Epoch 3/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3943 - acc: 0.8155Epoch 00003: val_loss improved from 0.39714 to 0.39180, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 18us/step - loss: 0.3945 - acc: 0.8154 - val_loss: 0.3918 - val_acc: 0.8143\n",
      "Epoch 4/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3860 - acc: 0.8212Epoch 00004: val_loss improved from 0.39180 to 0.38774, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 18us/step - loss: 0.3861 - acc: 0.8212 - val_loss: 0.3877 - val_acc: 0.8190\n",
      "Epoch 5/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3786 - acc: 0.8245Epoch 00005: val_loss improved from 0.38774 to 0.38526, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 3s 20us/step - loss: 0.3785 - acc: 0.8243 - val_loss: 0.3853 - val_acc: 0.8196\n",
      "Epoch 6/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3726 - acc: 0.8278Epoch 00006: val_loss improved from 0.38526 to 0.38384, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 16us/step - loss: 0.3725 - acc: 0.8279 - val_loss: 0.3838 - val_acc: 0.8209\n",
      "Epoch 7/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3668 - acc: 0.8312Epoch 00007: val_loss improved from 0.38384 to 0.38321, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3671 - acc: 0.8310 - val_loss: 0.3832 - val_acc: 0.8210\n",
      "Epoch 8/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3621 - acc: 0.8342Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3623 - acc: 0.8340 - val_loss: 0.3843 - val_acc: 0.8190\n",
      "Epoch 9/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3582 - acc: 0.8354Epoch 00009: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3583 - acc: 0.8354 - val_loss: 0.3833 - val_acc: 0.8203\n",
      "Epoch 10/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3544 - acc: 0.8383Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3543 - acc: 0.8383 - val_loss: 0.3881 - val_acc: 0.8185\n",
      "Epoch 00010: early stopping\n",
      "66668/66668 [==============================] - 6s 91us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.4668 - acc: 0.7645Epoch 00001: val_loss improved from inf to 0.42782, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 9s 64us/step - loss: 0.4657 - acc: 0.7654 - val_loss: 0.4278 - val_acc: 0.7925\n",
      "Epoch 2/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4026 - acc: 0.8099Epoch 00002: val_loss improved from 0.42782 to 0.41325, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.4025 - acc: 0.8099 - val_loss: 0.4132 - val_acc: 0.8028\n",
      "Epoch 3/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3892 - acc: 0.8169Epoch 00003: val_loss improved from 0.41325 to 0.40866, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3892 - acc: 0.8168 - val_loss: 0.4087 - val_acc: 0.8044\n",
      "Epoch 4/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3807 - acc: 0.8224Epoch 00004: val_loss improved from 0.40866 to 0.40346, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3807 - acc: 0.8223 - val_loss: 0.4035 - val_acc: 0.8076\n",
      "Epoch 5/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3739 - acc: 0.8268Epoch 00005: val_loss improved from 0.40346 to 0.40124, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3737 - acc: 0.8269 - val_loss: 0.4012 - val_acc: 0.8090\n",
      "Epoch 6/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3679 - acc: 0.8297Epoch 00006: val_loss improved from 0.40124 to 0.40060, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3680 - acc: 0.8296 - val_loss: 0.4006 - val_acc: 0.8084\n",
      "Epoch 7/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3629 - acc: 0.8328Epoch 00007: val_loss improved from 0.40060 to 0.39949, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3627 - acc: 0.8327 - val_loss: 0.3995 - val_acc: 0.8131\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 8/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3580 - acc: 0.8358Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3580 - acc: 0.8358 - val_loss: 0.3999 - val_acc: 0.8086\n",
      "Epoch 9/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3540 - acc: 0.8383Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3539 - acc: 0.8384 - val_loss: 0.3996 - val_acc: 0.8147\n",
      "Epoch 10/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3500 - acc: 0.8401Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3501 - acc: 0.8400 - val_loss: 0.4001 - val_acc: 0.8147\n",
      "Epoch 00010: early stopping\n",
      "66666/66666 [==============================] - 7s 106us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4674 - acc: 0.7646Epoch 00001: val_loss improved from inf to 0.42787, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 10s 74us/step - loss: 0.4656 - acc: 0.7659 - val_loss: 0.4279 - val_acc: 0.7946\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4004 - acc: 0.8111Epoch 00002: val_loss improved from 0.42787 to 0.41421, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 3s 23us/step - loss: 0.4005 - acc: 0.8110 - val_loss: 0.4142 - val_acc: 0.8017\n",
      "Epoch 3/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3864 - acc: 0.8191Epoch 00003: val_loss improved from 0.41421 to 0.40747, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 3s 20us/step - loss: 0.3865 - acc: 0.8190 - val_loss: 0.4075 - val_acc: 0.8078\n",
      "Epoch 4/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3782 - acc: 0.8235- ETA: 0s - loss: 0.3808 - Epoch 00004: val_loss improved from 0.40747 to 0.40484, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 3s 22us/step - loss: 0.3780 - acc: 0.8236 - val_loss: 0.4048 - val_acc: 0.8095\n",
      "Epoch 5/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3715 - acc: 0.8271Epoch 00005: val_loss improved from 0.40484 to 0.40391, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 3s 21us/step - loss: 0.3715 - acc: 0.8272 - val_loss: 0.4039 - val_acc: 0.8104\n",
      "Epoch 6/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3659 - acc: 0.8306Epoch 00006: val_loss improved from 0.40391 to 0.40095, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 18us/step - loss: 0.3658 - acc: 0.8308 - val_loss: 0.4010 - val_acc: 0.8131\n",
      "Epoch 7/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3602 - acc: 0.8334Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 3s 23us/step - loss: 0.3602 - acc: 0.8335 - val_loss: 0.4019 - val_acc: 0.8127\n",
      "Epoch 8/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3556 - acc: 0.8360Epoch 00008: val_loss improved from 0.40095 to 0.40004, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 18us/step - loss: 0.3558 - acc: 0.8360 - val_loss: 0.4000 - val_acc: 0.8123\n",
      "Epoch 9/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3510 - acc: 0.8385Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3511 - acc: 0.8384 - val_loss: 0.4012 - val_acc: 0.8133\n",
      "Epoch 10/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3471 - acc: 0.8403Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3472 - acc: 0.8402 - val_loss: 0.4027 - val_acc: 0.8131\n",
      "Epoch 11/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3437 - acc: 0.8423Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3438 - acc: 0.8422 - val_loss: 0.4028 - val_acc: 0.8115\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 6s 84us/step\n",
      "Model:  basic_model_adam\n",
      "0.82% (+/- 0.00%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.4734 - acc: 0.7616Epoch 00001: val_loss improved from inf to 0.41161, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 35s 261us/step - loss: 0.4721 - acc: 0.7625 - val_loss: 0.4116 - val_acc: 0.8037\n",
      "Epoch 2/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.4095 - acc: 0.8053Epoch 00002: val_loss improved from 0.41161 to 0.39745, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 16us/step - loss: 0.4095 - acc: 0.8053 - val_loss: 0.3974 - val_acc: 0.8109\n",
      "Epoch 3/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3952 - acc: 0.8142Epoch 00003: val_loss improved from 0.39745 to 0.39225, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3954 - acc: 0.8141 - val_loss: 0.3922 - val_acc: 0.8143\n",
      "Epoch 4/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3874 - acc: 0.8197Epoch 00004: val_loss improved from 0.39225 to 0.38933, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3874 - acc: 0.8197 - val_loss: 0.3893 - val_acc: 0.8181\n",
      "Epoch 5/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3800 - acc: 0.8235Epoch 00005: val_loss improved from 0.38933 to 0.38664, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3799 - acc: 0.8233 - val_loss: 0.3866 - val_acc: 0.8175\n",
      "Epoch 6/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3741 - acc: 0.8266Epoch 00006: val_loss improved from 0.38664 to 0.38522, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3740 - acc: 0.8266 - val_loss: 0.3852 - val_acc: 0.8170\n",
      "Epoch 7/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3684 - acc: 0.8298Epoch 00007: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3689 - acc: 0.8296 - val_loss: 0.3853 - val_acc: 0.8184\n",
      "Epoch 8/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3642 - acc: 0.8321Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3644 - acc: 0.8319 - val_loss: 0.3860 - val_acc: 0.8172\n",
      "Epoch 9/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3606 - acc: 0.8341Epoch 00009: val_loss improved from 0.38522 to 0.38448, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3607 - acc: 0.8341 - val_loss: 0.3845 - val_acc: 0.8180\n",
      "Epoch 10/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3566 - acc: 0.8375Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3565 - acc: 0.8376 - val_loss: 0.3893 - val_acc: 0.8172\n",
      "Epoch 11/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3527 - acc: 0.8394Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3532 - acc: 0.8390 - val_loss: 0.3868 - val_acc: 0.8185\n",
      "Epoch 12/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3499 - acc: 0.8413Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3499 - acc: 0.8413 - val_loss: 0.3873 - val_acc: 0.8179\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 00012: early stopping\n",
      "66668/66668 [==============================] - 6s 90us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4670 - acc: 0.7640Epoch 00001: val_loss improved from inf to 0.42878, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 8s 61us/step - loss: 0.4668 - acc: 0.7641 - val_loss: 0.4288 - val_acc: 0.7908\n",
      "Epoch 2/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.4034 - acc: 0.8092Epoch 00002: val_loss improved from 0.42878 to 0.41367, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.4033 - acc: 0.8091 - val_loss: 0.4137 - val_acc: 0.8007\n",
      "Epoch 3/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3902 - acc: 0.8161Epoch 00003: val_loss improved from 0.41367 to 0.40906, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3902 - acc: 0.8160 - val_loss: 0.4091 - val_acc: 0.8024\n",
      "Epoch 4/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3819 - acc: 0.8215Epoch 00004: val_loss improved from 0.40906 to 0.40398, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3819 - acc: 0.8215 - val_loss: 0.4040 - val_acc: 0.8063\n",
      "Epoch 5/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3751 - acc: 0.8257Epoch 00005: val_loss improved from 0.40398 to 0.40216, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3751 - acc: 0.8257 - val_loss: 0.4022 - val_acc: 0.8075\n",
      "Epoch 6/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3694 - acc: 0.8284Epoch 00006: val_loss improved from 0.40216 to 0.40144, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3694 - acc: 0.8284 - val_loss: 0.4014 - val_acc: 0.8068\n",
      "Epoch 7/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3642 - acc: 0.8321Epoch 00007: val_loss improved from 0.40144 to 0.40031, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3643 - acc: 0.8319 - val_loss: 0.4003 - val_acc: 0.8126\n",
      "Epoch 8/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3596 - acc: 0.8345Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3597 - acc: 0.8346 - val_loss: 0.4016 - val_acc: 0.8075\n",
      "Epoch 9/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3559 - acc: 0.8366Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3557 - acc: 0.8368 - val_loss: 0.4011 - val_acc: 0.8129\n",
      "Epoch 10/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3521 - acc: 0.8384Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3521 - acc: 0.8383 - val_loss: 0.4013 - val_acc: 0.8085\n",
      "Epoch 00010: early stopping\n",
      "66666/66666 [==============================] - 6s 84us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4669 - acc: 0.7643Epoch 00001: val_loss improved from inf to 0.42928, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 8s 62us/step - loss: 0.4665 - acc: 0.7645 - val_loss: 0.4293 - val_acc: 0.7930\n",
      "Epoch 2/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4016 - acc: 0.8097Epoch 00002: val_loss improved from 0.42928 to 0.41438, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.4014 - acc: 0.8098 - val_loss: 0.4144 - val_acc: 0.8007\n",
      "Epoch 3/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3874 - acc: 0.8179Epoch 00003: val_loss improved from 0.41438 to 0.40771, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 3s 19us/step - loss: 0.3875 - acc: 0.8177 - val_loss: 0.4077 - val_acc: 0.8070\n",
      "Epoch 4/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3796 - acc: 0.8219Epoch 00004: val_loss improved from 0.40771 to 0.40518, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3793 - acc: 0.8220 - val_loss: 0.4052 - val_acc: 0.8092\n",
      "Epoch 5/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3731 - acc: 0.8252Epoch 00005: val_loss improved from 0.40518 to 0.40376, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3731 - acc: 0.8253 - val_loss: 0.4038 - val_acc: 0.8098\n",
      "Epoch 6/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3675 - acc: 0.8294Epoch 00006: val_loss improved from 0.40376 to 0.40151, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3676 - acc: 0.8294 - val_loss: 0.4015 - val_acc: 0.8115\n",
      "Epoch 7/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3625 - acc: 0.8321Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3622 - acc: 0.8322 - val_loss: 0.4021 - val_acc: 0.8128\n",
      "Epoch 8/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3580 - acc: 0.8348Epoch 00008: val_loss improved from 0.40151 to 0.40066, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3580 - acc: 0.8349 - val_loss: 0.4007 - val_acc: 0.8113\n",
      "Epoch 9/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3538 - acc: 0.8365Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3536 - acc: 0.8367 - val_loss: 0.4016 - val_acc: 0.8114\n",
      "Epoch 10/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3496 - acc: 0.8388Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3499 - acc: 0.8386 - val_loss: 0.4039 - val_acc: 0.8109\n",
      "Epoch 11/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3470 - acc: 0.8408Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3470 - acc: 0.8408 - val_loss: 0.4042 - val_acc: 0.8111\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 6s 88us/step\n",
      "Model:  basic_model_adam\n",
      "0.81% (+/- 0.00%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "130048/133332 [============================>.] - ETA: 2s - loss: 0.4782 - acc: 0.7564Epoch 00001: val_loss improved from inf to 0.41757, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 86s 647us/step - loss: 0.4767 - acc: 0.7576 - val_loss: 0.4176 - val_acc: 0.8004\n",
      "Epoch 2/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.4157 - acc: 0.8006Epoch 00002: val_loss improved from 0.41757 to 0.40443, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 17us/step - loss: 0.4157 - acc: 0.8006 - val_loss: 0.4044 - val_acc: 0.8078\n",
      "Epoch 3/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4024 - acc: 0.8086Epoch 00003: val_loss improved from 0.40443 to 0.39805, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.4024 - acc: 0.8086 - val_loss: 0.3980 - val_acc: 0.8108\n",
      "Epoch 4/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3953 - acc: 0.8130Epoch 00004: val_loss improved from 0.39805 to 0.39484, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3953 - acc: 0.8129 - val_loss: 0.3948 - val_acc: 0.8144\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 5/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3882 - acc: 0.8170Epoch 00005: val_loss improved from 0.39484 to 0.39418, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3881 - acc: 0.8169 - val_loss: 0.3942 - val_acc: 0.8139\n",
      "Epoch 6/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3823 - acc: 0.8198Epoch 00006: val_loss improved from 0.39418 to 0.39089, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3822 - acc: 0.8197 - val_loss: 0.3909 - val_acc: 0.8149\n",
      "Epoch 7/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3767 - acc: 0.8237Epoch 00007: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3773 - acc: 0.8232 - val_loss: 0.3909 - val_acc: 0.8153\n",
      "Epoch 8/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3730 - acc: 0.8260Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3731 - acc: 0.8258 - val_loss: 0.3940 - val_acc: 0.8016\n",
      "Epoch 9/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3694 - acc: 0.8276Epoch 00009: val_loss improved from 0.39089 to 0.39018, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 18us/step - loss: 0.3697 - acc: 0.8274 - val_loss: 0.3902 - val_acc: 0.8147\n",
      "Epoch 10/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3653 - acc: 0.8303Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3655 - acc: 0.8302 - val_loss: 0.3942 - val_acc: 0.8146\n",
      "Epoch 11/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3621 - acc: 0.8322Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3626 - acc: 0.8321 - val_loss: 0.3928 - val_acc: 0.8143\n",
      "Epoch 12/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3592 - acc: 0.8341Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 19us/step - loss: 0.3592 - acc: 0.8341 - val_loss: 0.3950 - val_acc: 0.8032\n",
      "Epoch 00012: early stopping\n",
      "66668/66668 [==============================] - 6s 91us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4727 - acc: 0.7593Epoch 00001: val_loss improved from inf to 0.43437, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 10s 71us/step - loss: 0.4713 - acc: 0.7602 - val_loss: 0.4344 - val_acc: 0.7873\n",
      "Epoch 2/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.4098 - acc: 0.8038Epoch 00002: val_loss improved from 0.43437 to 0.41920, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.4097 - acc: 0.8039 - val_loss: 0.4192 - val_acc: 0.7971\n",
