LR and LSTM approach.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\tulasi ram\\Anaconda3\\lib\\site-packages\\sklearn\\cross_validation.py:41: DeprecationWarning: This module was deprecated in version 0.18 in favor of the model_selection module into which all the refactored classes and functions are moved. Also note that the interface of the new CV iterators are different from that of this module. This module will be removed in 0.20.\n",
      "  \"This module will be removed in 0.20.\", DeprecationWarning)\n",
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import re\n",
    "import time\n",
    "import warnings\n",
    "import numpy as np\n",
    "from nltk.corpus import stopwords\n",
    "from sklearn.decomposition import TruncatedSVD\n",
    "from sklearn.preprocessing import normalize\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.manifold import TSNE\n",
    "import seaborn as sns\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.metrics.classification import accuracy_score, log_loss\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.linear_model import SGDClassifier\n",
    "from imblearn.over_sampling import SMOTE\n",
    "from collections import Counter\n",
    "from scipy.sparse import hstack\n",
    "from sklearn.multiclass import OneVsRestClassifier\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.cross_validation import StratifiedKFold \n",
    "from collections import Counter, defaultdict\n",
    "from sklearn.calibration import CalibratedClassifierCV\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from sklearn.naive_bayes import GaussianNB\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "import math\n",
    "from sklearn.metrics import normalized_mutual_info_score\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "from mlxtend.classifier import StackingClassifier\n",
    "from tqdm import tqdm\n",
    "\n",
    "from sklearn import model_selection\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "import keras"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#after merging and preprocessed data\n",
    "result = pd.read_csv('cancer_data',index_col=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ID</th>\n",
       "      <th>Gene</th>\n",
       "      <th>Variation</th>\n",
       "      <th>Class</th>\n",
       "      <th>TEXT</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>FAM58A</td>\n",
       "      <td>Truncating Mutations</td>\n",
       "      <td>1</td>\n",
       "      <td>cyclin dependent kinases cdks regulate variety...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>CBL</td>\n",
       "      <td>W802*</td>\n",
       "      <td>2</td>\n",
       "      <td>abstract background non small cell lung cancer...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>CBL</td>\n",
       "      <td>Q249E</td>\n",
       "      <td>2</td>\n",
       "      <td>abstract background non small cell lung cancer...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>CBL</td>\n",
       "      <td>N454D</td>\n",
       "      <td>3</td>\n",
       "      <td>recent evidence demonstrated acquired uniparen...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>CBL</td>\n",
       "      <td>L399V</td>\n",
       "      <td>4</td>\n",
       "      <td>oncogenic mutations monomeric casitas b lineag...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   ID    Gene             Variation  Class  \\\n",
       "0   0  FAM58A  Truncating Mutations      1   \n",
       "1   1     CBL                 W802*      2   \n",
       "2   2     CBL                 Q249E      2   \n",
       "3   3     CBL                 N454D      3   \n",
       "4   4     CBL                 L399V      4   \n",
       "\n",
       "                                                TEXT  \n",
       "0  cyclin dependent kinases cdks regulate variety...  \n",
       "1  abstract background non small cell lung cancer...  \n",
       "2  abstract background non small cell lung cancer...  \n",
       "3  recent evidence demonstrated acquired uniparen...  \n",
       "4  oncogenic mutations monomeric casitas b lineag...  "
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Splitting data into train, test and cross validation (64:20:16)**\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_true = result['Class'].values\n",
    "result.Gene      = result.Gene.str.replace('\\s+', '_')\n",
    "result.Variation = result.Variation.str.replace('\\s+', '_')\n",
    "\n",
    "# split the data into test and train by maintaining same distribution of output varaible 'y_true' [stratify=y_true]\n",
    "X_train, test_df, y_train, y_test = train_test_split(result, y_true, stratify=y_true, test_size=0.2)\n",
    "# split the train data into train and cross validation by maintaining same distribution of output varaible 'y_train' [stratify=y_train]\n",
    "train_df, cv_df, y_train, y_cv = train_test_split(X_train, y_train, stratify=y_train, test_size=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# This function plots the confusion matrices given y_i, y_i_hat.\n",
    "def plot_confusion_matrix(test_y, predict_y):\n",
    "    C = confusion_matrix(test_y, predict_y)\n",
    "    A =(((C.T)/(C.sum(axis=1))).T)\n",
    "                   \n",
    "    B =(C/C.sum(axis=0))\n",
    " \n",
    "    labels = [1,2,3,4,5,6,7,8,9]\n",
    "    # representing A in heatmap format\n",
    "    print(\"-\"*20, \"Confusion matrix\", \"-\"*20)\n",
    "    plt.figure(figsize=(20,7))\n",
    "    sns.heatmap(C, annot=True, cmap=\"YlGnBu\", fmt=\".3f\", xticklabels=labels, yticklabels=labels)\n",
    "    plt.xlabel('Predicted Class')\n",
    "    plt.ylabel('Original Class')\n",
    "    plt.show()\n",
    "\n",
    "    print(\"-\"*20, \"Precision matrix (Columm Sum=1)\", \"-\"*20)\n",
    "    plt.figure(figsize=(20,7))\n",
    "    sns.heatmap(B, annot=True, cmap=\"YlGnBu\", fmt=\".3f\", xticklabels=labels, yticklabels=labels)\n",
    "    plt.xlabel('Predicted Class')\n",
    "    plt.ylabel('Original Class')\n",
    "    plt.show()\n",
    "    \n",
    "    # representing B in heatmap format\n",
    "    print(\"-\"*20, \"Recall matrix (Row sum=1)\", \"-\"*20)\n",
    "    plt.figure(figsize=(20,7))\n",
    "    sns.heatmap(A, annot=True, cmap=\"YlGnBu\", fmt=\".3f\", xticklabels=labels, yticklabels=labels)\n",
    "    plt.xlabel('Predicted Class')\n",
    "    plt.ylabel('Original Class')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# code for response coding with Laplace smoothing.\n",
    "# alpha : used for laplace smoothing\n",
    "# feature: ['gene', 'variation']\n",
    "# df: ['train_df', 'test_df', 'cv_df']\n",
    "# get_gv_fea_dict: Get Gene varaition Feature Dict\n",
    "def get_gv_fea_dict(alpha, feature, df):\n",
    "    value_count = train_df[feature].value_counts()\n",
    "    # gv_dict : Gene Variation Dict, which contains the probability array for each gene/variation\n",
    "    gv_dict = dict()\n",
    "    for i, denominator in value_count.items():\n",
    "        vec = []\n",
    "        for k in range(1,10):\n",
    "            cls_cnt = train_df.loc[(train_df['Class']==k) & (train_df[feature]==i)]\n",