      "Epoch 3/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3972 - acc: 0.8109Epoch 00003: val_loss improved from 0.41920 to 0.41567, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3972 - acc: 0.8110 - val_loss: 0.4157 - val_acc: 0.7994\n",
      "Epoch 4/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3895 - acc: 0.8159Epoch 00004: val_loss improved from 0.41567 to 0.41030, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3896 - acc: 0.8157 - val_loss: 0.4103 - val_acc: 0.8031\n",
      "Epoch 5/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3827 - acc: 0.8205Epoch 00005: val_loss improved from 0.41030 to 0.40916, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3828 - acc: 0.8205 - val_loss: 0.4092 - val_acc: 0.8033\n",
      "Epoch 6/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3769 - acc: 0.8231Epoch 00006: val_loss improved from 0.40916 to 0.40849, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3774 - acc: 0.8228 - val_loss: 0.4085 - val_acc: 0.8039\n",
      "Epoch 7/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3726 - acc: 0.8257Epoch 00007: val_loss improved from 0.40849 to 0.40727, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3725 - acc: 0.8256 - val_loss: 0.4073 - val_acc: 0.8044\n",
      "Epoch 8/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3681 - acc: 0.8286Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3682 - acc: 0.8286 - val_loss: 0.4079 - val_acc: 0.8053\n",
      "Epoch 9/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3645 - acc: 0.8304Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3644 - acc: 0.8304 - val_loss: 0.4088 - val_acc: 0.8040\n",
      "Epoch 10/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3609 - acc: 0.8324Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3609 - acc: 0.8324 - val_loss: 0.4085 - val_acc: 0.8056\n",
      "Epoch 00010: early stopping\n",
      "66666/66666 [==============================] - 7s 103us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4726 - acc: 0.7608  Epoch 00001: val_loss improved from inf to 0.43364, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 11s 86us/step - loss: 0.4709 - acc: 0.7621 - val_loss: 0.4336 - val_acc: 0.7862\n",
      "Epoch 2/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.4076 - acc: 0.8061Epoch 00002: val_loss improved from 0.43364 to 0.42044, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.4073 - acc: 0.8064 - val_loss: 0.4204 - val_acc: 0.7948\n",
      "Epoch 3/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3939 - acc: 0.8139Epoch 00003: val_loss improved from 0.42044 to 0.41426, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 17us/step - loss: 0.3939 - acc: 0.8137 - val_loss: 0.4143 - val_acc: 0.8000\n",
      "Epoch 4/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3863 - acc: 0.8183Epoch 00004: val_loss improved from 0.41426 to 0.41123, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 17us/step - loss: 0.3860 - acc: 0.8184 - val_loss: 0.4112 - val_acc: 0.8028\n",
      "Epoch 5/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3799 - acc: 0.8216Epoch 00005: val_loss improved from 0.41123 to 0.40850, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3799 - acc: 0.8217 - val_loss: 0.4085 - val_acc: 0.8041\n",
      "Epoch 6/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3746 - acc: 0.8250Epoch 00006: val_loss improved from 0.40850 to 0.40784, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 17us/step - loss: 0.3745 - acc: 0.8250 - val_loss: 0.4078 - val_acc: 0.8042\n",
      "Epoch 7/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3696 - acc: 0.8280Epoch 00007: val_loss improved from 0.40784 to 0.40769, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3696 - acc: 0.8280 - val_loss: 0.4077 - val_acc: 0.8064\n",
      "Epoch 8/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3654 - acc: 0.8304Epoch 00008: val_loss improved from 0.40769 to 0.40587, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3654 - acc: 0.8305 - val_loss: 0.4059 - val_acc: 0.8047\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 9/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3609 - acc: 0.8324Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3609 - acc: 0.8324 - val_loss: 0.4085 - val_acc: 0.8042\n",
      "Epoch 10/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3574 - acc: 0.8341Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3575 - acc: 0.8341 - val_loss: 0.4104 - val_acc: 0.8048\n",
      "Epoch 11/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3545 - acc: 0.8368Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3545 - acc: 0.8368 - val_loss: 0.4080 - val_acc: 0.8056\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 6s 88us/step\n",
      "Model:  basic_model_adam\n",
      "0.81% (+/- 0.00%)\n"
     ]
    }
   ],
   "source": [
    "accuracies=[]\n",
    "stds=[]\n",
    "\n",
    "for corpus in corpuses: \n",
    "    model_score=VP.classify_with_neural_networks(neural_nets, global_vectors, corpus, total_training_tweets, nr_pos_tweets, epochs=100, n_folds=3)\n",
    "    accuracies.append(model_score[0][0])\n",
    "    stds.append(model_score[0][1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Want to determine which preprocessing techniques that improved the accuracy, and keep them: "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The original corpus gave accuracy of:  0.812709927664 std: 0.00568187806062 \n",
      "\n",
      "IMPROVED:   SH_corpus , score: 0.816064928216 std: 0.00544780821751\n",
      "IMPROVED:   SHM_corpus , score: 0.816469933014 std: 0.00484066138544\n",
      "IMPROVED:   H_corpus , score: 0.814829949765 std: 0.0035532936432\n",
      "IMPROVED:   HK_corpus , score: 0.815369944064 std: 0.00395651181696\n",
      "Not better: PS_corpus , score: 0.812649932164 std: 0.00539160668946\n",
      "IMPROVED:   NS__corpus , score: 0.812724931564 std: 0.00525848291263\n",
      "IMPROVED:   OS_corpus , score: 0.812899931364 std: 0.00517227899144\n",
      "IMPROVED:   N_corpus , score: 0.812709927664 std: 0.00568187806062\n",
      "IMPROVED:   NM_corpus , score: 0.812879923665 std: 0.00583995960365\n",
      "Not better: ST_corpus , score: 0.804854960261 std: 0.00282527802715\n",
      "IMPROVED:   SP_corpus , score: 0.814269927511 std: 0.00545526475224\n",
      "IMPROVED:   E_corpus , score: 0.814399937815 std: 0.00497216733352\n",
      "IMPROVED:   SN_corpus , score: 0.816884942113 std: 0.00417103355424\n",
      "Not better: RS_corpus , score: 0.80443494016 std: 0.00452520062037\n",
      "Not better: EX_corpus , score: 0.812254931812 std: 0.00504595145149\n",
      "IMPROVED:   N-2_corpus , score: 0.815464945014 std: 0.00390164541785\n",
      "IMPROVED:   N-3_corpus , score: 0.813984959613 std: 0.00290504921046\n",
      "Not better: N-4_corpus , score: 0.807929932065 std: 0.00480581203173\n"
     ]
    }
   ],
   "source": [
    "corpuses_1=[]\n",
    "names_1=[]\n",
    "stds_1=[]\n",
    "acc_1=[]\n",
    "print('The original corpus gave accuracy of: ',accuracies[0], 'std:', stds[0],'\\n')\n",
    "for i in range(1,len(accuracies)):\n",
    "    if accuracies[i]>=accuracies[0]:\n",
    "        corpuses_1.append(corpuses[i])\n",
    "        names_1.append(names[i])\n",
    "        stds_1.append(stds[i])\n",
    "        acc_1.append(accuracies[i])\n",
    "        print('IMPROVED:  ',names[i],', score:',accuracies[i],'std:',stds[i])\n",
    "    else:\n",
    "        print('Not better:',names[i],', score:',accuracies[i],'std:',stds[i])\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# We do not keep SH, as SHM is better. \n",
    "# We do not keep N as NM is better \n",
    "# We do not keep SP due to time  (158.26 min)\n",
    "# We do not keep techniques that imporved less than 0.05 percentage points. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Apply all techniques that contributed positively"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "input_={'hashtag': True, 'segmentation_hash': True, 'hashtag_mention':True,\n",
    "        'hearts':True,'hugs_and_kisses':True,'elongation':True, 'set_to_not':True}\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reading twitter - 1grams ...\n",
      "Reading twitter - 2grams ...\n",
      "Time in min before starting first for loop: 0.18563808600107828\n",
      "Time in min after 1  tweets: 0.18638985157012938\n",
      "Time in min after 25001  tweets: 0.5039530992507935\n",
      "Time in min after 50001  tweets: 0.7553902665774027\n",
      "Time in min after 75001  tweets: 0.9880523641904195\n",
      "Time in min after 100001  tweets: 1.2328885157903036\n",
      "Time in min after 125001  tweets: 1.4751866857210796\n",
      "Time in min after 150001  tweets: 1.7036137024561564\n",
      "Time in min after 175001  tweets: 1.9359561999638875\n",
      "Time in min after 200001  tweets: 2.164798633257548\n",
      "Time in min total: 2.2595055023829143\n"
     ]
    }
   ],
   "source": [
    "#Applying the techniques that we have decided to keep:\n",
    "best_prepr_corpus=TO.preprocess_corpus(full_corpus, **input_)\n",
    "best_corpus=TO.creating_n_grams_corpus(2,best_prepr_corpus)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "131072/133332 [============================>.] - ETA: 1s - loss: 0.4641 - acc: 0.7698Epoch 00001: val_loss improved from inf to 0.40136, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 93s 700us/step - loss: 0.4629 - acc: 0.7706 - val_loss: 0.4014 - val_acc: 0.8091\n",
      "Epoch 2/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3990 - acc: 0.8131Epoch 00002: val_loss improved from 0.40136 to 0.38808, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3990 - acc: 0.8130 - val_loss: 0.3881 - val_acc: 0.8164\n",
      "Epoch 3/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3851 - acc: 0.8211Epoch 00003: val_loss improved from 0.38808 to 0.38216, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3851 - acc: 0.8211 - val_loss: 0.3822 - val_acc: 0.8199\n",
      "Epoch 4/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3765 - acc: 0.8259Epoch 00004: val_loss improved from 0.38216 to 0.37871, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3765 - acc: 0.8259 - val_loss: 0.3787 - val_acc: 0.8236\n",
      "Epoch 5/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3689 - acc: 0.8296Epoch 00005: val_loss improved from 0.37871 to 0.37640, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3690 - acc: 0.8296 - val_loss: 0.3764 - val_acc: 0.8240\n",
      "Epoch 6/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3629 - acc: 0.8336Epoch 00006: val_loss improved from 0.37640 to 0.37520, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3629 - acc: 0.8337 - val_loss: 0.3752 - val_acc: 0.8241\n",
      "Epoch 7/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3577 - acc: 0.8367Epoch 00007: val_loss improved from 0.37520 to 0.37419, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3577 - acc: 0.8367 - val_loss: 0.3742 - val_acc: 0.8249\n",
      "Epoch 8/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3529 - acc: 0.8396Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3530 - acc: 0.8394 - val_loss: 0.3747 - val_acc: 0.8260\n",
      "Epoch 9/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3487 - acc: 0.8416Epoch 00009: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3488 - acc: 0.8415 - val_loss: 0.3752 - val_acc: 0.8237\n",
      "Epoch 10/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3448 - acc: 0.8439Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3449 - acc: 0.8439 - val_loss: 0.3780 - val_acc: 0.8240\n",
      "Epoch 00010: early stopping\n",
      "66668/66668 [==============================] - 7s 110us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4603 - acc: 0.7696Epoch 00001: val_loss improved from inf to 0.41874, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 11s 80us/step - loss: 0.4586 - acc: 0.7707 - val_loss: 0.4187 - val_acc: 0.8007\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3932 - acc: 0.8160Epoch 00002: val_loss improved from 0.41874 to 0.40397, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3932 - acc: 0.8159 - val_loss: 0.4040 - val_acc: 0.8094\n",
      "Epoch 3/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3801 - acc: 0.8231Epoch 00003: val_loss improved from 0.40397 to 0.39908, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3801 - acc: 0.8230 - val_loss: 0.3991 - val_acc: 0.8110\n",
      "Epoch 4/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3714 - acc: 0.8283Epoch 00004: val_loss improved from 0.39908 to 0.39486, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3714 - acc: 0.8282 - val_loss: 0.3949 - val_acc: 0.8142\n",
      "Epoch 5/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3649 - acc: 0.8319Epoch 00005: val_loss improved from 0.39486 to 0.39157, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3645 - acc: 0.8322 - val_loss: 0.3916 - val_acc: 0.8134\n",
      "Epoch 6/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3586 - acc: 0.8356Epoch 00006: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3587 - acc: 0.8355 - val_loss: 0.3918 - val_acc: 0.8160\n",
      "Epoch 7/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3535 - acc: 0.8388Epoch 00007: val_loss improved from 0.39157 to 0.38999, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3536 - acc: 0.8387 - val_loss: 0.3900 - val_acc: 0.8149\n",
      "Epoch 8/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3485 - acc: 0.8418Epoch 00008: val_loss improved from 0.38999 to 0.38958, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3486 - acc: 0.8417 - val_loss: 0.3896 - val_acc: 0.8155\n",
      "Epoch 9/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3449 - acc: 0.8436Epoch 00009: val_loss improved from 0.38958 to 0.38955, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3447 - acc: 0.8436 - val_loss: 0.3895 - val_acc: 0.8207\n",
      "Epoch 10/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3411 - acc: 0.8458Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3410 - acc: 0.8457 - val_loss: 0.3897 - val_acc: 0.8198\n",
      "Epoch 11/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3374 - acc: 0.8477Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 3s 20us/step - loss: 0.3376 - acc: 0.8477 - val_loss: 0.3916 - val_acc: 0.8145\n",
      "Epoch 12/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3346 - acc: 0.8493Epoch 00012: val_loss improved from 0.38955 to 0.38850, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3346 - acc: 0.8493 - val_loss: 0.3885 - val_acc: 0.8166\n",
      "Epoch 13/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3311 - acc: 0.8510Epoch 00013: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3311 - acc: 0.8510 - val_loss: 0.3901 - val_acc: 0.8203\n",
      "Epoch 14/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3282 - acc: 0.8534Epoch 00014: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3281 - acc: 0.8535 - val_loss: 0.3905 - val_acc: 0.8201\n",
      "Epoch 15/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3255 - acc: 0.8537Epoch 00015: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3255 - acc: 0.8538 - val_loss: 0.3906 - val_acc: 0.8192\n",
      "Epoch 00015: early stopping\n",
      "66666/66666 [==============================] - 7s 101us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4604 - acc: 0.7691Epoch 00001: val_loss improved from inf to 0.41836, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 10s 73us/step - loss: 0.4584 - acc: 0.7704 - val_loss: 0.4184 - val_acc: 0.8017\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3912 - acc: 0.8169Epoch 00002: val_loss improved from 0.41836 to 0.40491, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3913 - acc: 0.8169 - val_loss: 0.4049 - val_acc: 0.8074\n",
      "Epoch 3/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3770 - acc: 0.8245Epoch 00003: val_loss improved from 0.40491 to 0.39778, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3771 - acc: 0.8245 - val_loss: 0.3978 - val_acc: 0.8140\n",
      "Epoch 4/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3685 - acc: 0.8288Epoch 00004: val_loss improved from 0.39778 to 0.39485, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3684 - acc: 0.8289 - val_loss: 0.3948 - val_acc: 0.8163\n",
      "Epoch 5/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3619 - acc: 0.8332Epoch 00005: val_loss improved from 0.39485 to 0.39428, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3618 - acc: 0.8332 - val_loss: 0.3943 - val_acc: 0.8165\n",
      "Epoch 6/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3560 - acc: 0.8358Epoch 00006: val_loss improved from 0.39428 to 0.39067, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3561 - acc: 0.8359 - val_loss: 0.3907 - val_acc: 0.8202\n",
      "Epoch 7/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3504 - acc: 0.8395Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3504 - acc: 0.8395 - val_loss: 0.3919 - val_acc: 0.8196\n",
      "Epoch 8/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3462 - acc: 0.8419Epoch 00008: val_loss improved from 0.39067 to 0.39044, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3461 - acc: 0.8420 - val_loss: 0.3904 - val_acc: 0.8207\n",
      "Epoch 9/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3412 - acc: 0.8450Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3412 - acc: 0.8450 - val_loss: 0.3916 - val_acc: 0.8194\n",
      "Epoch 10/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3376 - acc: 0.8468Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3376 - acc: 0.8468 - val_loss: 0.3932 - val_acc: 0.8187\n",
      "Epoch 11/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3338 - acc: 0.8488Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3340 - acc: 0.8487 - val_loss: 0.3954 - val_acc: 0.8184\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 6s 96us/step\n",
      "Model:  basic_model_adam\n",
      "0.82% (+/- 0.00%)\n"
     ]
    }
   ],
   "source": [
    "#Cross validating:\n",
    "model_score=VP.classify_with_neural_networks(neural_nets, global_vectors, best_corpus, \n",
    "                                             total_training_tweets, nr_pos_tweets, epochs=100, n_folds=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.820739958766 std: 0.00335801472691 \n",
      "\n"
     ]
    }
   ],
   "source": [
    "print('Accuracy:',  model_score[0][0], 'std:',model_score[0][1],'\\n' )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dynamic stopword list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "131072/133332 [============================>.] - ETA: 1s - loss: 0.4641 - acc: 0.7698Epoch 00001: val_loss improved from inf to 0.40136, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 104s 784us/step - loss: 0.4629 - acc: 0.7706 - val_loss: 0.4014 - val_acc: 0.8091\n",
      "Epoch 2/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3990 - acc: 0.8132Epoch 00002: val_loss improved from 0.40136 to 0.38808, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3990 - acc: 0.8130 - val_loss: 0.3881 - val_acc: 0.8164\n",
      "Epoch 3/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3849 - acc: 0.8212Epoch 00003: val_loss improved from 0.38808 to 0.38216, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3851 - acc: 0.8211 - val_loss: 0.3822 - val_acc: 0.8199\n",
      "Epoch 4/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3765 - acc: 0.8260Epoch 00004: val_loss improved from 0.38216 to 0.37871, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3765 - acc: 0.8259 - val_loss: 0.3787 - val_acc: 0.8236\n",
      "Epoch 5/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3690 - acc: 0.8298Epoch 00005: val_loss improved from 0.37871 to 0.37640, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3690 - acc: 0.8296 - val_loss: 0.3764 - val_acc: 0.8240\n",
      "Epoch 6/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3628 - acc: 0.8337Epoch 00006: val_loss improved from 0.37640 to 0.37520, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3629 - acc: 0.8337 - val_loss: 0.3752 - val_acc: 0.8241\n",
      "Epoch 7/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3570 - acc: 0.8372Epoch 00007: val_loss improved from 0.37520 to 0.37419, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3577 - acc: 0.8367 - val_loss: 0.3742 - val_acc: 0.8249\n",
      "Epoch 8/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3528 - acc: 0.8395Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3530 - acc: 0.8394 - val_loss: 0.3747 - val_acc: 0.8260\n",
      "Epoch 9/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3487 - acc: 0.8415Epoch 00009: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3488 - acc: 0.8415 - val_loss: 0.3752 - val_acc: 0.8237\n",
      "Epoch 10/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3450 - acc: 0.8439Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3449 - acc: 0.8439 - val_loss: 0.3780 - val_acc: 0.8240\n",
      "Epoch 00010: early stopping\n",
      "66668/66668 [==============================] - 8s 120us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4603 - acc: 0.7696Epoch 00001: val_loss improved from inf to 0.41874, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 11s 82us/step - loss: 0.4586 - acc: 0.7707 - val_loss: 0.4187 - val_acc: 0.8007\n",
      "Epoch 2/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3934 - acc: 0.8160Epoch 00002: val_loss improved from 0.41874 to 0.40397, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3932 - acc: 0.8159 - val_loss: 0.4040 - val_acc: 0.8094\n",
      "Epoch 3/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3801 - acc: 0.8230Epoch 00003: val_loss improved from 0.40397 to 0.39908, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3801 - acc: 0.8230 - val_loss: 0.3991 - val_acc: 0.8110\n",
      "Epoch 4/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3715 - acc: 0.8282Epoch 00004: val_loss improved from 0.39908 to 0.39486, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3714 - acc: 0.8282 - val_loss: 0.3949 - val_acc: 0.8142\n",
      "Epoch 5/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3646 - acc: 0.8321Epoch 00005: val_loss improved from 0.39486 to 0.39157, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 17us/step - loss: 0.3645 - acc: 0.8322 - val_loss: 0.3916 - val_acc: 0.8134\n",
      "Epoch 6/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3586 - acc: 0.8356Epoch 00006: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 17us/step - loss: 0.3587 - acc: 0.8355 - val_loss: 0.3918 - val_acc: 0.8160\n",
      "Epoch 7/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3535 - acc: 0.8388Epoch 00007: val_loss improved from 0.39157 to 0.38999, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 3s 19us/step - loss: 0.3536 - acc: 0.8387 - val_loss: 0.3900 - val_acc: 0.8149\n",
      "Epoch 8/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3484 - acc: 0.8419Epoch 00008: val_loss improved from 0.38999 to 0.38958, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 17us/step - loss: 0.3486 - acc: 0.8417 - val_loss: 0.3896 - val_acc: 0.8155\n",
      "Epoch 9/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3451 - acc: 0.8434Epoch 00009: val_loss improved from 0.38958 to 0.38955, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3447 - acc: 0.8436 - val_loss: 0.3895 - val_acc: 0.8207\n",
      "Epoch 10/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3410 - acc: 0.8459Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3410 - acc: 0.8457 - val_loss: 0.3897 - val_acc: 0.8198\n",
      "Epoch 11/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3376 - acc: 0.8477Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3376 - acc: 0.8477 - val_loss: 0.3916 - val_acc: 0.8145\n",
      "Epoch 12/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3345 - acc: 0.8494Epoch 00012: val_loss improved from 0.38955 to 0.38850, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 18us/step - loss: 0.3346 - acc: 0.8493 - val_loss: 0.3885 - val_acc: 0.8166\n",
      "Epoch 13/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3305 - acc: 0.8514Epoch 00013: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3311 - acc: 0.8510 - val_loss: 0.3901 - val_acc: 0.8203\n",
      "Epoch 14/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3282 - acc: 0.8534Epoch 00014: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3281 - acc: 0.8535 - val_loss: 0.3905 - val_acc: 0.8201\n",
      "Epoch 15/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3255 - acc: 0.8537Epoch 00015: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3255 - acc: 0.8538 - val_loss: 0.3906 - val_acc: 0.8192\n",
      "Epoch 00015: early stopping\n",
      "66666/66666 [==============================] - 7s 102us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4585 - acc: 0.7703Epoch 00001: val_loss improved from inf to 0.41836, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 10s 74us/step - loss: 0.4584 - acc: 0.7704 - val_loss: 0.4184 - val_acc: 0.8017\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3912 - acc: 0.8169Epoch 00002: val_loss improved from 0.41836 to 0.40491, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3913 - acc: 0.8169 - val_loss: 0.4049 - val_acc: 0.8074\n",
      "Epoch 3/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3770 - acc: 0.8245Epoch 00003: val_loss improved from 0.40491 to 0.39778, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3771 - acc: 0.8245 - val_loss: 0.3978 - val_acc: 0.8140\n",
      "Epoch 4/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3685 - acc: 0.8288Epoch 00004: val_loss improved from 0.39778 to 0.39485, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3684 - acc: 0.8289 - val_loss: 0.3948 - val_acc: 0.8163\n",
      "Epoch 5/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3619 - acc: 0.8332Epoch 00005: val_loss improved from 0.39485 to 0.39428, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3618 - acc: 0.8332 - val_loss: 0.3943 - val_acc: 0.8165\n",
      "Epoch 6/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3560 - acc: 0.8359Epoch 00006: val_loss improved from 0.39428 to 0.39067, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3561 - acc: 0.8359 - val_loss: 0.3907 - val_acc: 0.8202\n",
      "Epoch 7/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3504 - acc: 0.8395Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3504 - acc: 0.8395 - val_loss: 0.3919 - val_acc: 0.8196\n",
      "Epoch 8/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3462 - acc: 0.8419Epoch 00008: val_loss improved from 0.39067 to 0.39044, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3461 - acc: 0.8420 - val_loss: 0.3904 - val_acc: 0.8207\n",
      "Epoch 9/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3414 - acc: 0.8449Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3412 - acc: 0.8450 - val_loss: 0.3916 - val_acc: 0.8194\n",
      "Epoch 10/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3372 - acc: 0.8470Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3376 - acc: 0.8468 - val_loss: 0.3932 - val_acc: 0.8187\n",
      "Epoch 11/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3340 - acc: 0.8488Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3340 - acc: 0.8487 - val_loss: 0.3954 - val_acc: 0.8184\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 7s 100us/step\n",
      "Model:  basic_model_adam\n",
      "0.82% (+/- 0.00%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "131072/133332 [============================>.] - ETA: 2s - loss: 0.4626 - acc: 0.7694Epoch 00001: val_loss improved from inf to 0.39871, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 122s 913us/step - loss: 0.4614 - acc: 0.7703 - val_loss: 0.3987 - val_acc: 0.8106\n",
      "Epoch 2/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3968 - acc: 0.8141Epoch 00002: val_loss improved from 0.39871 to 0.38524, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3967 - acc: 0.8140 - val_loss: 0.3852 - val_acc: 0.8177\n",
      "Epoch 3/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3829 - acc: 0.8224Epoch 00003: val_loss improved from 0.38524 to 0.37989, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3829 - acc: 0.8222 - val_loss: 0.3799 - val_acc: 0.8217\n",
      "Epoch 4/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3745 - acc: 0.8279Epoch 00004: val_loss improved from 0.37989 to 0.37602, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3746 - acc: 0.8279 - val_loss: 0.3760 - val_acc: 0.8249\n",
      "Epoch 5/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3671 - acc: 0.8308Epoch 00005: val_loss improved from 0.37602 to 0.37363, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3672 - acc: 0.8305 - val_loss: 0.3736 - val_acc: 0.8238\n",
      "Epoch 6/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3611 - acc: 0.8345Epoch 00006: val_loss improved from 0.37363 to 0.37240, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3612 - acc: 0.8345 - val_loss: 0.3724 - val_acc: 0.8275\n",
      "Epoch 7/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3556 - acc: 0.8378Epoch 00007: val_loss improved from 0.37240 to 0.37124, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 16us/step - loss: 0.3561 - acc: 0.8375 - val_loss: 0.3712 - val_acc: 0.8266\n",
      "Epoch 8/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3513 - acc: 0.8400Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3515 - acc: 0.8400 - val_loss: 0.3715 - val_acc: 0.8272\n",
      "Epoch 9/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3477 - acc: 0.8417Epoch 00009: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3476 - acc: 0.8417 - val_loss: 0.3722 - val_acc: 0.8255\n",
      "Epoch 10/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3439 - acc: 0.8442Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 13us/step - loss: 0.3438 - acc: 0.8442 - val_loss: 0.3747 - val_acc: 0.8262\n",
      "Epoch 00010: early stopping\n",
      "66668/66668 [==============================] - 7s 99us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.4579 - acc: 0.7704Epoch 00001: val_loss improved from inf to 0.41539, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 11s 84us/step - loss: 0.4572 - acc: 0.7709 - val_loss: 0.4154 - val_acc: 0.8022\n",
      "Epoch 2/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3910 - acc: 0.8173Epoch 00002: val_loss improved from 0.41539 to 0.40180, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3909 - acc: 0.8173 - val_loss: 0.4018 - val_acc: 0.8114\n",
      "Epoch 3/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3778 - acc: 0.8243Epoch 00003: val_loss improved from 0.40180 to 0.39691, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3778 - acc: 0.8243 - val_loss: 0.3969 - val_acc: 0.8138\n",
      "Epoch 4/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3690 - acc: 0.8294Epoch 00004: val_loss improved from 0.39691 to 0.39298, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3693 - acc: 0.8293 - val_loss: 0.3930 - val_acc: 0.8159\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 5/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3627 - acc: 0.8326Epoch 00005: val_loss improved from 0.39298 to 0.38953, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3624 - acc: 0.8328 - val_loss: 0.3895 - val_acc: 0.8162\n",
      "Epoch 6/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3566 - acc: 0.8359Epoch 00006: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3568 - acc: 0.8359 - val_loss: 0.3900 - val_acc: 0.8162\n",
      "Epoch 7/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3519 - acc: 0.8390Epoch 00007: val_loss improved from 0.38953 to 0.38820, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3519 - acc: 0.8391 - val_loss: 0.3882 - val_acc: 0.8164\n",
      "Epoch 8/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3472 - acc: 0.8419Epoch 00008: val_loss improved from 0.38820 to 0.38745, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3472 - acc: 0.8418 - val_loss: 0.3875 - val_acc: 0.8181\n",
      "Epoch 9/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3434 - acc: 0.8438Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3432 - acc: 0.8438 - val_loss: 0.3879 - val_acc: 0.8209\n",
      "Epoch 10/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3396 - acc: 0.8461Epoch 00010: val_loss improved from 0.38745 to 0.38727, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3397 - acc: 0.8460 - val_loss: 0.3873 - val_acc: 0.8221\n",
      "Epoch 11/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3363 - acc: 0.8479Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3364 - acc: 0.8479 - val_loss: 0.3889 - val_acc: 0.8165\n",
      "Epoch 12/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3335 - acc: 0.8493Epoch 00012: val_loss improved from 0.38727 to 0.38540, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3337 - acc: 0.8492 - val_loss: 0.3854 - val_acc: 0.8197\n",
      "Epoch 13/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3297 - acc: 0.8511Epoch 00013: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3302 - acc: 0.8507 - val_loss: 0.3874 - val_acc: 0.8226\n",
      "Epoch 14/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3275 - acc: 0.8529Epoch 00014: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3272 - acc: 0.8531 - val_loss: 0.3876 - val_acc: 0.8219\n",
      "Epoch 15/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3247 - acc: 0.8544Epoch 00015: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3246 - acc: 0.8544 - val_loss: 0.3877 - val_acc: 0.8215\n",
      "Epoch 00015: early stopping\n",
      "66666/66666 [==============================] - 7s 98us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4572 - acc: 0.7711Epoch 00001: val_loss improved from inf to 0.41608, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 9s 70us/step - loss: 0.4569 - acc: 0.7714 - val_loss: 0.4161 - val_acc: 0.8019\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3886 - acc: 0.8183Epoch 00002: val_loss improved from 0.41608 to 0.40235, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3886 - acc: 0.8183 - val_loss: 0.4024 - val_acc: 0.8099\n",
      "Epoch 3/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3746 - acc: 0.8259Epoch 00003: val_loss improved from 0.40235 to 0.39545, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3746 - acc: 0.8258 - val_loss: 0.3954 - val_acc: 0.8156\n",
      "Epoch 4/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3666 - acc: 0.8295Epoch 00004: val_loss improved from 0.39545 to 0.39248, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3663 - acc: 0.8297 - val_loss: 0.3925 - val_acc: 0.8173\n",
      "Epoch 5/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3599 - acc: 0.8343Epoch 00005: val_loss improved from 0.39248 to 0.39184, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3600 - acc: 0.8343 - val_loss: 0.3918 - val_acc: 0.8175\n",
      "Epoch 6/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3545 - acc: 0.8372Epoch 00006: val_loss improved from 0.39184 to 0.38775, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3544 - acc: 0.8371 - val_loss: 0.3878 - val_acc: 0.8211\n",
      "Epoch 7/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3494 - acc: 0.8398Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3491 - acc: 0.8401 - val_loss: 0.3892 - val_acc: 0.8194\n",
      "Epoch 8/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3449 - acc: 0.8422Epoch 00008: val_loss improved from 0.38775 to 0.38749, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3450 - acc: 0.8421 - val_loss: 0.3875 - val_acc: 0.8204\n",
      "Epoch 9/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3402 - acc: 0.8449Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3404 - acc: 0.8450 - val_loss: 0.3886 - val_acc: 0.8194\n",
      "Epoch 10/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3370 - acc: 0.8465Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 13us/step - loss: 0.3370 - acc: 0.8463 - val_loss: 0.3894 - val_acc: 0.8206\n",
      "Epoch 11/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3332 - acc: 0.8485Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3335 - acc: 0.8486 - val_loss: 0.3914 - val_acc: 0.8189\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 6s 96us/step\n",
      "Model:  basic_model_adam\n",
      "0.82% (+/- 0.00%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.4611 - acc: 0.7705Epoch 00001: val_loss improved from inf to 0.39778, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 46s 346us/step - loss: 0.4605 - acc: 0.7709 - val_loss: 0.3978 - val_acc: 0.8115\n",
      "Epoch 2/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3957 - acc: 0.8147Epoch 00002: val_loss improved from 0.39778 to 0.38417, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3956 - acc: 0.8147 - val_loss: 0.3842 - val_acc: 0.8184\n",
      "Epoch 3/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3822 - acc: 0.8228Epoch 00003: val_loss improved from 0.38417 to 0.37944, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3822 - acc: 0.8228 - val_loss: 0.3794 - val_acc: 0.8234\n",
      "Epoch 4/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3742 - acc: 0.8276Epoch 00004: val_loss improved from 0.37944 to 0.37533, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3741 - acc: 0.8275 - val_loss: 0.3753 - val_acc: 0.8254\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 5/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3668 - acc: 0.8314Epoch 00005: val_loss improved from 0.37533 to 0.37286, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3668 - acc: 0.8313 - val_loss: 0.3729 - val_acc: 0.8263\n",
      "Epoch 6/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3609 - acc: 0.8344Epoch 00006: val_loss improved from 0.37286 to 0.37153, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3609 - acc: 0.8345 - val_loss: 0.3715 - val_acc: 0.8280\n",
      "Epoch 7/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3550 - acc: 0.8380Epoch 00007: val_loss improved from 0.37153 to 0.37037, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3559 - acc: 0.8376 - val_loss: 0.3704 - val_acc: 0.8278\n",
      "Epoch 8/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3515 - acc: 0.8403Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3516 - acc: 0.8402 - val_loss: 0.3706 - val_acc: 0.8280\n",