    "            vec.append((cls_cnt.shape[0] + alpha*10)/ (denominator + 90*alpha))\n",
    "\n",
    "        # we are adding the gene/variation to the dict as key and vec as value\n",
    "        gv_dict[i]=vec\n",
    "    return gv_dict\n",
    "\n",
    "# Get Gene variation feature\n",
    "def get_gv_feature(alpha, feature, df):\n",
    "\n",
    "    gv_dict = get_gv_fea_dict(alpha, feature, df)\n",
    "    # value_count is similar in get_gv_fea_dict\n",
    "    value_count = train_df[feature].value_counts()\n",
    "    \n",
    "    # gv_fea: Gene_variation feature, it will contain the feature for each feature value in the data\n",
    "    gv_fea = []\n",
    "    for index, row in df.iterrows():\n",
    "        if row[feature] in dict(value_count).keys():\n",
    "            gv_fea.append(gv_dict[row[feature]])\n",
    "        else:\n",
    "            gv_fea.append([1/9,1/9,1/9,1/9,1/9,1/9,1/9,1/9,1/9])\n",
    "#             gv_fea.append([-1,-1,-1,-1,-1,-1,-1,-1,-1])\n",
    "    return gv_fea"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "### one-hot encoding of Gene feature.\n",
    "gene_vectorizer = CountVectorizer(ngram_range=(1,2))\n",
    "train_gene_feature_onehotCoding = gene_vectorizer.fit_transform(train_df['Gene'])\n",
    "test_gene_feature_onehotCoding = gene_vectorizer.transform(test_df['Gene'])\n",
    "cv_gene_feature_onehotCoding = gene_vectorizer.transform(cv_df['Gene'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# one-hot encoding of variation feature.\n",
    "variation_vectorizer = CountVectorizer(ngram_range=(1,2))\n",
    "train_variation_feature_onehotCoding = variation_vectorizer.fit_transform(train_df['Variation'])\n",
    "test_variation_feature_onehotCoding = variation_vectorizer.transform(test_df['Variation'])\n",
    "cv_variation_feature_onehotCoding = variation_vectorizer.transform(cv_df['Variation'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total number of unique words in train data : 773113\n"
     ]
    }
   ],
   "source": [
    "# building a TfidfVectorizer with all the words that occured minimum 3 times in train data\n",
    "text_vectorizer = CountVectorizer(min_df=3,ngram_range=(1,2))\n",
    "train_text_feature_onehotCoding = text_vectorizer.fit_transform(train_df['TEXT'])\n",
    "# getting all the feature names (words)\n",
    "train_text_features= text_vectorizer.get_feature_names()\n",
    "\n",
    "# train_text_feature_onehotCoding.sum(axis=0).A1 will sum every row and returns (1*number of features) vector\n",
    "train_text_fea_counts = train_text_feature_onehotCoding.sum(axis=0).A1\n",
    "\n",
    "# zip(list(text_features),text_fea_counts) will zip a word with its number of times it occured\n",
    "text_fea_dict = dict(zip(list(train_text_features),train_text_fea_counts))\n",
    "\n",
    "\n",
    "print(\"Total number of unique words in train data :\", len(train_text_features))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# normalizing every feature\n",
    "train_text_feature_onehotCoding = normalize(train_text_feature_onehotCoding, axis=0)\n",
    "\n",
    "# we use the same vectorizer that was trained on train data\n",
    "test_text_feature_onehotCoding = text_vectorizer.transform(test_df['TEXT'])\n",
    "# don't forget to normalize every feature\n",
    "test_text_feature_onehotCoding = normalize(test_text_feature_onehotCoding, axis=0)\n",
    "\n",
    "# we use the same vectorizer that was trained on train data\n",
    "cv_text_feature_onehotCoding = text_vectorizer.transform(cv_df['TEXT'])\n",
    "#normalizing every feature\n",
    "cv_text_feature_onehotCoding = normalize(cv_text_feature_onehotCoding, axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# merging gene, variance and text features\n",
    "train_gene_var_onehotCoding = hstack((train_gene_feature_onehotCoding,train_variation_feature_onehotCoding))\n",
    "test_gene_var_onehotCoding = hstack((test_gene_feature_onehotCoding,test_variation_feature_onehotCoding))\n",
    "cv_gene_var_onehotCoding = hstack((cv_gene_feature_onehotCoding,cv_variation_feature_onehotCoding))\n",
    "\n",
    "train_x_onehotCoding = hstack((train_gene_var_onehotCoding, train_text_feature_onehotCoding)).tocsr()\n",
    "train_y = np.array(list(train_df['Class']))\n",
    "\n",
    "test_x_onehotCoding = hstack((test_gene_var_onehotCoding, test_text_feature_onehotCoding)).tocsr()\n",
    "test_y = np.array(list(test_df['Class']))\n",
    "\n",
    "cv_x_onehotCoding = hstack((cv_gene_var_onehotCoding, cv_text_feature_onehotCoding)).tocsr()\n",
    "cv_y = np.array(list(cv_df['Class']))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def predict_and_plot_confusion_matrix(train_x, train_y,test_x, test_y, clf):\n",
    "    clf.fit(train_x, train_y)\n",
    "    sig_clf = CalibratedClassifierCV(clf, method=\"sigmoid\")\n",
    "    sig_clf.fit(train_x, train_y)\n",
    "    pred_y = sig_clf.predict(test_x)\n",
    "\n",
    "    # for calculating log_loss we willl provide the array of probabilities belongs to each class\n",
    "    print(\"Log loss :\",log_loss(test_y, sig_clf.predict_proba(test_x)))\n",
    "    # calculating the number of data points that are misclassified\n",
    "    print(\"Number of mis-classified points :\", np.count_nonzero((pred_y- test_y))/test_y.shape[0])\n",
    "    plot_confusion_matrix(test_y, pred_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def report_log_loss(train_x, train_y, test_x, test_y,  clf):\n",
    "    clf.fit(train_x, train_y)\n",
    "    sig_clf = CalibratedClassifierCV(clf, method=\"sigmoid\")\n",
    "    sig_clf.fit(train_x, train_y)\n",
    "    sig_clf_probs = sig_clf.predict_proba(test_x)\n",
    "    return log_loss(test_y, sig_clf_probs, eps=1e-15)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Logistic Regression**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*With Class Balancing*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "for alpha = 1e-06\n",
      "Log Loss : 1.53298313437766\n",
      "for alpha = 1e-05\n",
      "Log Loss : 1.5464018591715154\n",
      "for alpha = 0.0001\n",
      "Log Loss : 1.543208771557003\n",
      "for alpha = 0.001\n",
      "Log Loss : 1.4616086103029957\n",
      "for alpha = 0.01\n",
      "Log Loss : 1.1759957762030604\n",
      "for alpha = 0.1\n",
      "Log Loss : 1.2022043275319736\n",
      "for alpha = 1\n",
      "Log Loss : 1.254903804227539\n",
      "for alpha = 10\n",
      "Log Loss : 1.3113137106070625\n",
      "for alpha = 100\n",
      "Log Loss : 1.325428127644812\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbwAAAEWCAYAAAAdNyJXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nLaTgRURELfwXeLiLiITKcOEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1c647364c88>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "For values of best alpha =  0.01 The train log loss is: 0.8350900757184301\n",
      "For values of best alpha =  0.01 The cross validation log loss is: 1.1759957762030604\n",
      "For values of best alpha =  0.01 The test log loss is: 1.1924056222620378\n"
     ]
    }
   ],
   "source": [
    "#hyperparameter tuning\n",
    "alpha = [10 ** x for x in range(-6, 3)]\n",
    "cv_log_error_array = []\n",
    "for i in alpha:\n",
    "    print(\"for alpha =\", i)\n",
    "    clf = SGDClassifier(class_weight='balanced', alpha=i, penalty='l2', loss='log', random_state=42)\n",
    "    clf.fit(train_x_onehotCoding, train_y)\n",
    "    sig_clf = CalibratedClassifierCV(clf, method=\"sigmoid\")\n",
    "    sig_clf.fit(train_x_onehotCoding, train_y)\n",
    "    sig_clf_probs = sig_clf.predict_proba(cv_x_onehotCoding)\n",
    "    cv_log_error_array.append(log_loss(cv_y, sig_clf_probs, labels=clf.classes_, eps=1e-15))\n",
    "    # to avoid rounding error while multiplying probabilites we use log-probability estimates\n",
    "    print(\"Log Loss :\",log_loss(cv_y, sig_clf_probs)) \n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax.plot(alpha, cv_log_error_array,c='g')\n",
    "for i, txt in enumerate(np.round(cv_log_error_array,3)):\n",
    "    ax.annotate((alpha[i],str(txt)), (alpha[i],cv_log_error_array[i]))\n",
    "plt.grid()\n",
    "plt.title(\"Cross Validation Error for each alpha\")\n",
    "plt.xlabel(\"Alpha i's\")\n",
    "plt.ylabel(\"Error measure\")\n",
    "plt.show()\n",
    "\n",
    "\n",
    "best_alpha = np.argmin(cv_log_error_array)\n",
    "clf = SGDClassifier(class_weight='balanced', alpha=alpha[best_alpha], penalty='l2', loss='log', random_state=42)\n",
    "clf.fit(train_x_onehotCoding, train_y)\n",
    "sig_clf = CalibratedClassifierCV(clf, method=\"sigmoid\")\n",
    "sig_clf.fit(train_x_onehotCoding, train_y)\n",
    "\n",
    "predict_y = sig_clf.predict_proba(train_x_onehotCoding)\n",
    "print('For values of best alpha = ', alpha[best_alpha], \"The train log loss is:\",log_loss(y_train, predict_y, labels=clf.classes_, eps=1e-15))\n",
    "predict_y = sig_clf.predict_proba(cv_x_onehotCoding)\n",
    "print('For values of best alpha = ', alpha[best_alpha], \"The cross validation log loss is:\",log_loss(y_cv, predict_y, labels=clf.classes_, eps=1e-15))\n",
    "predict_y = sig_clf.predict_proba(test_x_onehotCoding)\n",
    "print('For values of best alpha = ', alpha[best_alpha], \"The test log loss is:\",log_loss(y_test, predict_y, labels=clf.classes_, eps=1e-15))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Testing the model with best hyper paramters**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Log loss : 1.1759957762030604\n",
      "Number of mis-classified points : 0.40789473684210525\n",
      "-------------------- Confusion matrix --------------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABA4AAAGtCAYAAACWbjanAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZmZF/T9JLV30sGVkKAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1c647282d68>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------- Precision matrix (Columm Sum=1) --------------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABAsAAAGtCAYAAABwR/3jAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nNWOxQJIkSZIkNfP/AciJNq6W7G3TAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1c64c540828>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------- Recall matrix (Row sum=1) --------------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABBIAAAGtCAYAAACr/FfBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZmZmZmZmeXMhwczMzMzMzMzy9v/w/R51aRKngwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1c647a149b0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "clf = SGDClassifier(class_weight='balanced', alpha=alpha[best_alpha], penalty='l2', loss='log', random_state=42)\n",
    "predict_and_plot_confusion_matrix(train_x_onehotCoding, train_y, cv_x_onehotCoding, cv_y, clf)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**INFERENCE**\n",
    "\n",
    "*Log loss value is greater than the one with unigrams and the results are not much good.*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "**ASSIGNMENT 4**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from keras.datasets import imdb\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.layers import LSTM\n",
    "from keras.layers.embeddings import Embedding\n",
    "from keras.preprocessing import sequence\n",
    "from keras.preprocessing.text import Tokenizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def dictionary_paddle_text(cls_text):\n",
    "    dictionary = defaultdict(int)\n",
    "    for index, row in cls_text.iterrows():\n",
    "        for word in row['TEXT'].split():\n",
    "            dictionary[word] +=1\n",
    "    return dictionary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "text_dictionary = dictionary_paddle_text(result)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sorted_text_dictionary = dict()\n",
    "for w in sorted(text_dictionary, key=text_dictionary.get, reverse=True):\n",
    "    sorted_text_dictionary[w] = text_dictionary[w]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "i =1\n",
    "for k, v in sorted_text_dictionary.items():\n",
    "    sorted_text_dictionary[k] = i\n",
    "    i = i+1\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████████| 3321/3321 [03:39<00:00, 15.16it/s]\n"
     ]
    }
   ],
   "source": [
    "i =0\n",
    "for text_data in tqdm(result['TEXT'].values):\n",
    "    words = text_data.split()\n",
    "    result['TEXT'][i] = ' '.join(str(sorted_text_dictionary.get(word, word)) for word in words)\n",
    "    i = i+1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████████| 3321/3321 [03:32<00:00, 15.62it/s]\n"
     ]
    }
   ],
   "source": [
    "i = 0\n",
    "for numbers in tqdm(result['TEXT'].values):\n",
    "    list = []\n",
    "    words = numbers.split(' ')\n",
    "    for word in words:\n",
    "        try:\n",
    "            if int(word)<2000:\n",
    "                list.append(word)\n",
    "        except ValueError:\n",
    "            continue\n",
    "    result['TEXT'][i] = ' '.join(list)\n",
    "    i = i+1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "652"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len((result['TEXT'][30]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LabelBinarizer(neg_label=0, pos_label=1, sparse_output=False)"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn import preprocessing\n",
    "lb = preprocessing.LabelBinarizer()\n",
    "lb.fit(result['Class'].values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 2, 3, 4, 5, 6, 7, 8, 9], dtype=int64)"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lb.classes_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "Y = lb.transform(result['Class'].values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{1, 2, 3, 4, 5, 6, 7, 8, 9}"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "set(result['Class'].values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 0, 0, 0, 0, 0, 0, 0, 0])"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Y[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for i in range(result.shape[0]):\n",
    "    result['TEXT'][i] = result['TEXT'][i].split()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train, X_test, Y_train, Y_test = train_test_split(result['TEXT'], Y, test_size=0.2 )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(2656, 5000)\n",
      "[   0    0    0 ...   14   73 1588]\n"
     ]
    }
   ],
   "source": [
    "# truncate and/or pad input sequences\n",