      "Epoch 9/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3479 - acc: 0.8422Epoch 00009: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3479 - acc: 0.8422 - val_loss: 0.3714 - val_acc: 0.8268\n",
      "Epoch 10/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3438 - acc: 0.8445Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3442 - acc: 0.8443 - val_loss: 0.3729 - val_acc: 0.8278\n",
      "Epoch 00010: early stopping\n",
      "66668/66668 [==============================] - 7s 101us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4566 - acc: 0.7713Epoch 00001: val_loss improved from inf to 0.41392, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 10s 74us/step - loss: 0.4564 - acc: 0.7714 - val_loss: 0.4139 - val_acc: 0.8029\n",
      "Epoch 2/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3898 - acc: 0.8178Epoch 00002: val_loss improved from 0.41392 to 0.40090, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3896 - acc: 0.8178 - val_loss: 0.4009 - val_acc: 0.8113\n",
      "Epoch 3/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3768 - acc: 0.8243Epoch 00003: val_loss improved from 0.40090 to 0.39588, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3769 - acc: 0.8243 - val_loss: 0.3959 - val_acc: 0.8145\n",
      "Epoch 4/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3686 - acc: 0.8302Epoch 00004: val_loss improved from 0.39588 to 0.39158, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3686 - acc: 0.8302 - val_loss: 0.3916 - val_acc: 0.8159\n",
      "Epoch 5/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3620 - acc: 0.8332Epoch 00005: val_loss improved from 0.39158 to 0.38887, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3621 - acc: 0.8332 - val_loss: 0.3889 - val_acc: 0.8157\n",
      "Epoch 6/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3561 - acc: 0.8368Epoch 00006: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3564 - acc: 0.8367 - val_loss: 0.3891 - val_acc: 0.8172\n",
      "Epoch 7/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3516 - acc: 0.8390Epoch 00007: val_loss improved from 0.38887 to 0.38714, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3515 - acc: 0.8390 - val_loss: 0.3871 - val_acc: 0.8214\n",
      "Epoch 8/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3469 - acc: 0.8427Epoch 00008: val_loss improved from 0.38714 to 0.38669, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3470 - acc: 0.8426 - val_loss: 0.3867 - val_acc: 0.8221\n",
      "Epoch 9/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3433 - acc: 0.8439Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3429 - acc: 0.8442 - val_loss: 0.3876 - val_acc: 0.8200\n",
      "Epoch 10/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3396 - acc: 0.8463Epoch 00010: val_loss improved from 0.38669 to 0.38626, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3395 - acc: 0.8464 - val_loss: 0.3863 - val_acc: 0.8223\n",
      "Epoch 11/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3365 - acc: 0.8480Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3364 - acc: 0.8481 - val_loss: 0.3874 - val_acc: 0.8227\n",
      "Epoch 12/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3336 - acc: 0.8490Epoch 00012: val_loss improved from 0.38626 to 0.38441, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3336 - acc: 0.8490 - val_loss: 0.3844 - val_acc: 0.8237\n",
      "Epoch 13/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3303 - acc: 0.8509Epoch 00013: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3303 - acc: 0.8509 - val_loss: 0.3861 - val_acc: 0.8222\n",
      "Epoch 14/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3274 - acc: 0.8527Epoch 00014: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3272 - acc: 0.8528 - val_loss: 0.3861 - val_acc: 0.8230\n",
      "Epoch 15/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3245 - acc: 0.8545Epoch 00015: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3245 - acc: 0.8545 - val_loss: 0.3862 - val_acc: 0.8215\n",
      "Epoch 00015: early stopping\n",
      "66666/66666 [==============================] - 7s 109us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4565 - acc: 0.7719Epoch 00001: val_loss improved from inf to 0.41439, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 18s 136us/step - loss: 0.4564 - acc: 0.7719 - val_loss: 0.4144 - val_acc: 0.8022\n",
      "Epoch 2/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3883 - acc: 0.8182Epoch 00002: val_loss improved from 0.41439 to 0.40038, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3878 - acc: 0.8185 - val_loss: 0.4004 - val_acc: 0.8110\n",
      "Epoch 3/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3742 - acc: 0.8256Epoch 00003: val_loss improved from 0.40038 to 0.39367, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3742 - acc: 0.8255 - val_loss: 0.3937 - val_acc: 0.8168\n",
      "Epoch 4/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3664 - acc: 0.8298Epoch 00004: val_loss improved from 0.39367 to 0.39094, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3663 - acc: 0.8299 - val_loss: 0.3909 - val_acc: 0.8181\n",
      "Epoch 5/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3603 - acc: 0.8342Epoch 00005: val_loss improved from 0.39094 to 0.39058, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3602 - acc: 0.8342 - val_loss: 0.3906 - val_acc: 0.8171\n",
      "Epoch 6/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3546 - acc: 0.8364Epoch 00006: val_loss improved from 0.39058 to 0.38641, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3545 - acc: 0.8365 - val_loss: 0.3864 - val_acc: 0.8215\n",
      "Epoch 7/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3496 - acc: 0.8395Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3494 - acc: 0.8397 - val_loss: 0.3876 - val_acc: 0.8202\n",
      "Epoch 8/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3450 - acc: 0.8424Epoch 00008: val_loss improved from 0.38641 to 0.38639, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3452 - acc: 0.8423 - val_loss: 0.3864 - val_acc: 0.8209\n",
      "Epoch 9/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3406 - acc: 0.8449Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3406 - acc: 0.8450 - val_loss: 0.3872 - val_acc: 0.8207\n",
      "Epoch 10/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3372 - acc: 0.8462Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3374 - acc: 0.8461 - val_loss: 0.3875 - val_acc: 0.8206\n",
      "Epoch 11/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3339 - acc: 0.8486Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3339 - acc: 0.8487 - val_loss: 0.3894 - val_acc: 0.8203\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 7s 104us/step\n",
      "Model:  basic_model_adam\n",
      "0.82% (+/- 0.00%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "129024/133332 [============================>.] - ETA: 1s - loss: 0.4626 - acc: 0.7696Epoch 00001: val_loss improved from inf to 0.39788, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 51s 380us/step - loss: 0.4605 - acc: 0.7713 - val_loss: 0.3979 - val_acc: 0.8116\n",
      "Epoch 2/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3954 - acc: 0.8144Epoch 00002: val_loss improved from 0.39788 to 0.38421, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3953 - acc: 0.8144 - val_loss: 0.3842 - val_acc: 0.8180\n",
      "Epoch 3/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3821 - acc: 0.8221Epoch 00003: val_loss improved from 0.38421 to 0.37962, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3822 - acc: 0.8221 - val_loss: 0.3796 - val_acc: 0.8212\n",
      "Epoch 4/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3743 - acc: 0.8273Epoch 00004: val_loss improved from 0.37962 to 0.37543, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3743 - acc: 0.8273 - val_loss: 0.3754 - val_acc: 0.8265\n",
      "Epoch 5/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3669 - acc: 0.8315Epoch 00005: val_loss improved from 0.37543 to 0.37249, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3670 - acc: 0.8313 - val_loss: 0.3725 - val_acc: 0.8262\n",
      "Epoch 6/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3613 - acc: 0.8345Epoch 00006: val_loss improved from 0.37249 to 0.37109, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3612 - acc: 0.8345 - val_loss: 0.3711 - val_acc: 0.8276\n",
      "Epoch 7/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3555 - acc: 0.8377Epoch 00007: val_loss improved from 0.37109 to 0.37061, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3563 - acc: 0.8373 - val_loss: 0.3706 - val_acc: 0.8273\n",
      "Epoch 8/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3519 - acc: 0.8395Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3520 - acc: 0.8393 - val_loss: 0.3709 - val_acc: 0.8272\n",
      "Epoch 9/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3483 - acc: 0.8415Epoch 00009: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3483 - acc: 0.8416 - val_loss: 0.3707 - val_acc: 0.8256\n",
      "Epoch 10/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3446 - acc: 0.8430Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3446 - acc: 0.8431 - val_loss: 0.3726 - val_acc: 0.8275\n",
      "Epoch 00010: early stopping\n",
      "66668/66668 [==============================] - 7s 103us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4570 - acc: 0.7714Epoch 00001: val_loss improved from inf to 0.41313, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 10s 77us/step - loss: 0.4568 - acc: 0.7715 - val_loss: 0.4131 - val_acc: 0.8030\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3896 - acc: 0.8178Epoch 00002: val_loss improved from 0.41313 to 0.40019, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3896 - acc: 0.8178 - val_loss: 0.4002 - val_acc: 0.8108\n",
      "Epoch 3/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3770 - acc: 0.8244Epoch 00003: val_loss improved from 0.40019 to 0.39528, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3770 - acc: 0.8243 - val_loss: 0.3953 - val_acc: 0.8132\n",
      "Epoch 4/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3689 - acc: 0.8299Epoch 00004: val_loss improved from 0.39528 to 0.39085, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3689 - acc: 0.8299 - val_loss: 0.3908 - val_acc: 0.8158\n",
      "Epoch 5/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3626 - acc: 0.8330Epoch 00005: val_loss improved from 0.39085 to 0.38830, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3624 - acc: 0.8332 - val_loss: 0.3883 - val_acc: 0.8202\n",
      "Epoch 6/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3566 - acc: 0.8364Epoch 00006: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3569 - acc: 0.8362 - val_loss: 0.3884 - val_acc: 0.8161\n",
      "Epoch 7/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3523 - acc: 0.8386Epoch 00007: val_loss improved from 0.38830 to 0.38658, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3521 - acc: 0.8386 - val_loss: 0.3866 - val_acc: 0.8216\n",
      "Epoch 8/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3476 - acc: 0.8417Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3477 - acc: 0.8416 - val_loss: 0.3867 - val_acc: 0.8222\n",
      "Epoch 9/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3439 - acc: 0.8442Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3436 - acc: 0.8444 - val_loss: 0.3876 - val_acc: 0.8193\n",
      "Epoch 10/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3403 - acc: 0.8448Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3404 - acc: 0.8447 - val_loss: 0.3871 - val_acc: 0.8223\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 00010: early stopping\n",
      "66666/66666 [==============================] - 7s 107us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4569 - acc: 0.7717Epoch 00001: val_loss improved from inf to 0.41407, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 10s 73us/step - loss: 0.4566 - acc: 0.7719 - val_loss: 0.4141 - val_acc: 0.8031\n",
      "Epoch 2/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3880 - acc: 0.8182Epoch 00002: val_loss improved from 0.41407 to 0.40008, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3876 - acc: 0.8184 - val_loss: 0.4001 - val_acc: 0.8126\n",
      "Epoch 3/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3741 - acc: 0.8255Epoch 00003: val_loss improved from 0.40008 to 0.39354, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3742 - acc: 0.8253 - val_loss: 0.3935 - val_acc: 0.8161\n",
      "Epoch 4/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3664 - acc: 0.8299Epoch 00004: val_loss improved from 0.39354 to 0.39119, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3663 - acc: 0.8300 - val_loss: 0.3912 - val_acc: 0.8172\n",
      "Epoch 5/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3604 - acc: 0.8336Epoch 00005: val_loss improved from 0.39119 to 0.39036, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3603 - acc: 0.8337 - val_loss: 0.3904 - val_acc: 0.8174\n",
      "Epoch 6/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3550 - acc: 0.8366Epoch 00006: val_loss improved from 0.39036 to 0.38623, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3548 - acc: 0.8367 - val_loss: 0.3862 - val_acc: 0.8215\n",
      "Epoch 7/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3498 - acc: 0.8391Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3497 - acc: 0.8392 - val_loss: 0.3874 - val_acc: 0.8209\n",
      "Epoch 8/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3458 - acc: 0.8414Epoch 00008: val_loss improved from 0.38623 to 0.38615, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3459 - acc: 0.8414 - val_loss: 0.3861 - val_acc: 0.8201\n",
      "Epoch 9/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3415 - acc: 0.8438Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3414 - acc: 0.8440 - val_loss: 0.3870 - val_acc: 0.8216\n",
      "Epoch 10/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3381 - acc: 0.8452Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3383 - acc: 0.8451 - val_loss: 0.3876 - val_acc: 0.8205\n",
      "Epoch 11/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3348 - acc: 0.8474Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3349 - acc: 0.8474 - val_loss: 0.3890 - val_acc: 0.8189\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 7s 103us/step\n",
      "Model:  basic_model_adam\n",
      "0.82% (+/- 0.00%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "132096/133332 [============================>.] - ETA: 1s - loss: 0.4616 - acc: 0.7699Epoch 00001: val_loss improved from inf to 0.39807, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 120s 901us/step - loss: 0.4610 - acc: 0.7703 - val_loss: 0.3981 - val_acc: 0.8107\n",
      "Epoch 2/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3960 - acc: 0.8144Epoch 00002: val_loss improved from 0.39807 to 0.38483, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3958 - acc: 0.8143 - val_loss: 0.3848 - val_acc: 0.8158\n",
      "Epoch 3/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3828 - acc: 0.8222Epoch 00003: val_loss improved from 0.38483 to 0.38054, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3829 - acc: 0.8221 - val_loss: 0.3805 - val_acc: 0.8199\n",
      "Epoch 4/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3753 - acc: 0.8272Epoch 00004: val_loss improved from 0.38054 to 0.37618, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3753 - acc: 0.8273 - val_loss: 0.3762 - val_acc: 0.8252\n",
      "Epoch 5/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3682 - acc: 0.8308Epoch 00005: val_loss improved from 0.37618 to 0.37325, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3682 - acc: 0.8308 - val_loss: 0.3732 - val_acc: 0.8253\n",
      "Epoch 6/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3625 - acc: 0.8336Epoch 00006: val_loss improved from 0.37325 to 0.37194, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3626 - acc: 0.8336 - val_loss: 0.3719 - val_acc: 0.8268\n",
      "Epoch 7/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3579 - acc: 0.8362Epoch 00007: val_loss improved from 0.37194 to 0.37116, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3578 - acc: 0.8362 - val_loss: 0.3712 - val_acc: 0.8275\n",
      "Epoch 8/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3536 - acc: 0.8391Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3536 - acc: 0.8389 - val_loss: 0.3723 - val_acc: 0.8246\n",
      "Epoch 9/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3499 - acc: 0.8409Epoch 00009: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3501 - acc: 0.8408 - val_loss: 0.3713 - val_acc: 0.8257\n",
      "Epoch 10/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3460 - acc: 0.8420Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3465 - acc: 0.8417 - val_loss: 0.3729 - val_acc: 0.8261\n",
      "Epoch 00010: early stopping\n",
      "66668/66668 [==============================] - 8s 115us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4576 - acc: 0.7705Epoch 00001: val_loss improved from inf to 0.41317, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 88us/step - loss: 0.4574 - acc: 0.7706 - val_loss: 0.4132 - val_acc: 0.8029\n",
      "Epoch 2/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3902 - acc: 0.8173Epoch 00002: val_loss improved from 0.41317 to 0.40026, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3901 - acc: 0.8173 - val_loss: 0.4003 - val_acc: 0.8110\n",
      "Epoch 3/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3778 - acc: 0.8243Epoch 00003: val_loss improved from 0.40026 to 0.39538, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3778 - acc: 0.8242 - val_loss: 0.3954 - val_acc: 0.8131\n",
      "Epoch 4/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3699 - acc: 0.8289Epoch 00004: val_loss improved from 0.39538 to 0.39104, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3700 - acc: 0.8289 - val_loss: 0.3910 - val_acc: 0.8201\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 5/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3636 - acc: 0.8323Epoch 00005: val_loss improved from 0.39104 to 0.38862, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3636 - acc: 0.8323 - val_loss: 0.3886 - val_acc: 0.8201\n",
      "Epoch 6/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3580 - acc: 0.8358Epoch 00006: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3583 - acc: 0.8356 - val_loss: 0.3887 - val_acc: 0.8220\n",
      "Epoch 7/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3537 - acc: 0.8377Epoch 00007: val_loss improved from 0.38862 to 0.38706, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3537 - acc: 0.8377 - val_loss: 0.3871 - val_acc: 0.8223\n",
      "Epoch 8/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3493 - acc: 0.8410Epoch 00008: val_loss improved from 0.38706 to 0.38664, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3494 - acc: 0.8410 - val_loss: 0.3866 - val_acc: 0.8226\n",