    "max_review_length = 5000\n",
    "X_train = sequence.pad_sequences(X_train, maxlen=max_review_length)\n",
    "X_test = sequence.pad_sequences(X_test, maxlen=max_review_length)\n",
    "\n",
    "print(X_train.shape)\n",
    "print(X_train[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'numpy.ndarray'>\n",
      "5000\n"
     ]
    }
   ],
   "source": [
    "print(type(X_train[1]))\n",
    "print(len(X_train[1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_3 (Embedding)      (None, 5000, 32)          64000     \n",
      "_________________________________________________________________\n",
      "lstm_5 (LSTM)                (None, 100)               53200     \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 9)                 909       \n",
      "=================================================================\n",
      "Total params: 118,109\n",
      "Trainable params: 118,109\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "top_words = 2000\n",
    "# create the model\n",
    "embedding_vector_length = 32\n",
    "model = Sequential()\n",
    "model.add(Embedding(top_words, embedding_vector_length, input_length=5000))\n",
    "model.add(LSTM(100,dropout=0.4))\n",
    "model.add(Dense(9, activation='softmax'))\n",
    "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "print(model.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "2656/2656 [==============================] - ETA: 17:49 - loss: 2.1891 - acc: 0.24 - ETA: 15:13 - loss: 2.1862 - acc: 0.23 - ETA: 13:18 - loss: 2.1825 - acc: 0.26 - ETA: 11:51 - loss: 2.1796 - acc: 0.26 - ETA: 10:40 - loss: 2.1745 - acc: 0.27 - ETA: 9:43 - loss: 2.1710 - acc: 0.2656 - ETA: 8:54 - loss: 2.1661 - acc: 0.262 - ETA: 8:12 - loss: 2.1569 - acc: 0.265 - ETA: 7:36 - loss: 2.1463 - acc: 0.270 - ETA: 7:04 - loss: 2.1332 - acc: 0.265 - ETA: 6:35 - loss: 2.1074 - acc: 0.272 - ETA: 6:09 - loss: 2.1023 - acc: 0.269 - ETA: 5:46 - loss: 2.0875 - acc: 0.273 - ETA: 5:25 - loss: 2.0749 - acc: 0.279 - ETA: 5:04 - loss: 2.0616 - acc: 0.282 - ETA: 4:45 - loss: 2.0523 - acc: 0.278 - ETA: 4:26 - loss: 2.0408 - acc: 0.278 - ETA: 4:09 - loss: 2.0335 - acc: 0.276 - ETA: 3:52 - loss: 2.0252 - acc: 0.275 - ETA: 3:36 - loss: 2.0198 - acc: 0.272 - ETA: 3:21 - loss: 2.0107 - acc: 0.271 - ETA: 3:06 - loss: 2.0014 - acc: 0.269 - ETA: 2:52 - loss: 1.9952 - acc: 0.265 - ETA: 2:38 - loss: 1.9900 - acc: 0.264 - ETA: 2:25 - loss: 1.9838 - acc: 0.263 - ETA: 2:12 - loss: 1.9784 - acc: 0.265 - ETA: 1:59 - loss: 1.9748 - acc: 0.264 - ETA: 1:47 - loss: 1.9735 - acc: 0.263 - ETA: 1:35 - loss: 1.9662 - acc: 0.266 - ETA: 1:23 - loss: 1.9608 - acc: 0.263 - ETA: 1:11 - loss: 1.9592 - acc: 0.265 - ETA: 1:00 - loss: 1.9539 - acc: 0.266 - ETA: 48s - loss: 1.9527 - acc: 0.265 - ETA: 37s - loss: 1.9480 - acc: 0.26 - ETA: 26s - loss: 1.9450 - acc: 0.26 - ETA: 15s - loss: 1.9447 - acc: 0.26 - ETA: 5s - loss: 1.9408 - acc: 0.2675 - 438s 165ms/step - loss: 1.9380 - acc: 0.2688\n",
      "Epoch 2/10\n",
      "2656/2656 [==============================] - ETA: 5:15 - loss: 1.9704 - acc: 0.250 - ETA: 5:09 - loss: 1.8763 - acc: 0.296 - ETA: 5:00 - loss: 1.8393 - acc: 0.281 - ETA: 4:57 - loss: 1.8192 - acc: 0.285 - ETA: 4:50 - loss: 1.8239 - acc: 0.281 - ETA: 4:44 - loss: 1.8298 - acc: 0.273 - ETA: 4:37 - loss: 1.8270 - acc: 0.279 - ETA: 4:31 - loss: 1.8069 - acc: 0.283 - ETA: 4:25 - loss: 1.8086 - acc: 0.293 - ETA: 4:18 - loss: 1.7960 - acc: 0.298 - ETA: 4:10 - loss: 1.8018 - acc: 0.295 - ETA: 4:03 - loss: 1.8040 - acc: 0.294 - ETA: 3:55 - loss: 1.7944 - acc: 0.290 - ETA: 3:48 - loss: 1.7856 - acc: 0.294 - ETA: 3:40 - loss: 1.7880 - acc: 0.296 - ETA: 3:32 - loss: 1.7801 - acc: 0.300 - ETA: 3:25 - loss: 1.7767 - acc: 0.298 - ETA: 3:17 - loss: 1.7772 - acc: 0.296 - ETA: 3:09 - loss: 1.7725 - acc: 0.295 - ETA: 3:00 - loss: 1.7750 - acc: 0.293 - ETA: 2:52 - loss: 1.7730 - acc: 0.296 - ETA: 2:44 - loss: 1.7688 - acc: 0.303 - ETA: 2:36 - loss: 1.7689 - acc: 0.307 - ETA: 2:28 - loss: 1.7670 - acc: 0.313 - ETA: 2:20 - loss: 1.7560 - acc: 0.320 - ETA: 2:11 - loss: 1.7681 - acc: 0.319 - ETA: 2:03 - loss: 1.7739 - acc: 0.319 - ETA: 1:55 - loss: 1.7750 - acc: 0.320 - ETA: 1:47 - loss: 1.7778 - acc: 0.320 - ETA: 1:38 - loss: 1.7751 - acc: 0.322 - ETA: 1:30 - loss: 1.7748 - acc: 0.326 - ETA: 1:21 - loss: 1.7723 - acc: 0.326 - ETA: 1:13 - loss: 1.7709 - acc: 0.328 - ETA: 1:04 - loss: 1.7689 - acc: 0.327 - ETA: 56s - loss: 1.7641 - acc: 0.331 - ETA: 47s - loss: 1.7644 - acc: 0.33 - ETA: 38s - loss: 1.7594 - acc: 0.33 - ETA: 30s - loss: 1.7563 - acc: 0.33 - ETA: 21s - loss: 1.7547 - acc: 0.33 - ETA: 13s - loss: 1.7522 - acc: 0.33 - ETA: 4s - loss: 1.7498 - acc: 0.3399 - 360s 136ms/step - loss: 1.7515 - acc: 0.3389\n",
      "Epoch 3/10\n",
      "2656/2656 [==============================] - ETA: 5:51 - loss: 1.6860 - acc: 0.375 - ETA: 5:39 - loss: 1.6220 - acc: 0.367 - ETA: 5:32 - loss: 1.6381 - acc: 0.390 - ETA: 5:22 - loss: 1.6594 - acc: 0.375 - ETA: 5:16 - loss: 1.7152 - acc: 0.340 - ETA: 5:08 - loss: 1.6835 - acc: 0.346 - ETA: 5:02 - loss: 1.6678 - acc: 0.354 - ETA: 4:54 - loss: 1.6704 - acc: 0.351 - ETA: 4:45 - loss: 1.6571 - acc: 0.368 - ETA: 4:37 - loss: 1.6456 - acc: 0.384 - ETA: 4:28 - loss: 1.6425 - acc: 0.390 - ETA: 4:20 - loss: 1.6474 - acc: 0.394 - ETA: 4:11 - loss: 1.6400 - acc: 0.399 - ETA: 4:02 - loss: 1.6302 - acc: 0.411 - ETA: 3:54 - loss: 1.6326 - acc: 0.406 - ETA: 3:45 - loss: 1.6328 - acc: 0.410 - ETA: 3:37 - loss: 1.6342 - acc: 0.409 - ETA: 3:28 - loss: 1.6255 - acc: 0.412 - ETA: 3:19 - loss: 1.6254 - acc: 0.411 - ETA: 3:11 - loss: 1.6177 - acc: 0.411 - ETA: 3:02 - loss: 1.6169 - acc: 0.413 - ETA: 2:53 - loss: 1.6063 - acc: 0.417 - ETA: 2:44 - loss: 1.5987 - acc: 0.419 - ETA: 2:35 - loss: 1.5925 - acc: 0.425 - ETA: 2:27 - loss: 1.5757 - acc: 0.436 - ETA: 2:18 - loss: 1.5755 - acc: 0.438 - ETA: 2:09 - loss: 1.5699 - acc: 0.438 - ETA: 2:00 - loss: 1.5662 - acc: 0.438 - ETA: 1:52 - loss: 1.5651 - acc: 0.440 - ETA: 1:43 - loss: 1.5600 - acc: 0.439 - ETA: 1:34 - loss: 1.5494 - acc: 0.444 - ETA: 1:25 - loss: 1.5423 - acc: 0.448 - ETA: 1:16 - loss: 1.5404 - acc: 0.449 - ETA: 1:07 - loss: 1.5390 - acc: 0.449 - ETA: 58s - loss: 1.5361 - acc: 0.452 - ETA: 49s - loss: 1.5386 - acc: 0.45 - ETA: 40s - loss: 1.5363 - acc: 0.45 - ETA: 31s - loss: 1.5357 - acc: 0.45 - ETA: 22s - loss: 1.5361 - acc: 0.45 - ETA: 13s - loss: 1.5322 - acc: 0.45 - ETA: 4s - loss: 1.5277 - acc: 0.4577 - 376s 141ms/step - loss: 1.5260 - acc: 0.4593\n",
      "Epoch 4/10\n",
      "2656/2656 [==============================] - ETA: 5:51 - loss: 1.2523 - acc: 0.640 - ETA: 5:40 - loss: 1.2977 - acc: 0.578 - ETA: 5:35 - loss: 1.3142 - acc: 0.536 - ETA: 5:29 - loss: 1.3391 - acc: 0.519 - ETA: 5:21 - loss: 1.3943 - acc: 0.496 - ETA: 5:16 - loss: 1.3830 - acc: 0.494 - ETA: 5:10 - loss: 1.3558 - acc: 0.506 - ETA: 5:02 - loss: 1.3356 - acc: 0.515 - ETA: 4:55 - loss: 1.3297 - acc: 0.522 - ETA: 4:46 - loss: 1.3435 - acc: 0.517 - ETA: 4:36 - loss: 1.3532 - acc: 0.512 - ETA: 4:27 - loss: 1.3458 - acc: 0.519 - ETA: 4:18 - loss: 1.3489 - acc: 0.513 - ETA: 4:09 - loss: 1.3419 - acc: 0.517 - ETA: 4:00 - loss: 1.3360 - acc: 0.515 - ETA: 3:51 - loss: 1.3415 - acc: 0.512 - ETA: 3:42 - loss: 1.3461 - acc: 0.511 - ETA: 3:33 - loss: 1.3367 - acc: 0.516 - ETA: 3:24 - loss: 1.3331 - acc: 0.518 - ETA: 3:15 - loss: 1.3351 - acc: 0.520 - ETA: 3:06 - loss: 1.3266 - acc: 0.521 - ETA: 2:57 - loss: 1.3197 - acc: 0.522 - ETA: 2:48 - loss: 1.3195 - acc: 0.523 - ETA: 2:39 - loss: 1.3169 - acc: 0.524 - ETA: 2:30 - loss: 1.3168 - acc: 0.525 - ETA: 2:21 - loss: 1.3163 - acc: 0.523 - ETA: 2:12 - loss: 1.3131 - acc: 0.526 - ETA: 2:02 - loss: 1.3147 - acc: 0.529 - ETA: 1:53 - loss: 1.3126 - acc: 0.530 - ETA: 1:44 - loss: 1.3218 - acc: 0.529 - ETA: 1:35 - loss: 1.3182 - acc: 0.529 - ETA: 1:26 - loss: 1.3235 - acc: 0.529 - ETA: 1:17 - loss: 1.3277 - acc: 0.526 - ETA: 1:07 - loss: 1.3230 - acc: 0.528 - ETA: 58s - loss: 1.3215 - acc: 0.531 - ETA: 49s - loss: 1.3184 - acc: 0.53 - ETA: 40s - loss: 1.3165 - acc: 0.53 - ETA: 31s - loss: 1.3132 - acc: 0.53 - ETA: 22s - loss: 1.3114 - acc: 0.53 - ETA: 13s - loss: 1.3089 - acc: 0.53 - ETA: 4s - loss: 1.3139 - acc: 0.5366 - 375s 141ms/step - loss: 1.3148 - acc: 0.5369\n",
      "Epoch 5/10\n",
      "2656/2656 [==============================] - ETA: 5:41 - loss: 0.9514 - acc: 0.703 - ETA: 5:37 - loss: 1.0333 - acc: 0.679 - ETA: 5:32 - loss: 1.0723 - acc: 0.640 - ETA: 5:22 - loss: 1.1016 - acc: 0.628 - ETA: 5:15 - loss: 1.0981 - acc: 0.618 - ETA: 5:07 - loss: 1.1358 - acc: 0.604 - ETA: 4:59 - loss: 1.1428 - acc: 0.604 - ETA: 4:52 - loss: 1.1205 - acc: 0.617 - ETA: 4:44 - loss: 1.1444 - acc: 0.605 - ETA: 4:37 - loss: 1.1396 - acc: 0.603 - ETA: 4:29 - loss: 1.1502 - acc: 0.599 - ETA: 4:20 - loss: 1.1410 - acc: 0.606 - ETA: 4:12 - loss: 1.1230 - acc: 0.613 - ETA: 4:03 - loss: 1.1309 - acc: 0.611 - ETA: 3:54 - loss: 1.1372 - acc: 0.610 - ETA: 3:45 - loss: 1.1534 - acc: 0.599 - ETA: 3:37 - loss: 1.1441 - acc: 0.603 - ETA: 3:28 - loss: 1.1479 - acc: 0.604 - ETA: 3:19 - loss: 1.1514 - acc: 0.599 - ETA: 3:11 - loss: 1.1573 - acc: 0.596 - ETA: 3:02 - loss: 1.1630 - acc: 0.593 - ETA: 2:53 - loss: 1.1560 - acc: 0.597 - ETA: 2:45 - loss: 1.1555 - acc: 0.597 - ETA: 2:36 - loss: 1.1606 - acc: 0.592 - ETA: 2:28 - loss: 1.1629 - acc: 0.591 - ETA: 2:19 - loss: 1.1601 - acc: 0.594 - ETA: 2:10 - loss: 1.1557 - acc: 0.592 - ETA: 2:01 - loss: 1.1613 - acc: 0.592 - ETA: 1:53 - loss: 1.1647 - acc: 0.590 - ETA: 1:44 - loss: 1.1663 - acc: 0.588 - ETA: 1:35 - loss: 1.1665 - acc: 0.588 - ETA: 1:26 - loss: 1.1638 - acc: 0.590 - ETA: 1:17 - loss: 1.1593 - acc: 0.593 - ETA: 1:08 - loss: 1.1623 - acc: 0.591 - ETA: 59s - loss: 1.1592 - acc: 0.592 - ETA: 50s - loss: 1.1602 - acc: 0.59 - ETA: 41s - loss: 1.1533 - acc: 0.59 - ETA: 32s - loss: 1.1448 - acc: 0.59 - ETA: 22s - loss: 1.1500 - acc: 0.59 - ETA: 13s - loss: 1.1521 - acc: 0.59 - ETA: 4s - loss: 1.1508 - acc: 0.5957 - 380s 143ms/step - loss: 1.1502 - acc: 0.5953\n",
      "Epoch 6/10\n",
      "2656/2656 [==============================] - ETA: 6:18 - loss: 1.2775 - acc: 0.546 - ETA: 6:12 - loss: 1.0812 - acc: 0.656 - ETA: 6:04 - loss: 1.1048 - acc: 0.630 - ETA: 5:53 - loss: 1.0600 - acc: 0.652 - ETA: 5:45 - loss: 1.1128 - acc: 0.634 - ETA: 5:35 - loss: 1.1042 - acc: 0.635 - ETA: 5:26 - loss: 1.1458 - acc: 0.627 - ETA: 5:16 - loss: 1.1296 - acc: 0.632 - ETA: 5:07 - loss: 1.1168 - acc: 0.631 - ETA: 4:58 - loss: 1.1243 - acc: 0.620 - ETA: 4:49 - loss: 1.1079 - acc: 0.627 - ETA: 4:40 - loss: 1.0983 - acc: 0.626 - ETA: 4:30 - loss: 1.1024 - acc: 0.623 - ETA: 4:21 - loss: 1.0950 - acc: 0.623 - ETA: 4:12 - loss: 1.0958 - acc: 0.624 - ETA: 4:02 - loss: 1.0994 - acc: 0.618 - ETA: 3:53 - loss: 1.0817 - acc: 0.620 - ETA: 3:44 - loss: 1.0807 - acc: 0.617 - ETA: 3:34 - loss: 1.0796 - acc: 0.616 - ETA: 3:24 - loss: 1.0854 - acc: 0.613 - ETA: 3:15 - loss: 1.0896 - acc: 0.610 - ETA: 3:05 - loss: 1.0831 - acc: 0.612 - ETA: 2:56 - loss: 1.0886 - acc: 0.610 - ETA: 2:46 - loss: 1.0895 - acc: 0.609 - ETA: 2:37 - loss: 1.0886 - acc: 0.609 - ETA: 2:27 - loss: 1.0774 - acc: 0.614 - ETA: 2:18 - loss: 1.0736 - acc: 0.615 - ETA: 2:09 - loss: 1.0777 - acc: 0.613 - ETA: 2:00 - loss: 1.0734 - acc: 0.615 - ETA: 1:51 - loss: 1.0691 - acc: 0.616 - ETA: 1:41 - loss: 1.0651 - acc: 0.616 - ETA: 1:32 - loss: 1.0629 - acc: 0.619 - ETA: 1:22 - loss: 1.0666 - acc: 0.617 - ETA: 1:12 - loss: 1.0672 - acc: 0.619 - ETA: 1:03 - loss: 1.0657 - acc: 0.620 - ETA: 53s - loss: 1.0663 - acc: 0.619 - ETA: 44s - loss: 1.0632 - acc: 0.62 - ETA: 34s - loss: 1.0670 - acc: 0.61 - ETA: 24s - loss: 1.0645 - acc: 0.62 - ETA: 14s - loss: 1.0634 - acc: 0.62 - ETA: 4s - loss: 1.0602 - acc: 0.6239 - 413s 156ms/step - loss: 1.0562 - acc: 0.6254\n",
      "Epoch 7/10\n",
      "2656/2656 [==============================] - ETA: 7:32 - loss: 0.8642 - acc: 0.671 - ETA: 7:19 - loss: 0.8160 - acc: 0.710 - ETA: 6:48 - loss: 0.8778 - acc: 0.692 - ETA: 6:27 - loss: 0.9027 - acc: 0.695 - ETA: 6:17 - loss: 0.9442 - acc: 0.678 - ETA: 6:07 - loss: 0.9478 - acc: 0.674 - ETA: 6:01 - loss: 0.9704 - acc: 0.651 - ETA: 5:53 - loss: 0.9590 - acc: 0.646 - ETA: 5:42 - loss: 0.9720 - acc: 0.635 - ETA: 5:29 - loss: 0.9610 - acc: 0.642 - ETA: 5:21 - loss: 0.9627 - acc: 0.642 - ETA: 5:15 - loss: 0.9664 - acc: 0.641 - ETA: 5:02 - loss: 0.9603 - acc: 0.645 - ETA: 4:51 - loss: 0.9621 - acc: 0.642 - ETA: 4:39 - loss: 0.9521 - acc: 0.652 - ETA: 4:28 - loss: 0.9567 - acc: 0.650 - ETA: 4:20 - loss: 0.9553 - acc: 0.655 - ETA: 4:09 - loss: 0.9543 - acc: 0.658 - ETA: 3:58 - loss: 0.9563 - acc: 0.662 - ETA: 3:48 - loss: 0.9544 - acc: 0.663 - ETA: 3:37 - loss: 0.9508 - acc: 0.666 - ETA: 3:26 - loss: 0.9423 - acc: 0.670 - ETA: 3:16 - loss: 0.9377 - acc: 0.673 - ETA: 3:05 - loss: 0.9438 - acc: 0.670 - ETA: 2:54 - loss: 0.9376 - acc: 0.670 - ETA: 2:43 - loss: 0.9417 - acc: 0.665 - ETA: 2:34 - loss: 0.9486 - acc: 0.664 - ETA: 2:23 - loss: 0.9589 - acc: 0.660 - ETA: 2:12 - loss: 0.9654 - acc: 0.657 - ETA: 2:02 - loss: 0.9650 - acc: 0.656 - ETA: 1:51 - loss: 0.9611 - acc: 0.658 - ETA: 1:40 - loss: 0.9554 - acc: 0.660 - ETA: 1:30 - loss: 0.9544 - acc: 0.660 - ETA: 1:20 - loss: 0.9581 - acc: 0.659 - ETA: 1:09 - loss: 0.9590 - acc: 0.659 - ETA: 58s - loss: 0.9601 - acc: 0.659 - ETA: 47s - loss: 0.9553 - acc: 0.66 - ETA: 37s - loss: 0.9589 - acc: 0.66 - ETA: 26s - loss: 0.9618 - acc: 0.65 - ETA: 15s - loss: 0.9690 - acc: 0.65 - ETA: 5s - loss: 0.9744 - acc: 0.6574 - 441s 166ms/step - loss: 0.9761 - acc: 0.6559\n",
      "Epoch 8/10\n",
      "2656/2656 [==============================] - ETA: 6:35 - loss: 0.9163 - acc: 0.718 - ETA: 6:49 - loss: 0.8210 - acc: 0.734 - ETA: 6:30 - loss: 0.8389 - acc: 0.697 - ETA: 6:18 - loss: 0.8469 - acc: 0.691 - ETA: 6:06 - loss: 0.8592 - acc: 0.684 - ETA: 5:55 - loss: 0.8625 - acc: 0.692 - ETA: 5:43 - loss: 0.8736 - acc: 0.687 - ETA: 5:38 - loss: 0.9269 - acc: 0.673 - ETA: 5:28 - loss: 0.9299 - acc: 0.671 - ETA: 5:20 - loss: 0.9121 - acc: 0.682 - ETA: 5:12 - loss: 0.9056 - acc: 0.683 - ETA: 5:04 - loss: 0.8980 - acc: 0.687 - ETA: 4:54 - loss: 0.9036 - acc: 0.683 - ETA: 4:45 - loss: 0.8955 - acc: 0.687 - ETA: 4:33 - loss: 0.8853 - acc: 0.695 - ETA: 4:22 - loss: 0.8923 - acc: 0.692 - ETA: 4:11 - loss: 0.8999 - acc: 0.689 - ETA: 4:00 - loss: 0.8959 - acc: 0.685 - ETA: 3:50 - loss: 0.8879 - acc: 0.689 - ETA: 3:40 - loss: 0.8827 - acc: 0.689 - ETA: 3:29 - loss: 0.8782 - acc: 0.694 - ETA: 3:19 - loss: 0.8801 - acc: 0.691 - ETA: 3:09 - loss: 0.8817 - acc: 0.691 - ETA: 2:59 - loss: 0.8813 - acc: 0.692 - ETA: 2:50 - loss: 0.8821 - acc: 0.690 - ETA: 2:40 - loss: 0.8844 - acc: 0.690 - ETA: 2:30 - loss: 0.8787 - acc: 0.692 - ETA: 2:20 - loss: 0.8803 - acc: 0.689 - ETA: 2:10 - loss: 0.8727 - acc: 0.691 - ETA: 2:00 - loss: 0.8732 - acc: 0.691 - ETA: 1:50 - loss: 0.8791 - acc: 0.688 - ETA: 1:39 - loss: 0.8808 - acc: 0.687 - ETA: 1:28 - loss: 0.8864 - acc: 0.686 - ETA: 1:18 - loss: 0.8848 - acc: 0.687 - ETA: 1:07 - loss: 0.8885 - acc: 0.687 - ETA: 57s - loss: 0.8859 - acc: 0.688 - ETA: 46s - loss: 0.8856 - acc: 0.68 - ETA: 36s - loss: 0.8809 - acc: 0.69 - ETA: 26s - loss: 0.8806 - acc: 0.69 - ETA: 15s - loss: 0.8834 - acc: 0.68 - ETA: 5s - loss: 0.8849 - acc: 0.6886 - 431s 162ms/step - loss: 0.8852 - acc: 0.6879\n",