      "Epoch 9/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3454 - acc: 0.8427Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3453 - acc: 0.8428 - val_loss: 0.3879 - val_acc: 0.8199\n",
      "Epoch 10/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3420 - acc: 0.8445Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3421 - acc: 0.8443 - val_loss: 0.3867 - val_acc: 0.8231\n",
      "Epoch 11/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3388 - acc: 0.8463Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3390 - acc: 0.8461 - val_loss: 0.3876 - val_acc: 0.8215\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 8s 117us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4590 - acc: 0.7702Epoch 00001: val_loss improved from inf to 0.41424, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 90us/step - loss: 0.4569 - acc: 0.7716 - val_loss: 0.4142 - val_acc: 0.8029\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3883 - acc: 0.8179Epoch 00002: val_loss improved from 0.41424 to 0.40027, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3883 - acc: 0.8179 - val_loss: 0.4003 - val_acc: 0.8116\n",
      "Epoch 3/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3751 - acc: 0.8254Epoch 00003: val_loss improved from 0.40027 to 0.39373, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3753 - acc: 0.8253 - val_loss: 0.3937 - val_acc: 0.8156\n",
      "Epoch 4/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3677 - acc: 0.8293Epoch 00004: val_loss improved from 0.39373 to 0.39127, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3676 - acc: 0.8294 - val_loss: 0.3913 - val_acc: 0.8175\n",
      "Epoch 5/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3617 - acc: 0.8328Epoch 00005: val_loss improved from 0.39127 to 0.39126, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3616 - acc: 0.8328 - val_loss: 0.3913 - val_acc: 0.8158\n",
      "Epoch 6/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3563 - acc: 0.8355Epoch 00006: val_loss improved from 0.39126 to 0.38680, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3562 - acc: 0.8356 - val_loss: 0.3868 - val_acc: 0.8202\n",
      "Epoch 7/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3512 - acc: 0.8386Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3512 - acc: 0.8386 - val_loss: 0.3879 - val_acc: 0.8190\n",
      "Epoch 8/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3471 - acc: 0.8407Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3473 - acc: 0.8406 - val_loss: 0.3869 - val_acc: 0.8193\n",
      "Epoch 9/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3427 - acc: 0.8434Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3428 - acc: 0.8435 - val_loss: 0.3881 - val_acc: 0.8203\n",
      "Epoch 00009: early stopping\n",
      "66666/66666 [==============================] - 8s 123us/step\n",
      "Model:  basic_model_adam\n",
      "0.82% (+/- 0.00%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "129024/133332 [============================>.] - ETA: 3s - loss: 0.4643 - acc: 0.7679Epoch 00001: val_loss improved from inf to 0.39902, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 122s 913us/step - loss: 0.4621 - acc: 0.7695 - val_loss: 0.3990 - val_acc: 0.8107\n",
      "Epoch 2/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3971 - acc: 0.8132Epoch 00002: val_loss improved from 0.39902 to 0.38543, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3970 - acc: 0.8132 - val_loss: 0.3854 - val_acc: 0.8170\n",
      "Epoch 3/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3840 - acc: 0.8210Epoch 00003: val_loss improved from 0.38543 to 0.38176, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 16us/step - loss: 0.3841 - acc: 0.8210 - val_loss: 0.3818 - val_acc: 0.8212\n",
      "Epoch 4/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3766 - acc: 0.8259Epoch 00004: val_loss improved from 0.38176 to 0.37637, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3766 - acc: 0.8259 - val_loss: 0.3764 - val_acc: 0.8248\n",
      "Epoch 5/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3695 - acc: 0.8300Epoch 00005: val_loss improved from 0.37637 to 0.37385, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3697 - acc: 0.8297 - val_loss: 0.3739 - val_acc: 0.8267\n",
      "Epoch 6/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3642 - acc: 0.8325Epoch 00006: val_loss improved from 0.37385 to 0.37262, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3641 - acc: 0.8325 - val_loss: 0.3726 - val_acc: 0.8268\n",
      "Epoch 7/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3594 - acc: 0.8346Epoch 00007: val_loss improved from 0.37262 to 0.37197, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3593 - acc: 0.8346 - val_loss: 0.3720 - val_acc: 0.8267\n",
      "Epoch 8/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3551 - acc: 0.8368Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3553 - acc: 0.8367 - val_loss: 0.3731 - val_acc: 0.8253\n",
      "Epoch 9/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3518 - acc: 0.8389Epoch 00009: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3519 - acc: 0.8388 - val_loss: 0.3723 - val_acc: 0.8260\n",
      "Epoch 10/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3483 - acc: 0.8406Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3482 - acc: 0.8407 - val_loss: 0.3736 - val_acc: 0.8258\n",
      "Epoch 00010: early stopping\n",
      "66668/66668 [==============================] - 8s 119us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4601 - acc: 0.7691Epoch 00001: val_loss improved from inf to 0.41375, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 13s 95us/step - loss: 0.4585 - acc: 0.7700 - val_loss: 0.4137 - val_acc: 0.8025\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3911 - acc: 0.8163Epoch 00002: val_loss improved from 0.41375 to 0.40104, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3912 - acc: 0.8163 - val_loss: 0.4010 - val_acc: 0.8105\n",
      "Epoch 3/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3789 - acc: 0.8232Epoch 00003: val_loss improved from 0.40104 to 0.39632, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3790 - acc: 0.8232 - val_loss: 0.3963 - val_acc: 0.8131\n",
      "Epoch 4/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3713 - acc: 0.8285Epoch 00004: val_loss improved from 0.39632 to 0.39135, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3713 - acc: 0.8286 - val_loss: 0.3914 - val_acc: 0.8188\n",
      "Epoch 5/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3653 - acc: 0.8317Epoch 00005: val_loss improved from 0.39135 to 0.38927, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3651 - acc: 0.8319 - val_loss: 0.3893 - val_acc: 0.8199\n",
      "Epoch 6/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3596 - acc: 0.8348Epoch 00006: val_loss improved from 0.38927 to 0.38868, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 19us/step - loss: 0.3598 - acc: 0.8348 - val_loss: 0.3887 - val_acc: 0.8213\n",
      "Epoch 7/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3553 - acc: 0.8371Epoch 00007: val_loss improved from 0.38868 to 0.38736, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 19us/step - loss: 0.3552 - acc: 0.8371 - val_loss: 0.3874 - val_acc: 0.8207\n",
      "Epoch 8/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3510 - acc: 0.8399Epoch 00008: val_loss improved from 0.38736 to 0.38642, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 17us/step - loss: 0.3510 - acc: 0.8398 - val_loss: 0.3864 - val_acc: 0.8219\n",
      "Epoch 9/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3473 - acc: 0.8414Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3470 - acc: 0.8415 - val_loss: 0.3875 - val_acc: 0.8205\n",
      "Epoch 10/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3436 - acc: 0.8438Epoch 00010: val_loss improved from 0.38642 to 0.38594, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3437 - acc: 0.8438 - val_loss: 0.3859 - val_acc: 0.8227\n",
      "Epoch 11/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3408 - acc: 0.8450Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3408 - acc: 0.8451 - val_loss: 0.3867 - val_acc: 0.8220\n",
      "Epoch 12/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3380 - acc: 0.8465Epoch 00012: val_loss improved from 0.38594 to 0.38417, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3382 - acc: 0.8463 - val_loss: 0.3842 - val_acc: 0.8230\n",
      "Epoch 13/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3348 - acc: 0.8483Epoch 00013: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3350 - acc: 0.8480 - val_loss: 0.3869 - val_acc: 0.8211\n",
      "Epoch 14/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3324 - acc: 0.8499Epoch 00014: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3323 - acc: 0.8499 - val_loss: 0.3862 - val_acc: 0.8229\n",
      "Epoch 15/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3300 - acc: 0.8514Epoch 00015: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3298 - acc: 0.8515 - val_loss: 0.3852 - val_acc: 0.8228\n",
      "Epoch 00015: early stopping\n",
      "66666/66666 [==============================] - 10s 152us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4580 - acc: 0.7703Epoch 00001: val_loss improved from inf to 0.41523, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 15s 110us/step - loss: 0.4579 - acc: 0.7704 - val_loss: 0.4152 - val_acc: 0.8018\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3891 - acc: 0.8164Epoch 00002: val_loss improved from 0.41523 to 0.40139, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 3s 20us/step - loss: 0.3892 - acc: 0.8164 - val_loss: 0.4014 - val_acc: 0.8102\n",
      "Epoch 3/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3762 - acc: 0.8246Epoch 00003: val_loss improved from 0.40139 to 0.39509, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 18us/step - loss: 0.3761 - acc: 0.8244 - val_loss: 0.3951 - val_acc: 0.8142\n",
      "Epoch 4/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3685 - acc: 0.8289Epoch 00004: val_loss improved from 0.39509 to 0.39265, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 18us/step - loss: 0.3684 - acc: 0.8290 - val_loss: 0.3926 - val_acc: 0.8161\n",
      "Epoch 5/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3628 - acc: 0.8324Epoch 00005: val_loss improved from 0.39265 to 0.39219, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 17us/step - loss: 0.3627 - acc: 0.8324 - val_loss: 0.3922 - val_acc: 0.8153\n",
      "Epoch 6/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3577 - acc: 0.8349Epoch 00006: val_loss improved from 0.39219 to 0.38782, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 18us/step - loss: 0.3575 - acc: 0.8351 - val_loss: 0.3878 - val_acc: 0.8188\n",
      "Epoch 7/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3528 - acc: 0.8377Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 3s 19us/step - loss: 0.3525 - acc: 0.8379 - val_loss: 0.3887 - val_acc: 0.8179\n",
      "Epoch 8/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3486 - acc: 0.8402Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 18us/step - loss: 0.3489 - acc: 0.8401 - val_loss: 0.3880 - val_acc: 0.8190\n",
      "Epoch 9/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3448 - acc: 0.8414Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 18us/step - loss: 0.3447 - acc: 0.8415 - val_loss: 0.3888 - val_acc: 0.8189\n",
      "Epoch 00009: early stopping\n",
      "66666/66666 [==============================] - 10s 147us/step\n",
      "Model:  basic_model_adam\n",
      "0.82% (+/- 0.00%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4635 - acc: 0.7682Epoch 00001: val_loss improved from inf to 0.40006, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 129s 965us/step - loss: 0.4634 - acc: 0.7682 - val_loss: 0.4001 - val_acc: 0.8094\n",
      "Epoch 2/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3983 - acc: 0.8123Epoch 00002: val_loss improved from 0.40006 to 0.38657, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3983 - acc: 0.8123 - val_loss: 0.3866 - val_acc: 0.8157\n",
      "Epoch 3/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3853 - acc: 0.8201Epoch 00003: val_loss improved from 0.38657 to 0.38319, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3855 - acc: 0.8201 - val_loss: 0.3832 - val_acc: 0.8184\n",
      "Epoch 4/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3783 - acc: 0.8246Epoch 00004: val_loss improved from 0.38319 to 0.37768, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3783 - acc: 0.8246 - val_loss: 0.3777 - val_acc: 0.8253\n",
      "Epoch 5/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3711 - acc: 0.8289Epoch 00005: val_loss improved from 0.37768 to 0.37525, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3713 - acc: 0.8286 - val_loss: 0.3753 - val_acc: 0.8248\n",
      "Epoch 6/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3658 - acc: 0.8314Epoch 00006: val_loss improved from 0.37525 to 0.37405, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3658 - acc: 0.8314 - val_loss: 0.3740 - val_acc: 0.8264\n",
      "Epoch 7/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3605 - acc: 0.8347Epoch 00007: val_loss improved from 0.37405 to 0.37378, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3611 - acc: 0.8342 - val_loss: 0.3738 - val_acc: 0.8250\n",
      "Epoch 8/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3569 - acc: 0.8360Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3571 - acc: 0.8358 - val_loss: 0.3746 - val_acc: 0.8244\n",
      "Epoch 9/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3534 - acc: 0.8386Epoch 00009: val_loss improved from 0.37378 to 0.37357, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3536 - acc: 0.8386 - val_loss: 0.3736 - val_acc: 0.8246\n",
      "Epoch 10/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3496 - acc: 0.8401Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3500 - acc: 0.8398 - val_loss: 0.3755 - val_acc: 0.8250\n",
      "Epoch 11/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3470 - acc: 0.8411Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3471 - acc: 0.8410 - val_loss: 0.3745 - val_acc: 0.8250\n",
      "Epoch 12/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3438 - acc: 0.8429Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 16us/step - loss: 0.3435 - acc: 0.8430 - val_loss: 0.3740 - val_acc: 0.8264\n",
      "Epoch 00012: early stopping\n",
      "66668/66668 [==============================] - 9s 128us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4612 - acc: 0.7680Epoch 00001: val_loss improved from inf to 0.41528, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 15s 109us/step - loss: 0.4596 - acc: 0.7689 - val_loss: 0.4153 - val_acc: 0.8017\n",
      "Epoch 2/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3924 - acc: 0.8151Epoch 00002: val_loss improved from 0.41528 to 0.40276, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3924 - acc: 0.8152 - val_loss: 0.4028 - val_acc: 0.8089\n",
      "Epoch 3/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3801 - acc: 0.8221Epoch 00003: val_loss improved from 0.40276 to 0.39807, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3802 - acc: 0.8219 - val_loss: 0.3981 - val_acc: 0.8121\n",
      "Epoch 4/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3727 - acc: 0.8277Epoch 00004: val_loss improved from 0.39807 to 0.39343, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3727 - acc: 0.8278 - val_loss: 0.3934 - val_acc: 0.8185\n",
      "Epoch 5/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3668 - acc: 0.8309Epoch 00005: val_loss improved from 0.39343 to 0.39117, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3666 - acc: 0.8311 - val_loss: 0.3912 - val_acc: 0.8179\n",
      "Epoch 6/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3611 - acc: 0.8336Epoch 00006: val_loss improved from 0.39117 to 0.39084, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3613 - acc: 0.8336 - val_loss: 0.3908 - val_acc: 0.8203\n",
      "Epoch 7/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3571 - acc: 0.8363Epoch 00007: val_loss improved from 0.39084 to 0.38930, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3569 - acc: 0.8362 - val_loss: 0.3893 - val_acc: 0.8203\n",
      "Epoch 8/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3526 - acc: 0.8389Epoch 00008: val_loss improved from 0.38930 to 0.38920, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3528 - acc: 0.8388 - val_loss: 0.3892 - val_acc: 0.8203\n",
      "Epoch 9/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3490 - acc: 0.8404Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3488 - acc: 0.8406 - val_loss: 0.3896 - val_acc: 0.8206\n",
      "Epoch 10/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3457 - acc: 0.8424Epoch 00010: val_loss improved from 0.38920 to 0.38830, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3458 - acc: 0.8424 - val_loss: 0.3883 - val_acc: 0.8205\n",
      "Epoch 11/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3428 - acc: 0.8441Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3430 - acc: 0.8441 - val_loss: 0.3883 - val_acc: 0.8211\n",
      "Epoch 12/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3404 - acc: 0.8451Epoch 00012: val_loss improved from 0.38830 to 0.38694, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 3s 19us/step - loss: 0.3404 - acc: 0.8451 - val_loss: 0.3869 - val_acc: 0.8218\n",
      "Epoch 13/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3372 - acc: 0.8471Epoch 00013: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 18us/step - loss: 0.3372 - acc: 0.8470 - val_loss: 0.3893 - val_acc: 0.8188\n",
      "Epoch 14/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3350 - acc: 0.8489Epoch 00014: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3345 - acc: 0.8490 - val_loss: 0.3886 - val_acc: 0.8215\n",
      "Epoch 15/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3324 - acc: 0.8498Epoch 00015: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3322 - acc: 0.8500 - val_loss: 0.3875 - val_acc: 0.8213\n",
      "Epoch 00015: early stopping\n",
      "66666/66666 [==============================] - 9s 134us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4596 - acc: 0.7696Epoch 00001: val_loss improved from inf to 0.41647, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 13s 96us/step - loss: 0.4593 - acc: 0.7698 - val_loss: 0.4165 - val_acc: 0.8013\n",
      "Epoch 2/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3909 - acc: 0.8148Epoch 00002: val_loss improved from 0.41647 to 0.40259, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 18us/step - loss: 0.3907 - acc: 0.8150 - val_loss: 0.4026 - val_acc: 0.8097\n",
      "Epoch 3/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3778 - acc: 0.8246Epoch 00003: val_loss improved from 0.40259 to 0.39637, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 18us/step - loss: 0.3779 - acc: 0.8246 - val_loss: 0.3964 - val_acc: 0.8134\n",
      "Epoch 4/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3705 - acc: 0.8283Epoch 00004: val_loss improved from 0.39637 to 0.39402, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 18us/step - loss: 0.3703 - acc: 0.8283 - val_loss: 0.3940 - val_acc: 0.8155\n",