      "Epoch 9/10\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2656/2656 [==============================] - ETA: 7:31 - loss: 0.7328 - acc: 0.718 - ETA: 8:11 - loss: 0.7636 - acc: 0.718 - ETA: 7:27 - loss: 0.8727 - acc: 0.682 - ETA: 7:02 - loss: 0.8791 - acc: 0.675 - ETA: 6:43 - loss: 0.8714 - acc: 0.687 - ETA: 6:36 - loss: 0.8717 - acc: 0.687 - ETA: 6:21 - loss: 0.8575 - acc: 0.694 - ETA: 6:02 - loss: 0.8371 - acc: 0.705 - ETA: 5:49 - loss: 0.8439 - acc: 0.708 - ETA: 5:42 - loss: 0.8410 - acc: 0.714 - ETA: 5:33 - loss: 0.8441 - acc: 0.715 - ETA: 5:19 - loss: 0.8405 - acc: 0.714 - ETA: 5:05 - loss: 0.8429 - acc: 0.712 - ETA: 4:54 - loss: 0.8306 - acc: 0.716 - ETA: 4:42 - loss: 0.8307 - acc: 0.712 - ETA: 4:32 - loss: 0.8233 - acc: 0.715 - ETA: 4:22 - loss: 0.8262 - acc: 0.717 - ETA: 4:10 - loss: 0.8313 - acc: 0.716 - ETA: 3:58 - loss: 0.8371 - acc: 0.715 - ETA: 3:46 - loss: 0.8342 - acc: 0.711 - ETA: 3:35 - loss: 0.8294 - acc: 0.710 - ETA: 3:25 - loss: 0.8282 - acc: 0.712 - ETA: 3:15 - loss: 0.8260 - acc: 0.714 - ETA: 3:05 - loss: 0.8316 - acc: 0.709 - ETA: 2:55 - loss: 0.8280 - acc: 0.710 - ETA: 2:44 - loss: 0.8294 - acc: 0.709 - ETA: 2:32 - loss: 0.8357 - acc: 0.707 - ETA: 2:22 - loss: 0.8393 - acc: 0.705 - ETA: 2:11 - loss: 0.8381 - acc: 0.705 - ETA: 2:00 - loss: 0.8453 - acc: 0.703 - ETA: 1:49 - loss: 0.8445 - acc: 0.703 - ETA: 1:38 - loss: 0.8408 - acc: 0.704 - ETA: 1:28 - loss: 0.8444 - acc: 0.704 - ETA: 1:17 - loss: 0.8387 - acc: 0.707 - ETA: 1:07 - loss: 0.8345 - acc: 0.707 - ETA: 56s - loss: 0.8343 - acc: 0.707 - ETA: 46s - loss: 0.8413 - acc: 0.70 - ETA: 36s - loss: 0.8427 - acc: 0.70 - ETA: 25s - loss: 0.8460 - acc: 0.70 - ETA: 15s - loss: 0.8434 - acc: 0.70 - ETA: 5s - loss: 0.8437 - acc: 0.7050 - 426s 160ms/step - loss: 0.8453 - acc: 0.7048\n",
      "Epoch 10/10\n",
      "2656/2656 [==============================] - ETA: 6:16 - loss: 0.5932 - acc: 0.828 - ETA: 6:15 - loss: 0.6998 - acc: 0.773 - ETA: 6:03 - loss: 0.7572 - acc: 0.744 - ETA: 5:54 - loss: 0.7220 - acc: 0.765 - ETA: 5:44 - loss: 0.7495 - acc: 0.753 - ETA: 5:35 - loss: 0.7598 - acc: 0.742 - ETA: 5:24 - loss: 0.7943 - acc: 0.734 - ETA: 5:14 - loss: 0.7967 - acc: 0.732 - ETA: 5:04 - loss: 0.7831 - acc: 0.732 - ETA: 4:55 - loss: 0.7959 - acc: 0.725 - ETA: 4:46 - loss: 0.7973 - acc: 0.725 - ETA: 4:37 - loss: 0.7875 - acc: 0.734 - ETA: 4:28 - loss: 0.8050 - acc: 0.724 - ETA: 4:18 - loss: 0.8154 - acc: 0.719 - ETA: 4:09 - loss: 0.8130 - acc: 0.720 - ETA: 4:00 - loss: 0.8120 - acc: 0.724 - ETA: 3:52 - loss: 0.8118 - acc: 0.724 - ETA: 3:42 - loss: 0.8009 - acc: 0.727 - ETA: 3:33 - loss: 0.7938 - acc: 0.731 - ETA: 3:24 - loss: 0.7873 - acc: 0.734 - ETA: 3:15 - loss: 0.7813 - acc: 0.737 - ETA: 3:06 - loss: 0.7787 - acc: 0.740 - ETA: 2:57 - loss: 0.7840 - acc: 0.737 - ETA: 2:49 - loss: 0.7840 - acc: 0.736 - ETA: 2:39 - loss: 0.7832 - acc: 0.735 - ETA: 2:30 - loss: 0.7818 - acc: 0.736 - ETA: 2:20 - loss: 0.7820 - acc: 0.734 - ETA: 2:11 - loss: 0.7787 - acc: 0.735 - ETA: 2:01 - loss: 0.7825 - acc: 0.732 - ETA: 1:52 - loss: 0.7798 - acc: 0.732 - ETA: 1:43 - loss: 0.7769 - acc: 0.732 - ETA: 1:33 - loss: 0.7796 - acc: 0.730 - ETA: 1:23 - loss: 0.7848 - acc: 0.729 - ETA: 1:14 - loss: 0.7845 - acc: 0.728 - ETA: 1:04 - loss: 0.7815 - acc: 0.729 - ETA: 54s - loss: 0.7780 - acc: 0.731 - ETA: 44s - loss: 0.7748 - acc: 0.73 - ETA: 34s - loss: 0.7722 - acc: 0.73 - ETA: 24s - loss: 0.7676 - acc: 0.73 - ETA: 14s - loss: 0.7642 - acc: 0.73 - ETA: 5s - loss: 0.7644 - acc: 0.7359 - 416s 157ms/step - loss: 0.7679 - acc: 0.7353\n",
      "Accuracy: 56.84%\n"
     ]
    }
   ],
   "source": [
    "model.fit(X_train, Y_train, nb_epoch=10, batch_size=64)\n",
    "# Final evaluation of the model\n",
    "scores = model.evaluate(X_test, Y_test, verbose=0)\n",
    "print(\"Accuracy: %.2f%%\" % (scores[1]*100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "2656/2656 [==============================] - ETA: 4:41 - loss: 2.1984 - acc: 0.156 - ETA: 4:14 - loss: 2.1961 - acc: 0.179 - ETA: 4:02 - loss: 2.1937 - acc: 0.218 - ETA: 3:56 - loss: 2.1920 - acc: 0.234 - ETA: 3:50 - loss: 2.1900 - acc: 0.246 - ETA: 3:43 - loss: 2.1873 - acc: 0.260 - ETA: 3:37 - loss: 2.1854 - acc: 0.254 - ETA: 3:30 - loss: 2.1819 - acc: 0.257 - ETA: 3:23 - loss: 2.1787 - acc: 0.260 - ETA: 3:16 - loss: 2.1746 - acc: 0.265 - ETA: 3:10 - loss: 2.1708 - acc: 0.268 - ETA: 3:03 - loss: 2.1632 - acc: 0.277 - ETA: 2:57 - loss: 2.1545 - acc: 0.276 - ETA: 2:51 - loss: 2.1450 - acc: 0.275 - ETA: 2:45 - loss: 2.1312 - acc: 0.275 - ETA: 2:38 - loss: 2.1178 - acc: 0.277 - ETA: 2:32 - loss: 2.1066 - acc: 0.279 - ETA: 2:26 - loss: 2.0976 - acc: 0.276 - ETA: 2:20 - loss: 2.0829 - acc: 0.276 - ETA: 2:14 - loss: 2.0764 - acc: 0.272 - ETA: 2:08 - loss: 2.0688 - acc: 0.272 - ETA: 2:01 - loss: 2.0610 - acc: 0.274 - ETA: 1:55 - loss: 2.0530 - acc: 0.276 - ETA: 1:49 - loss: 2.0415 - acc: 0.281 - ETA: 1:43 - loss: 2.0292 - acc: 0.281 - ETA: 1:37 - loss: 2.0244 - acc: 0.277 - ETA: 1:31 - loss: 2.0169 - acc: 0.273 - ETA: 1:25 - loss: 2.0076 - acc: 0.272 - ETA: 1:18 - loss: 2.0032 - acc: 0.273 - ETA: 1:12 - loss: 1.9951 - acc: 0.273 - ETA: 1:06 - loss: 1.9914 - acc: 0.278 - ETA: 1:00 - loss: 1.9881 - acc: 0.279 - ETA: 53s - loss: 1.9826 - acc: 0.281 - ETA: 47s - loss: 1.9739 - acc: 0.28 - ETA: 41s - loss: 1.9701 - acc: 0.28 - ETA: 34s - loss: 1.9657 - acc: 0.28 - ETA: 28s - loss: 1.9583 - acc: 0.28 - ETA: 22s - loss: 1.9599 - acc: 0.28 - ETA: 15s - loss: 1.9564 - acc: 0.28 - ETA: 9s - loss: 1.9555 - acc: 0.2844 - ETA: 3s - loss: 1.9527 - acc: 0.285 - 266s 100ms/step - loss: 1.9517 - acc: 0.2843\n",
      "Epoch 2/50\n",
      "2656/2656 [==============================] - ETA: 4:34 - loss: 1.8228 - acc: 0.171 - ETA: 4:38 - loss: 1.8125 - acc: 0.257 - ETA: 4:29 - loss: 1.7806 - acc: 0.260 - ETA: 4:24 - loss: 1.8257 - acc: 0.234 - ETA: 4:19 - loss: 1.8244 - acc: 0.234 - ETA: 4:12 - loss: 1.8204 - acc: 0.250 - ETA: 4:05 - loss: 1.8165 - acc: 0.263 - ETA: 3:58 - loss: 1.8223 - acc: 0.271 - ETA: 3:51 - loss: 1.8215 - acc: 0.270 - ETA: 3:46 - loss: 1.8141 - acc: 0.276 - ETA: 3:39 - loss: 1.8013 - acc: 0.284 - ETA: 3:32 - loss: 1.8017 - acc: 0.282 - ETA: 3:24 - loss: 1.8042 - acc: 0.287 - ETA: 3:17 - loss: 1.7978 - acc: 0.286 - ETA: 3:10 - loss: 1.7986 - acc: 0.287 - ETA: 3:03 - loss: 1.7996 - acc: 0.286 - ETA: 2:56 - loss: 1.7986 - acc: 0.284 - ETA: 2:49 - loss: 1.8026 - acc: 0.279 - ETA: 2:42 - loss: 1.7989 - acc: 0.282 - ETA: 2:35 - loss: 1.7983 - acc: 0.287 - ETA: 2:27 - loss: 1.8061 - acc: 0.283 - ETA: 2:20 - loss: 1.7994 - acc: 0.286 - ETA: 2:13 - loss: 1.7975 - acc: 0.292 - ETA: 2:06 - loss: 1.7941 - acc: 0.296 - ETA: 1:59 - loss: 1.7938 - acc: 0.297 - ETA: 1:51 - loss: 1.7899 - acc: 0.299 - ETA: 1:44 - loss: 1.7868 - acc: 0.299 - ETA: 1:37 - loss: 1.7789 - acc: 0.305 - ETA: 1:30 - loss: 1.7748 - acc: 0.310 - ETA: 1:22 - loss: 1.7719 - acc: 0.311 - ETA: 1:15 - loss: 1.7657 - acc: 0.314 - ETA: 1:08 - loss: 1.7647 - acc: 0.312 - ETA: 1:01 - loss: 1.7639 - acc: 0.311 - ETA: 53s - loss: 1.7653 - acc: 0.310 - ETA: 46s - loss: 1.7638 - acc: 0.30 - ETA: 39s - loss: 1.7624 - acc: 0.31 - ETA: 32s - loss: 1.7571 - acc: 0.31 - ETA: 25s - loss: 1.7521 - acc: 0.31 - ETA: 18s - loss: 1.7506 - acc: 0.32 - ETA: 10s - loss: 1.7489 - acc: 0.32 - ETA: 3s - loss: 1.7440 - acc: 0.3285 - 299s 113ms/step - loss: 1.7441 - acc: 0.3283\n",
      "Epoch 3/50\n",
      "2656/2656 [==============================] - ETA: 4:37 - loss: 1.5765 - acc: 0.375 - ETA: 4:35 - loss: 1.5819 - acc: 0.398 - ETA: 4:30 - loss: 1.5584 - acc: 0.406 - ETA: 4:25 - loss: 1.5642 - acc: 0.406 - ETA: 4:18 - loss: 1.5727 - acc: 0.396 - ETA: 4:12 - loss: 1.5793 - acc: 0.403 - ETA: 4:05 - loss: 1.5678 - acc: 0.408 - ETA: 3:59 - loss: 1.5386 - acc: 0.421 - ETA: 3:52 - loss: 1.5176 - acc: 0.428 - ETA: 3:46 - loss: 1.5011 - acc: 0.439 - ETA: 3:39 - loss: 1.5250 - acc: 0.437 - ETA: 3:32 - loss: 1.5177 - acc: 0.438 - ETA: 3:25 - loss: 1.5158 - acc: 0.437 - ETA: 3:18 - loss: 1.5081 - acc: 0.445 - ETA: 3:11 - loss: 1.5050 - acc: 0.451 - ETA: 3:04 - loss: 1.5123 - acc: 0.449 - ETA: 2:57 - loss: 1.5062 - acc: 0.449 - ETA: 2:50 - loss: 1.5068 - acc: 0.448 - ETA: 2:43 - loss: 1.5031 - acc: 0.448 - ETA: 2:35 - loss: 1.5007 - acc: 0.450 - ETA: 2:28 - loss: 1.4977 - acc: 0.449 - ETA: 2:21 - loss: 1.4959 - acc: 0.449 - ETA: 2:14 - loss: 1.4913 - acc: 0.453 - ETA: 2:07 - loss: 1.4883 - acc: 0.456 - ETA: 2:00 - loss: 1.4874 - acc: 0.456 - ETA: 1:52 - loss: 1.4821 - acc: 0.459 - ETA: 1:45 - loss: 1.4792 - acc: 0.461 - ETA: 1:38 - loss: 1.4710 - acc: 0.464 - ETA: 1:31 - loss: 1.4626 - acc: 0.467 - ETA: 1:23 - loss: 1.4866 - acc: 0.464 - ETA: 1:16 - loss: 1.4911 - acc: 0.466 - ETA: 1:09 - loss: 1.4906 - acc: 0.466 - ETA: 1:02 - loss: 1.4877 - acc: 0.468 - ETA: 54s - loss: 1.4805 - acc: 0.472 - ETA: 47s - loss: 1.4823 - acc: 0.47 - ETA: 40s - loss: 1.4835 - acc: 0.47 - ETA: 32s - loss: 1.4859 - acc: 0.47 - ETA: 25s - loss: 1.4803 - acc: 0.47 - ETA: 18s - loss: 1.4761 - acc: 0.47 - ETA: 10s - loss: 1.4705 - acc: 0.47 - ETA: 3s - loss: 1.4689 - acc: 0.4783 - 304s 114ms/step - loss: 1.4666 - acc: 0.4785\n",
      "Epoch 4/50\n",
      "2656/2656 [==============================] - ETA: 4:51 - loss: 1.2548 - acc: 0.531 - ETA: 4:46 - loss: 1.4808 - acc: 0.460 - ETA: 4:43 - loss: 1.4063 - acc: 0.479 - ETA: 4:39 - loss: 1.3586 - acc: 0.492 - ETA: 4:32 - loss: 1.3155 - acc: 0.503 - ETA: 4:22 - loss: 1.3310 - acc: 0.507 - ETA: 4:14 - loss: 1.3278 - acc: 0.508 - ETA: 4:06 - loss: 1.3442 - acc: 0.500 - ETA: 3:59 - loss: 1.3399 - acc: 0.503 - ETA: 3:52 - loss: 1.3218 - acc: 0.514 - ETA: 3:45 - loss: 1.3068 - acc: 0.518 - ETA: 3:38 - loss: 1.2890 - acc: 0.522 - ETA: 3:31 - loss: 1.2665 - acc: 0.527 - ETA: 3:23 - loss: 1.2690 - acc: 0.531 - ETA: 3:15 - loss: 1.2555 - acc: 0.533 - ETA: 3:08 - loss: 1.2634 - acc: 0.534 - ETA: 3:01 - loss: 1.2568 - acc: 0.531 - ETA: 2:53 - loss: 1.2643 - acc: 0.531 - ETA: 2:46 - loss: 1.2628 - acc: 0.528 - ETA: 2:39 - loss: 1.2531 - acc: 0.530 - ETA: 2:31 - loss: 1.2491 - acc: 0.533 - ETA: 2:24 - loss: 1.2499 - acc: 0.534 - ETA: 2:17 - loss: 1.2514 - acc: 0.538 - ETA: 2:09 - loss: 1.2543 - acc: 0.536 - ETA: 2:02 - loss: 1.2504 - acc: 0.540 - ETA: 1:55 - loss: 1.2392 - acc: 0.544 - ETA: 1:47 - loss: 1.2432 - acc: 0.546 - ETA: 1:40 - loss: 1.2424 - acc: 0.548 - ETA: 1:32 - loss: 1.2479 - acc: 0.544 - ETA: 1:25 - loss: 1.2458 - acc: 0.542 - ETA: 1:18 - loss: 1.2424 - acc: 0.542 - ETA: 1:10 - loss: 1.2424 - acc: 0.542 - ETA: 1:03 - loss: 1.2418 - acc: 0.541 - ETA: 55s - loss: 1.2393 - acc: 0.543 - ETA: 48s - loss: 1.2406 - acc: 0.54 - ETA: 41s - loss: 1.2362 - acc: 0.54 - ETA: 33s - loss: 1.2386 - acc: 0.54 - ETA: 26s - loss: 1.2386 - acc: 0.54 - ETA: 18s - loss: 1.2424 - acc: 0.54 - ETA: 11s - loss: 1.2443 - acc: 0.54 - ETA: 3s - loss: 1.2451 - acc: 0.5423 - 311s 117ms/step - loss: 1.2444 - acc: 0.5433\n",
      "Epoch 5/50\n",
      "2656/2656 [==============================] - ETA: 4:50 - loss: 1.3609 - acc: 0.562 - ETA: 4:55 - loss: 1.2608 - acc: 0.554 - ETA: 4:50 - loss: 1.2501 - acc: 0.546 - ETA: 4:44 - loss: 1.1994 - acc: 0.562 - ETA: 4:38 - loss: 1.1780 - acc: 0.575 - ETA: 4:34 - loss: 1.1693 - acc: 0.575 - ETA: 4:26 - loss: 1.1267 - acc: 0.600 - ETA: 4:19 - loss: 1.1217 - acc: 0.587 - ETA: 4:11 - loss: 1.1499 - acc: 0.569 - ETA: 4:04 - loss: 1.1420 - acc: 0.573 - ETA: 3:56 - loss: 1.1335 - acc: 0.573 - ETA: 3:48 - loss: 1.1270 - acc: 0.575 - ETA: 3:40 - loss: 1.1306 - acc: 0.573 - ETA: 3:33 - loss: 1.1384 - acc: 0.574 - ETA: 3:25 - loss: 1.1327 - acc: 0.578 - ETA: 3:17 - loss: 1.1351 - acc: 0.573 - ETA: 3:09 - loss: 1.1235 - acc: 0.576 - ETA: 3:01 - loss: 1.1265 - acc: 0.577 - ETA: 2:53 - loss: 1.1333 - acc: 0.575 - ETA: 2:45 - loss: 1.1318 - acc: 0.578 - ETA: 2:38 - loss: 1.1307 - acc: 0.578 - ETA: 2:30 - loss: 1.1346 - acc: 0.575 - ETA: 2:22 - loss: 1.1334 - acc: 0.577 - ETA: 2:14 - loss: 1.1303 - acc: 0.580 - ETA: 2:07 - loss: 1.1281 - acc: 0.583 - ETA: 1:59 - loss: 1.1314 - acc: 0.586 - ETA: 1:51 - loss: 1.1317 - acc: 0.586 - ETA: 1:43 - loss: 1.1328 - acc: 0.588 - ETA: 1:36 - loss: 1.1205 - acc: 0.593 - ETA: 1:28 - loss: 1.1216 - acc: 0.594 - ETA: 1:20 - loss: 1.1298 - acc: 0.590 - ETA: 1:13 - loss: 1.1349 - acc: 0.584 - ETA: 1:05 - loss: 1.1392 - acc: 0.582 - ETA: 57s - loss: 1.1387 - acc: 0.582 - ETA: 50s - loss: 1.1360 - acc: 0.58 - ETA: 42s - loss: 1.1353 - acc: 0.58 - ETA: 34s - loss: 1.1382 - acc: 0.58 - ETA: 26s - loss: 1.1396 - acc: 0.58 - ETA: 19s - loss: 1.1398 - acc: 0.58 - ETA: 11s - loss: 1.1385 - acc: 0.58 - ETA: 3s - loss: 1.1377 - acc: 0.5838 - 319s 120ms/step - loss: 1.1392 - acc: 0.5832\n",
      "Epoch 6/50\n",
      "2656/2656 [==============================] - ETA: 4:56 - loss: 1.0510 - acc: 0.656 - ETA: 4:52 - loss: 1.0148 - acc: 0.617 - ETA: 4:46 - loss: 0.9574 - acc: 0.635 - ETA: 4:39 - loss: 1.0103 - acc: 0.625 - ETA: 4:31 - loss: 0.9975 - acc: 0.628 - ETA: 4:23 - loss: 0.9752 - acc: 0.640 - ETA: 4:14 - loss: 0.9934 - acc: 0.640 - ETA: 4:06 - loss: 0.9741 - acc: 0.648 - ETA: 3:59 - loss: 0.9909 - acc: 0.642 - ETA: 3:51 - loss: 1.0047 - acc: 0.635 - ETA: 3:44 - loss: 1.0095 - acc: 0.636 - ETA: 3:37 - loss: 1.0105 - acc: 0.631 - ETA: 3:30 - loss: 1.0215 - acc: 0.625 - ETA: 3:23 - loss: 1.0274 - acc: 0.622 - ETA: 3:16 - loss: 1.0260 - acc: 0.619 - ETA: 3:08 - loss: 1.0310 - acc: 0.625 - ETA: 3:01 - loss: 1.0349 - acc: 0.626 - ETA: 2:54 - loss: 1.0366 - acc: 0.626 - ETA: 2:46 - loss: 1.0471 - acc: 0.623 - ETA: 2:39 - loss: 1.0465 - acc: 0.623 - ETA: 2:32 - loss: 1.0468 - acc: 0.624 - ETA: 2:25 - loss: 1.0397 - acc: 0.626 - ETA: 2:17 - loss: 1.0387 - acc: 0.629 - ETA: 2:10 - loss: 1.0405 - acc: 0.628 - ETA: 2:03 - loss: 1.0423 - acc: 0.623 - ETA: 1:55 - loss: 1.0359 - acc: 0.624 - ETA: 1:48 - loss: 1.0317 - acc: 0.624 - ETA: 1:41 - loss: 1.0337 - acc: 0.624 - ETA: 1:33 - loss: 1.0283 - acc: 0.626 - ETA: 1:26 - loss: 1.0281 - acc: 0.625 - ETA: 1:19 - loss: 1.0207 - acc: 0.627 - ETA: 1:11 - loss: 1.0161 - acc: 0.629 - ETA: 1:04 - loss: 1.0168 - acc: 0.628 - ETA: 56s - loss: 1.0110 - acc: 0.631 - ETA: 49s - loss: 1.0187 - acc: 0.62 - ETA: 41s - loss: 1.0185 - acc: 0.62 - ETA: 34s - loss: 1.0254 - acc: 0.62 - ETA: 26s - loss: 1.0206 - acc: 0.62 - ETA: 18s - loss: 1.0195 - acc: 0.62 - ETA: 11s - loss: 1.0254 - acc: 0.62 - ETA: 3s - loss: 1.0281 - acc: 0.6265 - 313s 118ms/step - loss: 1.0292 - acc: 0.6269\n",