      "Epoch 5/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3644 - acc: 0.8319Epoch 00005: val_loss improved from 0.39402 to 0.39299, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 18us/step - loss: 0.3645 - acc: 0.8317 - val_loss: 0.3930 - val_acc: 0.8146\n",
      "Epoch 6/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3596 - acc: 0.8340Epoch 00006: val_loss improved from 0.39299 to 0.38852, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 18us/step - loss: 0.3593 - acc: 0.8342 - val_loss: 0.3885 - val_acc: 0.8184\n",
      "Epoch 7/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3544 - acc: 0.8367Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 3s 19us/step - loss: 0.3543 - acc: 0.8367 - val_loss: 0.3891 - val_acc: 0.8175\n",
      "Epoch 8/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3501 - acc: 0.8389Epoch 00008: val_loss improved from 0.38852 to 0.38838, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 3s 19us/step - loss: 0.3504 - acc: 0.8387 - val_loss: 0.3884 - val_acc: 0.8182\n",
      "Epoch 9/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3464 - acc: 0.8413Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 3s 19us/step - loss: 0.3464 - acc: 0.8414 - val_loss: 0.3894 - val_acc: 0.8179\n",
      "Epoch 10/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3435 - acc: 0.8427Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 3s 19us/step - loss: 0.3435 - acc: 0.8427 - val_loss: 0.3897 - val_acc: 0.8180\n",
      "Epoch 11/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3402 - acc: 0.8445Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 3s 19us/step - loss: 0.3403 - acc: 0.8445 - val_loss: 0.3899 - val_acc: 0.8169\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 8s 123us/step\n",
      "Model:  basic_model_adam\n",
      "0.82% (+/- 0.00%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4655 - acc: 0.7669Epoch 00001: val_loss improved from inf to 0.40158, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 121s 905us/step - loss: 0.4654 - acc: 0.7670 - val_loss: 0.4016 - val_acc: 0.8062\n",
      "Epoch 2/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4006 - acc: 0.8106Epoch 00002: val_loss improved from 0.40158 to 0.38851, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.4007 - acc: 0.8106 - val_loss: 0.3885 - val_acc: 0.8166\n",
      "Epoch 3/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3879 - acc: 0.8185Epoch 00003: val_loss improved from 0.38851 to 0.38568, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3882 - acc: 0.8183 - val_loss: 0.3857 - val_acc: 0.8191\n",
      "Epoch 4/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3811 - acc: 0.8227Epoch 00004: val_loss improved from 0.38568 to 0.37968, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3811 - acc: 0.8228 - val_loss: 0.3797 - val_acc: 0.8232\n",
      "Epoch 5/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3742 - acc: 0.8266Epoch 00005: val_loss improved from 0.37968 to 0.37771, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3742 - acc: 0.8265 - val_loss: 0.3777 - val_acc: 0.8224\n",
      "Epoch 6/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3689 - acc: 0.8296Epoch 00006: val_loss improved from 0.37771 to 0.37668, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3688 - acc: 0.8295 - val_loss: 0.3767 - val_acc: 0.8236\n",
      "Epoch 7/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3638 - acc: 0.8322Epoch 00007: val_loss improved from 0.37668 to 0.37620, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3642 - acc: 0.8320 - val_loss: 0.3762 - val_acc: 0.8244\n",
      "Epoch 8/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3603 - acc: 0.8341Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3603 - acc: 0.8341 - val_loss: 0.3776 - val_acc: 0.8226\n",
      "Epoch 9/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3570 - acc: 0.8363Epoch 00009: val_loss improved from 0.37620 to 0.37573, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3571 - acc: 0.8363 - val_loss: 0.3757 - val_acc: 0.8235\n",
      "Epoch 10/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3537 - acc: 0.8374Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3536 - acc: 0.8374 - val_loss: 0.3775 - val_acc: 0.8243\n",
      "Epoch 11/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3505 - acc: 0.8397Epoch 00011: val_loss improved from 0.37573 to 0.37497, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3506 - acc: 0.8396 - val_loss: 0.3750 - val_acc: 0.8254\n",
      "Epoch 12/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3473 - acc: 0.8410Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3472 - acc: 0.8411 - val_loss: 0.3759 - val_acc: 0.8242\n",
      "Epoch 13/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3441 - acc: 0.8430Epoch 00013: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3441 - acc: 0.8429 - val_loss: 0.3769 - val_acc: 0.8215\n",
      "Epoch 14/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3415 - acc: 0.8445Epoch 00014: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3416 - acc: 0.8444 - val_loss: 0.3790 - val_acc: 0.8240\n",
      "Epoch 00014: early stopping\n",
      "66668/66668 [==============================] - 9s 132us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4616 - acc: 0.7674Epoch 00001: val_loss improved from inf to 0.41746, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 91us/step - loss: 0.4615 - acc: 0.7675 - val_loss: 0.4175 - val_acc: 0.7996\n",
      "Epoch 2/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3945 - acc: 0.8133Epoch 00002: val_loss improved from 0.41746 to 0.40504, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3945 - acc: 0.8133 - val_loss: 0.4050 - val_acc: 0.8069\n",
      "Epoch 3/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3824 - acc: 0.8215Epoch 00003: val_loss improved from 0.40504 to 0.40080, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3825 - acc: 0.8214 - val_loss: 0.4008 - val_acc: 0.8099\n",
      "Epoch 4/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3755 - acc: 0.8261Epoch 00004: val_loss improved from 0.40080 to 0.39614, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3754 - acc: 0.8261 - val_loss: 0.3961 - val_acc: 0.8177\n",
      "Epoch 5/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3697 - acc: 0.8290Epoch 00005: val_loss improved from 0.39614 to 0.39405, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3695 - acc: 0.8292 - val_loss: 0.3941 - val_acc: 0.8157\n",
      "Epoch 6/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3643 - acc: 0.8322Epoch 00006: val_loss improved from 0.39405 to 0.39381, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3644 - acc: 0.8321 - val_loss: 0.3938 - val_acc: 0.8176\n",
      "Epoch 7/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3604 - acc: 0.8346Epoch 00007: val_loss improved from 0.39381 to 0.39223, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3602 - acc: 0.8346 - val_loss: 0.3922 - val_acc: 0.8172\n",
      "Epoch 8/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3561 - acc: 0.8363Epoch 00008: val_loss improved from 0.39223 to 0.39163, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3562 - acc: 0.8362 - val_loss: 0.3916 - val_acc: 0.8176\n",
      "Epoch 9/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3522 - acc: 0.8390Epoch 00009: val_loss improved from 0.39163 to 0.39094, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3522 - acc: 0.8390 - val_loss: 0.3909 - val_acc: 0.8194\n",
      "Epoch 10/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3494 - acc: 0.8404Epoch 00010: val_loss improved from 0.39094 to 0.39004, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3494 - acc: 0.8404 - val_loss: 0.3900 - val_acc: 0.8198\n",
      "Epoch 11/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3463 - acc: 0.8419Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3466 - acc: 0.8419 - val_loss: 0.3907 - val_acc: 0.8183\n",
      "Epoch 12/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3441 - acc: 0.8424Epoch 00012: val_loss improved from 0.39004 to 0.38857, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3442 - acc: 0.8424 - val_loss: 0.3886 - val_acc: 0.8195\n",
      "Epoch 13/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3405 - acc: 0.8452Epoch 00013: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3405 - acc: 0.8452 - val_loss: 0.3925 - val_acc: 0.8177\n",
      "Epoch 14/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3386 - acc: 0.8462Epoch 00014: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3381 - acc: 0.8463 - val_loss: 0.3922 - val_acc: 0.8193\n",
      "Epoch 15/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3361 - acc: 0.8474Epoch 00015: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3359 - acc: 0.8476 - val_loss: 0.3904 - val_acc: 0.8187\n",
      "Epoch 00015: early stopping\n",
      "66666/66666 [==============================] - 8s 123us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4616 - acc: 0.7676Epoch 00001: val_loss improved from inf to 0.41864, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 91us/step - loss: 0.4612 - acc: 0.7679 - val_loss: 0.4186 - val_acc: 0.7995\n",
      "Epoch 2/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3930 - acc: 0.8137Epoch 00002: val_loss improved from 0.41864 to 0.40466, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3929 - acc: 0.8138 - val_loss: 0.4047 - val_acc: 0.8080\n",
      "Epoch 3/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3802 - acc: 0.8228Epoch 00003: val_loss improved from 0.40466 to 0.39845, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3802 - acc: 0.8226 - val_loss: 0.3985 - val_acc: 0.8119\n",
      "Epoch 4/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3728 - acc: 0.8266Epoch 00004: val_loss improved from 0.39845 to 0.39626, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3727 - acc: 0.8266 - val_loss: 0.3963 - val_acc: 0.8149\n",
      "Epoch 5/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3673 - acc: 0.8297Epoch 00005: val_loss improved from 0.39626 to 0.39563, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3672 - acc: 0.8297 - val_loss: 0.3956 - val_acc: 0.8141\n",
      "Epoch 6/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3624 - acc: 0.8320Epoch 00006: val_loss improved from 0.39563 to 0.39138, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3620 - acc: 0.8322 - val_loss: 0.3914 - val_acc: 0.8167\n",
      "Epoch 7/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3574 - acc: 0.8341Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3574 - acc: 0.8341 - val_loss: 0.3920 - val_acc: 0.8170\n",
      "Epoch 8/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3534 - acc: 0.8369Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3536 - acc: 0.8369 - val_loss: 0.3920 - val_acc: 0.8151\n",
      "Epoch 9/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3498 - acc: 0.8391Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3498 - acc: 0.8392 - val_loss: 0.3924 - val_acc: 0.8172\n",
      "Epoch 00009: early stopping\n",
      "66666/66666 [==============================] - 8s 117us/step\n",
      "Model:  basic_model_adam\n",
      "0.82% (+/- 0.00%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "129024/133332 [============================>.] - ETA: 4s - loss: 0.4687 - acc: 0.7646Epoch 00001: val_loss improved from inf to 0.40318, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 128s 962us/step - loss: 0.4666 - acc: 0.7664 - val_loss: 0.4032 - val_acc: 0.8064\n",
      "Epoch 2/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.4021 - acc: 0.8097Epoch 00002: val_loss improved from 0.40318 to 0.39078, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 16us/step - loss: 0.4020 - acc: 0.8098 - val_loss: 0.3908 - val_acc: 0.8163\n",
      "Epoch 3/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3896 - acc: 0.8175Epoch 00003: val_loss improved from 0.39078 to 0.38756, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3897 - acc: 0.8174 - val_loss: 0.3876 - val_acc: 0.8174\n",
      "Epoch 4/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3827 - acc: 0.8219Epoch 00004: val_loss improved from 0.38756 to 0.38149, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3827 - acc: 0.8219 - val_loss: 0.3815 - val_acc: 0.8216\n",
      "Epoch 5/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3759 - acc: 0.8250Epoch 00005: val_loss improved from 0.38149 to 0.38006, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3761 - acc: 0.8249 - val_loss: 0.3801 - val_acc: 0.8186\n",
      "Epoch 6/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3710 - acc: 0.8277Epoch 00006: val_loss improved from 0.38006 to 0.37902, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3710 - acc: 0.8277 - val_loss: 0.3790 - val_acc: 0.8213\n",
      "Epoch 7/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3666 - acc: 0.8302Epoch 00007: val_loss improved from 0.37902 to 0.37850, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3665 - acc: 0.8302 - val_loss: 0.3785 - val_acc: 0.8220\n",
      "Epoch 8/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3626 - acc: 0.8324Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3627 - acc: 0.8324 - val_loss: 0.3795 - val_acc: 0.8208\n",
      "Epoch 9/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3592 - acc: 0.8346Epoch 00009: val_loss improved from 0.37850 to 0.37770, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3595 - acc: 0.8346 - val_loss: 0.3777 - val_acc: 0.8220\n",
      "Epoch 10/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3562 - acc: 0.8355Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3564 - acc: 0.8354 - val_loss: 0.3795 - val_acc: 0.8224\n",
      "Epoch 11/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3532 - acc: 0.8376Epoch 00011: val_loss improved from 0.37770 to 0.37706, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3533 - acc: 0.8376 - val_loss: 0.3771 - val_acc: 0.8224\n",
      "Epoch 12/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3502 - acc: 0.8384Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3501 - acc: 0.8386 - val_loss: 0.3781 - val_acc: 0.8214\n",
      "Epoch 13/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3471 - acc: 0.8410Epoch 00013: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 14us/step - loss: 0.3470 - acc: 0.8410 - val_loss: 0.3792 - val_acc: 0.8192\n",
      "Epoch 14/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3447 - acc: 0.8423Epoch 00014: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3445 - acc: 0.8423 - val_loss: 0.3805 - val_acc: 0.8207\n",
      "Epoch 00014: early stopping\n",
      "66668/66668 [==============================] - 9s 136us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.4641 - acc: 0.7658Epoch 00001: val_loss improved from inf to 0.41908, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 13s 97us/step - loss: 0.4626 - acc: 0.7668 - val_loss: 0.4191 - val_acc: 0.7997\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3961 - acc: 0.8127Epoch 00002: val_loss improved from 0.41908 to 0.40660, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3962 - acc: 0.8127 - val_loss: 0.4066 - val_acc: 0.8065\n",
      "Epoch 3/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3842 - acc: 0.8204Epoch 00003: val_loss improved from 0.40660 to 0.40285, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3843 - acc: 0.8204 - val_loss: 0.4028 - val_acc: 0.8084\n",
      "Epoch 4/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3774 - acc: 0.8248Epoch 00004: val_loss improved from 0.40285 to 0.39836, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3774 - acc: 0.8248 - val_loss: 0.3984 - val_acc: 0.8112\n",
      "Epoch 5/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3723 - acc: 0.8269Epoch 00005: val_loss improved from 0.39836 to 0.39664, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3716 - acc: 0.8274 - val_loss: 0.3966 - val_acc: 0.8103\n",
      "Epoch 6/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3665 - acc: 0.8301Epoch 00006: val_loss improved from 0.39664 to 0.39659, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3666 - acc: 0.8301 - val_loss: 0.3966 - val_acc: 0.8105\n",
      "Epoch 7/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3623 - acc: 0.8329Epoch 00007: val_loss improved from 0.39659 to 0.39476, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3626 - acc: 0.8327 - val_loss: 0.3948 - val_acc: 0.8118\n",
      "Epoch 8/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3585 - acc: 0.8348Epoch 00008: val_loss improved from 0.39476 to 0.39399, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3586 - acc: 0.8346 - val_loss: 0.3940 - val_acc: 0.8158\n",
      "Epoch 9/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3549 - acc: 0.8370Epoch 00009: val_loss improved from 0.39399 to 0.39371, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3548 - acc: 0.8371 - val_loss: 0.3937 - val_acc: 0.8186\n",
      "Epoch 10/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3518 - acc: 0.8389Epoch 00010: val_loss improved from 0.39371 to 0.39346, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3520 - acc: 0.8387 - val_loss: 0.3935 - val_acc: 0.8172\n",
      "Epoch 11/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3490 - acc: 0.8402Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3493 - acc: 0.8401 - val_loss: 0.3935 - val_acc: 0.8166\n",
      "Epoch 12/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3471 - acc: 0.8410Epoch 00012: val_loss improved from 0.39346 to 0.39195, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3471 - acc: 0.8410 - val_loss: 0.3920 - val_acc: 0.8182\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 13/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3436 - acc: 0.8427Epoch 00013: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3437 - acc: 0.8427 - val_loss: 0.3962 - val_acc: 0.8154\n",
      "Epoch 14/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3416 - acc: 0.8435Epoch 00014: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3414 - acc: 0.8435 - val_loss: 0.3954 - val_acc: 0.8180\n",
      "Epoch 15/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3393 - acc: 0.8455Epoch 00015: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3390 - acc: 0.8457 - val_loss: 0.3927 - val_acc: 0.8171\n",
      "Epoch 00015: early stopping\n",
      "66666/66666 [==============================] - 8s 121us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.4634 - acc: 0.7664Epoch 00001: val_loss improved from inf to 0.41944, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 91us/step - loss: 0.4627 - acc: 0.7670 - val_loss: 0.4194 - val_acc: 0.7999\n",
      "Epoch 2/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3947 - acc: 0.8127Epoch 00002: val_loss improved from 0.41944 to 0.40572, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3946 - acc: 0.8128 - val_loss: 0.4057 - val_acc: 0.8076\n",
      "Epoch 3/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3820 - acc: 0.8217Epoch 00003: val_loss improved from 0.40572 to 0.39951, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3822 - acc: 0.8216 - val_loss: 0.3995 - val_acc: 0.8128\n",
      "Epoch 4/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3751 - acc: 0.8248Epoch 00004: val_loss improved from 0.39951 to 0.39767, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3748 - acc: 0.8249 - val_loss: 0.3977 - val_acc: 0.8142\n",
      "Epoch 5/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3695 - acc: 0.8280Epoch 00005: val_loss improved from 0.39767 to 0.39724, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3695 - acc: 0.8279 - val_loss: 0.3972 - val_acc: 0.8142\n",
      "Epoch 6/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3649 - acc: 0.8303Epoch 00006: val_loss improved from 0.39724 to 0.39299, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3646 - acc: 0.8305 - val_loss: 0.3930 - val_acc: 0.8168\n",
      "Epoch 7/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3603 - acc: 0.8320Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3601 - acc: 0.8322 - val_loss: 0.3934 - val_acc: 0.8160\n",
      "Epoch 8/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3559 - acc: 0.8357Epoch 00008: val_loss improved from 0.39299 to 0.39291, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3562 - acc: 0.8356 - val_loss: 0.3929 - val_acc: 0.8168\n",
      "Epoch 9/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3527 - acc: 0.8365Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3526 - acc: 0.8366 - val_loss: 0.3933 - val_acc: 0.8171\n",
      "Epoch 10/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3499 - acc: 0.8385Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3500 - acc: 0.8385 - val_loss: 0.3930 - val_acc: 0.8163\n",
      "Epoch 11/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3466 - acc: 0.8406Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3468 - acc: 0.8405 - val_loss: 0.3940 - val_acc: 0.8158\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 8s 126us/step\n",
      "Model:  basic_model_adam\n",
      "0.82% (+/- 0.00%)\n",
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "132096/133332 [============================>.] - ETA: 1s - loss: 0.4693 - acc: 0.7644Epoch 00001: val_loss improved from inf to 0.40579, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 138s 1ms/step - loss: 0.4687 - acc: 0.7649 - val_loss: 0.4058 - val_acc: 0.8069\n",
      "Epoch 2/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.4049 - acc: 0.8082Epoch 00002: val_loss improved from 0.40579 to 0.39375, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 3s 20us/step - loss: 0.4048 - acc: 0.8082 - val_loss: 0.3937 - val_acc: 0.8140\n",
      "Epoch 3/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3925 - acc: 0.8155Epoch 00003: val_loss improved from 0.39375 to 0.39108, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 17us/step - loss: 0.3928 - acc: 0.8154 - val_loss: 0.3911 - val_acc: 0.8148\n",
      "Epoch 4/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3858 - acc: 0.8200Epoch 00004: val_loss improved from 0.39108 to 0.38455, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 3s 20us/step - loss: 0.3858 - acc: 0.8200 - val_loss: 0.3845 - val_acc: 0.8204\n",
      "Epoch 5/100\n",
      "131072/133332 [============================>.] - ETA: 0s - loss: 0.3788 - acc: 0.8237Epoch 00005: val_loss improved from 0.38455 to 0.38339, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 3s 20us/step - loss: 0.3793 - acc: 0.8232 - val_loss: 0.3834 - val_acc: 0.8172\n",
      "Epoch 6/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3743 - acc: 0.8262Epoch 00006: val_loss improved from 0.38339 to 0.38266, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 18us/step - loss: 0.3743 - acc: 0.8261 - val_loss: 0.3827 - val_acc: 0.8197\n",
      "Epoch 7/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3694 - acc: 0.8287Epoch 00007: val_loss improved from 0.38266 to 0.38225, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3698 - acc: 0.8285 - val_loss: 0.3823 - val_acc: 0.8194\n",
      "Epoch 8/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3662 - acc: 0.8302Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3662 - acc: 0.8302 - val_loss: 0.3829 - val_acc: 0.8192\n",
      "Epoch 9/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3628 - acc: 0.8326Epoch 00009: val_loss improved from 0.38225 to 0.38173, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3629 - acc: 0.8326 - val_loss: 0.3817 - val_acc: 0.8199\n",
      "Epoch 10/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3594 - acc: 0.8345Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3599 - acc: 0.8341 - val_loss: 0.3834 - val_acc: 0.8200\n",
      "Epoch 11/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3568 - acc: 0.8358Epoch 00011: val_loss improved from 0.38173 to 0.38086, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3568 - acc: 0.8358 - val_loss: 0.3809 - val_acc: 0.8208\n",
      "Epoch 12/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3538 - acc: 0.8378Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3536 - acc: 0.8380 - val_loss: 0.3819 - val_acc: 0.8185\n",
      "Epoch 13/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3505 - acc: 0.8397Epoch 00013: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3504 - acc: 0.8396 - val_loss: 0.3823 - val_acc: 0.8187\n",
      "Epoch 14/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3480 - acc: 0.8404Epoch 00014: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3480 - acc: 0.8404 - val_loss: 0.3846 - val_acc: 0.8183\n",
      "Epoch 00014: early stopping\n",
      "66668/66668 [==============================] - 10s 145us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.4652 - acc: 0.7655Epoch 00001: val_loss improved from inf to 0.42133, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 13s 95us/step - loss: 0.4646 - acc: 0.7659 - val_loss: 0.4213 - val_acc: 0.7957\n",
      "Epoch 2/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3988 - acc: 0.8116Epoch 00002: val_loss improved from 0.42133 to 0.40892, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3989 - acc: 0.8117 - val_loss: 0.4089 - val_acc: 0.8026\n",
      "Epoch 3/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3872 - acc: 0.8182Epoch 00003: val_loss improved from 0.40892 to 0.40471, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3872 - acc: 0.8182 - val_loss: 0.4047 - val_acc: 0.8065\n",
      "Epoch 4/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3801 - acc: 0.8225Epoch 00004: val_loss improved from 0.40471 to 0.40029, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3802 - acc: 0.8224 - val_loss: 0.4003 - val_acc: 0.8136\n",
      "Epoch 5/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3744 - acc: 0.8259Epoch 00005: val_loss improved from 0.40029 to 0.39882, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3744 - acc: 0.8259 - val_loss: 0.3988 - val_acc: 0.8133\n",
      "Epoch 6/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3693 - acc: 0.8289Epoch 00006: val_loss improved from 0.39882 to 0.39861, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3693 - acc: 0.8289 - val_loss: 0.3986 - val_acc: 0.8092\n",
      "Epoch 7/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3654 - acc: 0.8314Epoch 00007: val_loss improved from 0.39861 to 0.39650, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3654 - acc: 0.8314 - val_loss: 0.3965 - val_acc: 0.8141\n",
      "Epoch 8/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3613 - acc: 0.8333Epoch 00008: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3614 - acc: 0.8331 - val_loss: 0.3966 - val_acc: 0.8147\n",
      "Epoch 9/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3576 - acc: 0.8350Epoch 00009: val_loss improved from 0.39650 to 0.39641, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3576 - acc: 0.8351 - val_loss: 0.3964 - val_acc: 0.8148\n",
      "Epoch 10/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3548 - acc: 0.8367Epoch 00010: val_loss improved from 0.39641 to 0.39569, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3548 - acc: 0.8368 - val_loss: 0.3957 - val_acc: 0.8156\n",
      "Epoch 11/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3523 - acc: 0.8375Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3523 - acc: 0.8376 - val_loss: 0.3959 - val_acc: 0.8147\n",
      "Epoch 12/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3497 - acc: 0.8387Epoch 00012: val_loss improved from 0.39569 to 0.39449, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3498 - acc: 0.8387 - val_loss: 0.3945 - val_acc: 0.8163\n",
      "Epoch 13/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3462 - acc: 0.8406Epoch 00013: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3465 - acc: 0.8404 - val_loss: 0.3988 - val_acc: 0.8126\n",
      "Epoch 14/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3449 - acc: 0.8418Epoch 00014: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3444 - acc: 0.8419 - val_loss: 0.3972 - val_acc: 0.8147\n",
      "Epoch 15/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3422 - acc: 0.8436Epoch 00015: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3420 - acc: 0.8437 - val_loss: 0.3955 - val_acc: 0.8133\n",
      "Epoch 00015: early stopping\n",
      "66666/66666 [==============================] - 8s 125us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.4653 - acc: 0.7649Epoch 00001: val_loss improved from inf to 0.42196, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 91us/step - loss: 0.4646 - acc: 0.7654 - val_loss: 0.4220 - val_acc: 0.7984\n",
      "Epoch 2/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3976 - acc: 0.8112Epoch 00002: val_loss improved from 0.42196 to 0.40874, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3975 - acc: 0.8113 - val_loss: 0.4087 - val_acc: 0.8061\n",
      "Epoch 3/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3852 - acc: 0.8191Epoch 00003: val_loss improved from 0.40874 to 0.40278, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3852 - acc: 0.8190 - val_loss: 0.4028 - val_acc: 0.8112\n",
      "Epoch 4/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3781 - acc: 0.8226Epoch 00004: val_loss improved from 0.40278 to 0.40049, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3779 - acc: 0.8227 - val_loss: 0.4005 - val_acc: 0.8122\n",
      "Epoch 5/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3726 - acc: 0.8259Epoch 00005: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3726 - acc: 0.8259 - val_loss: 0.4005 - val_acc: 0.8120\n",
      "Epoch 6/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3682 - acc: 0.8286Epoch 00006: val_loss improved from 0.40049 to 0.39609, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3676 - acc: 0.8289 - val_loss: 0.3961 - val_acc: 0.8148\n",
      "Epoch 7/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3636 - acc: 0.8311Epoch 00007: val_loss improved from 0.39609 to 0.39604, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3633 - acc: 0.8314 - val_loss: 0.3960 - val_acc: 0.8150\n",
      "Epoch 8/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3593 - acc: 0.8337Epoch 00008: val_loss improved from 0.39604 to 0.39564, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3595 - acc: 0.8335 - val_loss: 0.3956 - val_acc: 0.8147\n",
      "Epoch 9/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3556 - acc: 0.8356Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3556 - acc: 0.8357 - val_loss: 0.3964 - val_acc: 0.8151\n",
      "Epoch 10/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.3529 - acc: 0.8371Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3532 - acc: 0.8369 - val_loss: 0.3957 - val_acc: 0.8158\n",
      "Epoch 11/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3501 - acc: 0.8386Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3502 - acc: 0.8386 - val_loss: 0.3959 - val_acc: 0.8157\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 8s 125us/step\n",
      "Model:  basic_model_adam\n",
      "0.82% (+/- 0.00%)\n"
     ]
    }
   ],
   "source": [
    "mindfs= [0,2,5,10,20, 40, 60,100,140,200]\n",
    "maxdfs=[0.8] \n",
    "accuracies_stop=[]\n",
    "stds_stop=[]\n",
    "stop_lens=[]\n",
    "vocabs=[]\n",
    "\n",
    "\n",
    "for max_ in maxdfs:\n",
    "    for min_ in mindfs: \n",
    "        stopwords, vocab= TO.get_dynamic_stopwords(best_corpus, MinDf=min_, MaxDf=max_)\n",
    "        stopword_corpus=TO.remove_stopwords(best_corpus, stopwords)\n",
    "        model_score=VP.classify_with_neural_networks(neural_nets, global_vectors, stopword_corpus, total_training_tweets, nr_pos_tweets, epochs=100, n_folds=3)\n",
    "        accuracies_stop.append(model_score[0][0])\n",
    "        stds_stop.append(model_score[0][1])\n",
    "        stop_lens.append(len(stopwords))\n",
    "        vocabs.append(vocab)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "min 0 accuracy 0.820739958766 std 0.00335801472691 \n",
      "\n",
      "min 2 accuracy 0.822214956417 std 0.00309534377653 \n",
      "\n",
      "min 5 accuracy 0.824164963467 std 0.00282831843706 \n",
      "\n",
      "min 10 accuracy 0.823009957167 std 0.00308777109608 \n",
      "\n",
      "min 20 accuracy 0.823419959017 std 0.00306118761546 \n",
      "\n",
      "min 40 accuracy 0.822874961367 std 0.00322542863439 \n",
      "\n",
      "min 60 accuracy 0.821529969366 std 0.002610905471 \n",
      "\n",
      "min 100 accuracy 0.820544951416 std 0.00361748132732 \n",
      "\n",
      "min 140 accuracy 0.819149967466 std 0.00237155277619 \n",
      "\n",
      "min 200 accuracy 0.817264964516 std 0.00259651604751 \n",
      "\n",
      "number of stop words 0 lenght of vocabolary 825916 sum for checking 825916 \n",
      "\n",
      "number of stop words 569652 lenght of vocabolary 256264 sum for checking 825916 \n",
      "\n",
      "number of stop words 738291 lenght of vocabolary 87625 sum for checking 825916 \n",
      "\n",
      "number of stop words 780215 lenght of vocabolary 45701 sum for checking 825916 \n",
      "\n",
      "number of stop words 801552 lenght of vocabolary 24364 sum for checking 825916 \n",
      "\n",
      "number of stop words 812855 lenght of vocabolary 13061 sum for checking 825916 \n",
      "\n",
      "number of stop words 816894 lenght of vocabolary 9022 sum for checking 825916 \n",
      "\n",
      "number of stop words 820312 lenght of vocabolary 5604 sum for checking 825916 \n",
      "\n",
      "number of stop words 821875 lenght of vocabolary 4041 sum for checking 825916 \n",
      "\n",
      "number of stop words 823113 lenght of vocabolary 2803 sum for checking 825916 \n",
      "\n"
     ]
    }
   ],
   "source": [
    "for i in range(len(stop_lens)):\n",
    "    print('min', mindfs[i], 'accuracy', accuracies_stop[i],'std',stds_stop[i],  '\\n')\n",
    "for i in range(len(stop_lens)):\n",
    "    print('number of stop words', stop_lens[i], 'lenght of vocabolary', len(vocabs[i]),'sum for checking',stop_lens[i]+len(vocabs[i]),  '\\n')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Testing the best combination with the simple and the complex neural net"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "stopwords, vocab= TO.get_dynamic_stopwords(best_corpus, MinDf=5, MaxDf=0.8)\n",
    "final_corpus= TO.remove_stopwords(best_corpus, stopwords)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "130048/133332 [============================>.] - ETA: 1s - loss: 0.4621 - acc: 0.7696Epoch 00001: val_loss improved from inf to 0.39778, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 73s 544us/step - loss: 0.4605 - acc: 0.7709 - val_loss: 0.3978 - val_acc: 0.8115\n",
      "Epoch 2/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3955 - acc: 0.8147Epoch 00002: val_loss improved from 0.39778 to 0.38417, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 17us/step - loss: 0.3956 - acc: 0.8147 - val_loss: 0.3842 - val_acc: 0.8184\n",
      "Epoch 3/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3819 - acc: 0.8230Epoch 00003: val_loss improved from 0.38417 to 0.37944, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 17us/step - loss: 0.3822 - acc: 0.8228 - val_loss: 0.3794 - val_acc: 0.8234\n",
      "Epoch 4/100\n",
      "129024/133332 [============================>.] - ETA: 0s - loss: 0.3741 - acc: 0.8276Epoch 00004: val_loss improved from 0.37944 to 0.37533, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 17us/step - loss: 0.3741 - acc: 0.8275 - val_loss: 0.3753 - val_acc: 0.8254\n",
      "Epoch 5/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3668 - acc: 0.8314Epoch 00005: val_loss improved from 0.37533 to 0.37286, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3668 - acc: 0.8313 - val_loss: 0.3729 - val_acc: 0.8263\n",
      "Epoch 6/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3609 - acc: 0.8344Epoch 00006: val_loss improved from 0.37286 to 0.37153, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3609 - acc: 0.8345 - val_loss: 0.3715 - val_acc: 0.8280\n",
      "Epoch 7/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3559 - acc: 0.8375Epoch 00007: val_loss improved from 0.37153 to 0.37037, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3559 - acc: 0.8376 - val_loss: 0.3704 - val_acc: 0.8278\n",
      "Epoch 8/100\n",
      "130048/133332 [============================>.] - ETA: 0s - loss: 0.3516 - acc: 0.8403Epoch 00008: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3516 - acc: 0.8402 - val_loss: 0.3706 - val_acc: 0.8280\n",
      "Epoch 9/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3477 - acc: 0.8423Epoch 00009: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3479 - acc: 0.8422 - val_loss: 0.3714 - val_acc: 0.8268\n",
      "Epoch 10/100\n",
      "132096/133332 [============================>.] - ETA: 0s - loss: 0.3441 - acc: 0.8442Epoch 00010: val_loss did not improve\n",
      "133332/133332 [==============================] - 2s 15us/step - loss: 0.3442 - acc: 0.8443 - val_loss: 0.3729 - val_acc: 0.8278\n",
      "Epoch 00010: early stopping\n",
      "66668/66668 [==============================] - 12s 183us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "131072/133334 [============================>.] - ETA: 0s - loss: 0.4571 - acc: 0.7709Epoch 00001: val_loss improved from inf to 0.41392, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 15s 116us/step - loss: 0.4564 - acc: 0.7714 - val_loss: 0.4139 - val_acc: 0.8029\n",
      "Epoch 2/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3898 - acc: 0.8178Epoch 00002: val_loss improved from 0.41392 to 0.40090, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3896 - acc: 0.8178 - val_loss: 0.4009 - val_acc: 0.8113\n",
      "Epoch 3/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3768 - acc: 0.8244Epoch 00003: val_loss improved from 0.40090 to 0.39588, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3769 - acc: 0.8243 - val_loss: 0.3959 - val_acc: 0.8145\n",
      "Epoch 4/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3686 - acc: 0.8303Epoch 00004: val_loss improved from 0.39588 to 0.39158, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3686 - acc: 0.8302 - val_loss: 0.3916 - val_acc: 0.8159\n",
      "Epoch 5/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3623 - acc: 0.8329Epoch 00005: val_loss improved from 0.39158 to 0.38887, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3621 - acc: 0.8332 - val_loss: 0.3889 - val_acc: 0.8157\n",
      "Epoch 6/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3563 - acc: 0.8368Epoch 00006: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3564 - acc: 0.8367 - val_loss: 0.3891 - val_acc: 0.8172\n",
      "Epoch 7/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3516 - acc: 0.8390Epoch 00007: val_loss improved from 0.38887 to 0.38714, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3515 - acc: 0.8390 - val_loss: 0.3871 - val_acc: 0.8214\n",
      "Epoch 8/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3469 - acc: 0.8427Epoch 00008: val_loss improved from 0.38714 to 0.38669, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 16us/step - loss: 0.3470 - acc: 0.8426 - val_loss: 0.3867 - val_acc: 0.8221\n",
      "Epoch 9/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3434 - acc: 0.8439Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3429 - acc: 0.8442 - val_loss: 0.3876 - val_acc: 0.8200\n",
      "Epoch 10/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3396 - acc: 0.8463Epoch 00010: val_loss improved from 0.38669 to 0.38626, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3395 - acc: 0.8464 - val_loss: 0.3863 - val_acc: 0.8223\n",
      "Epoch 11/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3365 - acc: 0.8480Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3364 - acc: 0.8481 - val_loss: 0.3874 - val_acc: 0.8227\n",
      "Epoch 12/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3336 - acc: 0.8490Epoch 00012: val_loss improved from 0.38626 to 0.38441, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3336 - acc: 0.8490 - val_loss: 0.3844 - val_acc: 0.8237\n",
      "Epoch 13/100\n",
      "130048/133334 [============================>.] - ETA: 0s - loss: 0.3301 - acc: 0.8512Epoch 00013: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3303 - acc: 0.8509 - val_loss: 0.3861 - val_acc: 0.8222\n",
      "Epoch 14/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3274 - acc: 0.8527Epoch 00014: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3272 - acc: 0.8528 - val_loss: 0.3861 - val_acc: 0.8230\n",
      "Epoch 15/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3245 - acc: 0.8544Epoch 00015: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3245 - acc: 0.8545 - val_loss: 0.3862 - val_acc: 0.8215\n",
      "Epoch 00015: early stopping\n",
      "66666/66666 [==============================] - 9s 131us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.4567 - acc: 0.7716Epoch 00001: val_loss improved from inf to 0.41439, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 11s 81us/step - loss: 0.4564 - acc: 0.7719 - val_loss: 0.4144 - val_acc: 0.8022\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3878 - acc: 0.8185Epoch 00002: val_loss improved from 0.41439 to 0.40038, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3878 - acc: 0.8185 - val_loss: 0.4004 - val_acc: 0.8110\n",
      "Epoch 3/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3741 - acc: 0.8255Epoch 00003: val_loss improved from 0.40038 to 0.39367, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3742 - acc: 0.8255 - val_loss: 0.3937 - val_acc: 0.8168\n",
      "Epoch 4/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3664 - acc: 0.8298Epoch 00004: val_loss improved from 0.39367 to 0.39094, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3663 - acc: 0.8299 - val_loss: 0.3909 - val_acc: 0.8181\n",
      "Epoch 5/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3601 - acc: 0.8342Epoch 00005: val_loss improved from 0.39094 to 0.39058, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3602 - acc: 0.8342 - val_loss: 0.3906 - val_acc: 0.8171\n",
      "Epoch 6/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3546 - acc: 0.8364Epoch 00006: val_loss improved from 0.39058 to 0.38641, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3545 - acc: 0.8365 - val_loss: 0.3864 - val_acc: 0.8215\n",
      "Epoch 7/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3494 - acc: 0.8396Epoch 00007: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3494 - acc: 0.8397 - val_loss: 0.3876 - val_acc: 0.8202\n",
      "Epoch 8/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3451 - acc: 0.8424Epoch 00008: val_loss improved from 0.38641 to 0.38639, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3452 - acc: 0.8423 - val_loss: 0.3864 - val_acc: 0.8209\n",
      "Epoch 9/100\n",
      "132096/133334 [============================>.] - ETA: 0s - loss: 0.3407 - acc: 0.8448Epoch 00009: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3406 - acc: 0.8450 - val_loss: 0.3872 - val_acc: 0.8207\n",
      "Epoch 10/100\n",
      "129024/133334 [============================>.] - ETA: 0s - loss: 0.3375 - acc: 0.8460Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 14us/step - loss: 0.3374 - acc: 0.8461 - val_loss: 0.3875 - val_acc: 0.8206\n",
      "Epoch 11/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3339 - acc: 0.8487Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 2s 15us/step - loss: 0.3339 - acc: 0.8487 - val_loss: 0.3894 - val_acc: 0.8203\n",
      "Epoch 00011: early stopping\n",
      "66666/66666 [==============================] - 9s 132us/step\n",
      "Model:  basic_model_adam\n",
      "0.82% (+/- 0.00%)\n",
      "Applying the best combination of preprocessing, using pre-trained global vectors with 200 dimensions, an cros-validatino accuracy of 0.824164963467 +- 0.00282831843706 was achieved using the simple neural net\n"
     ]
    }
   ],
   "source": [
    "model_score=VP.classify_with_neural_networks([NN.basic_model_adam], global_vectors, final_corpus, total_training_tweets, nr_pos_tweets, epochs=100, n_folds=3)\n",
    "\n",
    "print('Applying the best combination of preprocessing, using pre-trained global vectors with 200 dimensions, an cross-validatino accuracy of', model_score[0][0],'+-',model_score[0][1], 'was achieved using the simple neural net')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tweets processed: 0  of total number of tweets: 200000\n",
      "tweets processed: 50000  of total number of tweets: 200000\n",
      "tweets processed: 100000  of total number of tweets: 200000\n",
      "tweets processed: 150000  of total number of tweets: 200000\n",
      "Train on 133332 samples, validate on 66668 samples\n",
      "Epoch 1/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.4548 - acc: 0.7747Epoch 00001: val_loss improved from inf to 0.38662, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 176s 1ms/step - loss: 0.4547 - acc: 0.7748 - val_loss: 0.3866 - val_acc: 0.8192\n",
      "Epoch 2/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3990 - acc: 0.8125Epoch 00002: val_loss improved from 0.38662 to 0.37698, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 12s 88us/step - loss: 0.3989 - acc: 0.8126 - val_loss: 0.3770 - val_acc: 0.8231\n",
      "Epoch 3/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3869 - acc: 0.8188Epoch 00003: val_loss improved from 0.37698 to 0.37164, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 12s 89us/step - loss: 0.3869 - acc: 0.8188 - val_loss: 0.3716 - val_acc: 0.8255\n",
      "Epoch 4/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3797 - acc: 0.8231Epoch 00004: val_loss improved from 0.37164 to 0.36867, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 12s 89us/step - loss: 0.3797 - acc: 0.8231 - val_loss: 0.3687 - val_acc: 0.8281\n",
      "Epoch 5/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3725 - acc: 0.8267Epoch 00005: val_loss improved from 0.36867 to 0.36650, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 12s 90us/step - loss: 0.3725 - acc: 0.8267 - val_loss: 0.3665 - val_acc: 0.8307\n",
      "Epoch 6/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3683 - acc: 0.8299Epoch 00006: val_loss improved from 0.36650 to 0.36597, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 12s 88us/step - loss: 0.3682 - acc: 0.8300 - val_loss: 0.3660 - val_acc: 0.8325\n",
      "Epoch 7/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3628 - acc: 0.8327Epoch 00007: val_loss improved from 0.36597 to 0.36153, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 12s 89us/step - loss: 0.3627 - acc: 0.8327 - val_loss: 0.3615 - val_acc: 0.8305\n",
      "Epoch 8/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3592 - acc: 0.8345Epoch 00008: val_loss improved from 0.36153 to 0.36114, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 12s 90us/step - loss: 0.3592 - acc: 0.8345 - val_loss: 0.3611 - val_acc: 0.8324\n",
      "Epoch 9/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3541 - acc: 0.8370Epoch 00009: val_loss did not improve\n",
      "133332/133332 [==============================] - 12s 89us/step - loss: 0.3541 - acc: 0.8370 - val_loss: 0.3620 - val_acc: 0.8322\n",
      "Epoch 10/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3504 - acc: 0.8383Epoch 00010: val_loss improved from 0.36114 to 0.35975, saving model to best_neural_model_save.hdf5\n",
      "133332/133332 [==============================] - 12s 90us/step - loss: 0.3504 - acc: 0.8384 - val_loss: 0.3598 - val_acc: 0.8328\n",
      "Epoch 11/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3480 - acc: 0.8404Epoch 00011: val_loss did not improve\n",
      "133332/133332 [==============================] - 12s 90us/step - loss: 0.3481 - acc: 0.8404 - val_loss: 0.3610 - val_acc: 0.8332\n",
      "Epoch 12/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3439 - acc: 0.8409Epoch 00012: val_loss did not improve\n",
      "133332/133332 [==============================] - 12s 89us/step - loss: 0.3439 - acc: 0.8408 - val_loss: 0.3598 - val_acc: 0.8340\n",
      "Epoch 13/100\n",
      "133120/133332 [============================>.] - ETA: 0s - loss: 0.3405 - acc: 0.8430Epoch 00013: val_loss did not improve\n",
      "133332/133332 [==============================] - 12s 89us/step - loss: 0.3405 - acc: 0.8430 - val_loss: 0.3603 - val_acc: 0.8339\n",
      "Epoch 00013: early stopping\n",
      "66668/66668 [==============================] - 11s 159us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4496 - acc: 0.7757Epoch 00001: val_loss improved from inf to 0.40529, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 25s 190us/step - loss: 0.4494 - acc: 0.7758 - val_loss: 0.4053 - val_acc: 0.8085\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3927 - acc: 0.8155Epoch 00002: val_loss improved from 0.40529 to 0.39402, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 89us/step - loss: 0.3928 - acc: 0.8154 - val_loss: 0.3940 - val_acc: 0.8141\n",
      "Epoch 3/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3814 - acc: 0.8220Epoch 00003: val_loss improved from 0.39402 to 0.39227, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 13s 98us/step - loss: 0.3814 - acc: 0.8220 - val_loss: 0.3923 - val_acc: 0.8141\n",
      "Epoch 4/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3734 - acc: 0.8258Epoch 00004: val_loss improved from 0.39227 to 0.38705, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 92us/step - loss: 0.3734 - acc: 0.8258 - val_loss: 0.3871 - val_acc: 0.8172\n",
      "Epoch 5/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3671 - acc: 0.8298Epoch 00005: val_loss improved from 0.38705 to 0.38410, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 92us/step - loss: 0.3671 - acc: 0.8298 - val_loss: 0.3841 - val_acc: 0.8197\n",
      "Epoch 6/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3622 - acc: 0.8319Epoch 00006: val_loss improved from 0.38410 to 0.38104, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 91us/step - loss: 0.3621 - acc: 0.8319 - val_loss: 0.3810 - val_acc: 0.8202\n",
      "Epoch 7/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3571 - acc: 0.8356Epoch 00007: val_loss improved from 0.38104 to 0.37873, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 91us/step - loss: 0.3571 - acc: 0.8355 - val_loss: 0.3787 - val_acc: 0.8209\n",
      "Epoch 8/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3530 - acc: 0.8376Epoch 00008: val_loss improved from 0.37873 to 0.37775, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 90us/step - loss: 0.3531 - acc: 0.8376 - val_loss: 0.3778 - val_acc: 0.8214\n",
      "Epoch 9/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3509 - acc: 0.8373Epoch 00009: val_loss improved from 0.37775 to 0.37597, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 90us/step - loss: 0.3509 - acc: 0.8373 - val_loss: 0.3760 - val_acc: 0.8224\n",
      "Epoch 10/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3448 - acc: 0.8412Epoch 00010: val_loss improved from 0.37597 to 0.37493, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 89us/step - loss: 0.3448 - acc: 0.8411 - val_loss: 0.3749 - val_acc: 0.8216\n",
      "Epoch 11/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3427 - acc: 0.8423Epoch 00011: val_loss improved from 0.37493 to 0.37454, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 89us/step - loss: 0.3427 - acc: 0.8423 - val_loss: 0.3745 - val_acc: 0.8281\n",
      "Epoch 12/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3386 - acc: 0.8442Epoch 00012: val_loss did not improve\n",
      "133334/133334 [==============================] - 12s 90us/step - loss: 0.3385 - acc: 0.8442 - val_loss: 0.3756 - val_acc: 0.8226\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 13/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3349 - acc: 0.8456Epoch 00013: val_loss improved from 0.37454 to 0.37249, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 90us/step - loss: 0.3349 - acc: 0.8457 - val_loss: 0.3725 - val_acc: 0.8245\n",
      "Epoch 14/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3320 - acc: 0.8475Epoch 00014: val_loss did not improve\n",
      "133334/133334 [==============================] - 12s 88us/step - loss: 0.3320 - acc: 0.8475 - val_loss: 0.3734 - val_acc: 0.8291\n",
      "Epoch 15/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3299 - acc: 0.8492Epoch 00015: val_loss did not improve\n",
      "133334/133334 [==============================] - 12s 89us/step - loss: 0.3298 - acc: 0.8493 - val_loss: 0.3727 - val_acc: 0.8243\n",
      "Epoch 16/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3264 - acc: 0.8506Epoch 00016: val_loss did not improve\n",
      "133334/133334 [==============================] - 12s 89us/step - loss: 0.3265 - acc: 0.8505 - val_loss: 0.3744 - val_acc: 0.8239\n",
      "Epoch 00016: early stopping\n",
      "66666/66666 [==============================] - 11s 165us/step\n",
      "Train on 133334 samples, validate on 66666 samples\n",
      "Epoch 1/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.4470 - acc: 0.7786Epoch 00001: val_loss improved from inf to 0.40483, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 23s 175us/step - loss: 0.4469 - acc: 0.7787 - val_loss: 0.4048 - val_acc: 0.8081\n",
      "Epoch 2/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3899 - acc: 0.8173Epoch 00002: val_loss improved from 0.40483 to 0.39308, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 90us/step - loss: 0.3899 - acc: 0.8174 - val_loss: 0.3931 - val_acc: 0.8158\n",
      "Epoch 3/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3790 - acc: 0.8229Epoch 00003: val_loss improved from 0.39308 to 0.38830, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 90us/step - loss: 0.3789 - acc: 0.8229 - val_loss: 0.3883 - val_acc: 0.8182\n",
      "Epoch 4/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3715 - acc: 0.8261Epoch 00004: val_loss improved from 0.38830 to 0.38492, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 88us/step - loss: 0.3715 - acc: 0.8261 - val_loss: 0.3849 - val_acc: 0.8209\n",
      "Epoch 5/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3655 - acc: 0.8302Epoch 00005: val_loss improved from 0.38492 to 0.38154, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 88us/step - loss: 0.3655 - acc: 0.8302 - val_loss: 0.3815 - val_acc: 0.8229\n",
      "Epoch 6/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3600 - acc: 0.8326Epoch 00006: val_loss improved from 0.38154 to 0.37997, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 88us/step - loss: 0.3599 - acc: 0.8327 - val_loss: 0.3800 - val_acc: 0.8237\n",
      "Epoch 7/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3551 - acc: 0.8352Epoch 00007: val_loss improved from 0.37997 to 0.37920, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 88us/step - loss: 0.3551 - acc: 0.8351 - val_loss: 0.3792 - val_acc: 0.8248\n",
      "Epoch 8/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3516 - acc: 0.8370Epoch 00008: val_loss improved from 0.37920 to 0.37697, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 89us/step - loss: 0.3516 - acc: 0.8369 - val_loss: 0.3770 - val_acc: 0.8257\n",
      "Epoch 9/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3481 - acc: 0.8380Epoch 00009: val_loss improved from 0.37697 to 0.37572, saving model to best_neural_model_save.hdf5\n",
      "133334/133334 [==============================] - 12s 91us/step - loss: 0.3481 - acc: 0.8380 - val_loss: 0.3757 - val_acc: 0.8265\n",
      "Epoch 10/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3439 - acc: 0.8406Epoch 00010: val_loss did not improve\n",
      "133334/133334 [==============================] - 13s 94us/step - loss: 0.3439 - acc: 0.8407 - val_loss: 0.3762 - val_acc: 0.8257\n",
      "Epoch 11/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3414 - acc: 0.8417Epoch 00011: val_loss did not improve\n",
      "133334/133334 [==============================] - 12s 90us/step - loss: 0.3415 - acc: 0.8416 - val_loss: 0.3758 - val_acc: 0.8265\n",
      "Epoch 12/100\n",
      "133120/133334 [============================>.] - ETA: 0s - loss: 0.3376 - acc: 0.8441Epoch 00012: val_loss did not improve\n",
      "133334/133334 [==============================] - 12s 89us/step - loss: 0.3376 - acc: 0.8441 - val_loss: 0.3769 - val_acc: 0.8277\n",
      "Epoch 00012: early stopping\n",
      "66666/66666 [==============================] - 12s 177us/step\n",
      "Model:  complex_model\n",
      "0.83% (+/- 0.00%)\n",
      "Applying the best combination of preprocessing, using pre-trained global vectors with 200 dimensions, an cros-validatino accuracy of 0.827934951815 +- 0.00350598164425 was achieved using the compelx neural net\n"
     ]
    }
   ],
   "source": [
    "model_score=VP.classify_with_neural_networks([NN.complex_model], global_vectors, final_corpus, total_training_tweets, nr_pos_tweets, epochs=100, n_folds=3)\n",
    "\n",
    "print('Applying the best combination of preprocessing, using pre-trained global vectors with 200 dimensions, an cross-validatino accuracy of', model_score[0][0],'+-',model_score[0][1], 'was achieved using the compelx neural net')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}