      "Epoch 7/50\n",
      "2656/2656 [==============================] - ETA: 4:56 - loss: 0.8193 - acc: 0.796 - ETA: 4:56 - loss: 0.8107 - acc: 0.773 - ETA: 4:51 - loss: 0.8555 - acc: 0.734 - ETA: 4:46 - loss: 0.8774 - acc: 0.710 - ETA: 4:38 - loss: 0.9094 - acc: 0.690 - ETA: 4:31 - loss: 0.9000 - acc: 0.692 - ETA: 4:24 - loss: 0.9079 - acc: 0.680 - ETA: 4:17 - loss: 0.9008 - acc: 0.679 - ETA: 4:10 - loss: 0.9091 - acc: 0.682 - ETA: 4:03 - loss: 0.9099 - acc: 0.676 - ETA: 3:56 - loss: 0.9281 - acc: 0.671 - ETA: 3:48 - loss: 0.9403 - acc: 0.661 - ETA: 3:41 - loss: 0.9536 - acc: 0.657 - ETA: 3:33 - loss: 0.9397 - acc: 0.661 - ETA: 3:25 - loss: 0.9426 - acc: 0.660 - ETA: 3:17 - loss: 0.9436 - acc: 0.657 - ETA: 3:10 - loss: 0.9362 - acc: 0.662 - ETA: 3:02 - loss: 0.9308 - acc: 0.666 - ETA: 2:54 - loss: 0.9297 - acc: 0.671 - ETA: 2:46 - loss: 0.9322 - acc: 0.672 - ETA: 2:38 - loss: 0.9335 - acc: 0.673 - ETA: 2:31 - loss: 0.9435 - acc: 0.671 - ETA: 2:23 - loss: 0.9443 - acc: 0.673 - ETA: 2:15 - loss: 0.9368 - acc: 0.678 - ETA: 2:07 - loss: 0.9413 - acc: 0.676 - ETA: 2:00 - loss: 0.9432 - acc: 0.676 - ETA: 1:52 - loss: 0.9466 - acc: 0.674 - ETA: 1:44 - loss: 0.9429 - acc: 0.673 - ETA: 1:36 - loss: 0.9429 - acc: 0.673 - ETA: 1:29 - loss: 0.9551 - acc: 0.666 - ETA: 1:21 - loss: 0.9525 - acc: 0.666 - ETA: 1:13 - loss: 0.9550 - acc: 0.667 - ETA: 1:05 - loss: 0.9554 - acc: 0.667 - ETA: 58s - loss: 0.9527 - acc: 0.669 - ETA: 50s - loss: 0.9465 - acc: 0.67 - ETA: 42s - loss: 0.9468 - acc: 0.66 - ETA: 35s - loss: 0.9496 - acc: 0.66 - ETA: 27s - loss: 0.9446 - acc: 0.66 - ETA: 19s - loss: 0.9403 - acc: 0.67 - ETA: 11s - loss: 0.9447 - acc: 0.67 - ETA: 3s - loss: 0.9411 - acc: 0.6738 - 322s 121ms/step - loss: 0.9393 - acc: 0.6751\n",
      "Epoch 8/50\n",
      "2656/2656 [==============================] - ETA: 4:56 - loss: 0.9355 - acc: 0.656 - ETA: 4:52 - loss: 0.8145 - acc: 0.718 - ETA: 4:49 - loss: 0.8026 - acc: 0.724 - ETA: 4:43 - loss: 0.8058 - acc: 0.707 - ETA: 4:36 - loss: 0.8702 - acc: 0.696 - ETA: 4:29 - loss: 0.8927 - acc: 0.690 - ETA: 4:22 - loss: 0.8688 - acc: 0.705 - ETA: 4:14 - loss: 0.8657 - acc: 0.699 - ETA: 4:06 - loss: 0.8605 - acc: 0.703 - ETA: 4:00 - loss: 0.8486 - acc: 0.707 - ETA: 3:52 - loss: 0.8526 - acc: 0.708 - ETA: 3:45 - loss: 0.8664 - acc: 0.709 - ETA: 3:37 - loss: 0.8617 - acc: 0.711 - ETA: 3:29 - loss: 0.8626 - acc: 0.713 - ETA: 3:21 - loss: 0.8722 - acc: 0.704 - ETA: 3:14 - loss: 0.8679 - acc: 0.706 - ETA: 3:06 - loss: 0.8673 - acc: 0.708 - ETA: 2:59 - loss: 0.8632 - acc: 0.711 - ETA: 2:52 - loss: 0.8687 - acc: 0.712 - ETA: 2:44 - loss: 0.8726 - acc: 0.706 - ETA: 2:36 - loss: 0.8731 - acc: 0.705 - ETA: 2:29 - loss: 0.8673 - acc: 0.707 - ETA: 2:21 - loss: 0.8650 - acc: 0.707 - ETA: 2:13 - loss: 0.8702 - acc: 0.707 - ETA: 2:06 - loss: 0.8765 - acc: 0.705 - ETA: 1:58 - loss: 0.8741 - acc: 0.704 - ETA: 1:51 - loss: 0.8802 - acc: 0.702 - ETA: 1:43 - loss: 0.8855 - acc: 0.698 - ETA: 1:35 - loss: 0.8888 - acc: 0.696 - ETA: 1:28 - loss: 0.8946 - acc: 0.692 - ETA: 1:20 - loss: 0.8930 - acc: 0.693 - ETA: 1:13 - loss: 0.8918 - acc: 0.696 - ETA: 1:05 - loss: 0.8910 - acc: 0.697 - ETA: 57s - loss: 0.8853 - acc: 0.699 - ETA: 49s - loss: 0.8845 - acc: 0.69 - ETA: 42s - loss: 0.8858 - acc: 0.69 - ETA: 34s - loss: 0.8846 - acc: 0.69 - ETA: 26s - loss: 0.8843 - acc: 0.69 - ETA: 19s - loss: 0.8798 - acc: 0.69 - ETA: 11s - loss: 0.8765 - acc: 0.69 - ETA: 3s - loss: 0.8727 - acc: 0.7001 - 319s 120ms/step - loss: 0.8733 - acc: 0.6999\n",
      "Epoch 9/50\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2656/2656 [==============================] - ETA: 4:53 - loss: 0.8190 - acc: 0.750 - ETA: 4:49 - loss: 0.7948 - acc: 0.757 - ETA: 4:46 - loss: 0.7635 - acc: 0.755 - ETA: 4:42 - loss: 0.7657 - acc: 0.730 - ETA: 4:38 - loss: 0.7632 - acc: 0.731 - ETA: 4:31 - loss: 0.7642 - acc: 0.742 - ETA: 4:24 - loss: 0.7648 - acc: 0.736 - ETA: 4:17 - loss: 0.7647 - acc: 0.736 - ETA: 4:10 - loss: 0.7623 - acc: 0.732 - ETA: 4:03 - loss: 0.7709 - acc: 0.725 - ETA: 3:55 - loss: 0.7651 - acc: 0.727 - ETA: 3:48 - loss: 0.7933 - acc: 0.717 - ETA: 3:40 - loss: 0.7896 - acc: 0.713 - ETA: 3:33 - loss: 0.7982 - acc: 0.708 - ETA: 3:25 - loss: 0.7986 - acc: 0.709 - ETA: 3:17 - loss: 0.7956 - acc: 0.713 - ETA: 3:09 - loss: 0.8016 - acc: 0.712 - ETA: 3:01 - loss: 0.7973 - acc: 0.714 - ETA: 2:54 - loss: 0.8117 - acc: 0.711 - ETA: 2:46 - loss: 0.8097 - acc: 0.716 - ETA: 2:38 - loss: 0.8117 - acc: 0.713 - ETA: 2:31 - loss: 0.8107 - acc: 0.715 - ETA: 2:23 - loss: 0.8097 - acc: 0.714 - ETA: 2:16 - loss: 0.8062 - acc: 0.714 - ETA: 2:08 - loss: 0.8115 - acc: 0.712 - ETA: 2:00 - loss: 0.8045 - acc: 0.715 - ETA: 1:53 - loss: 0.8052 - acc: 0.717 - ETA: 1:45 - loss: 0.8042 - acc: 0.718 - ETA: 1:37 - loss: 0.7981 - acc: 0.721 - ETA: 1:30 - loss: 0.8045 - acc: 0.718 - ETA: 1:22 - loss: 0.7989 - acc: 0.720 - ETA: 1:14 - loss: 0.7905 - acc: 0.725 - ETA: 1:06 - loss: 0.7920 - acc: 0.725 - ETA: 58s - loss: 0.7933 - acc: 0.726 - ETA: 50s - loss: 0.7945 - acc: 0.72 - ETA: 43s - loss: 0.7936 - acc: 0.72 - ETA: 35s - loss: 0.7906 - acc: 0.72 - ETA: 27s - loss: 0.7975 - acc: 0.72 - ETA: 19s - loss: 0.8048 - acc: 0.72 - ETA: 11s - loss: 0.8065 - acc: 0.72 - ETA: 3s - loss: 0.8080 - acc: 0.7207 - 325s 122ms/step - loss: 0.8087 - acc: 0.7206\n",
      "Epoch 10/50\n",
      "2656/2656 [==============================] - ETA: 5:10 - loss: 0.6967 - acc: 0.796 - ETA: 5:04 - loss: 0.7531 - acc: 0.734 - ETA: 4:56 - loss: 0.7243 - acc: 0.755 - ETA: 4:48 - loss: 0.7359 - acc: 0.750 - ETA: 4:41 - loss: 0.7312 - acc: 0.737 - ETA: 4:35 - loss: 0.7097 - acc: 0.752 - ETA: 4:29 - loss: 0.7200 - acc: 0.743 - ETA: 4:23 - loss: 0.7420 - acc: 0.726 - ETA: 4:16 - loss: 0.7295 - acc: 0.732 - ETA: 4:09 - loss: 0.7460 - acc: 0.725 - ETA: 4:02 - loss: 0.7460 - acc: 0.728 - ETA: 3:55 - loss: 0.7301 - acc: 0.735 - ETA: 3:47 - loss: 0.7156 - acc: 0.742 - ETA: 3:39 - loss: 0.7128 - acc: 0.743 - ETA: 3:32 - loss: 0.7085 - acc: 0.742 - ETA: 3:24 - loss: 0.7039 - acc: 0.746 - ETA: 3:16 - loss: 0.7128 - acc: 0.742 - ETA: 3:08 - loss: 0.7226 - acc: 0.740 - ETA: 3:00 - loss: 0.7213 - acc: 0.741 - ETA: 2:52 - loss: 0.7192 - acc: 0.744 - ETA: 2:44 - loss: 0.7234 - acc: 0.744 - ETA: 2:36 - loss: 0.7238 - acc: 0.745 - ETA: 2:28 - loss: 0.7325 - acc: 0.741 - ETA: 2:20 - loss: 0.7297 - acc: 0.742 - ETA: 2:12 - loss: 0.7304 - acc: 0.743 - ETA: 2:04 - loss: 0.7300 - acc: 0.742 - ETA: 1:56 - loss: 0.7403 - acc: 0.742 - ETA: 1:48 - loss: 0.7389 - acc: 0.743 - ETA: 1:40 - loss: 0.7390 - acc: 0.743 - ETA: 1:32 - loss: 0.7374 - acc: 0.743 - ETA: 1:24 - loss: 0.7380 - acc: 0.741 - ETA: 1:16 - loss: 0.7360 - acc: 0.743 - ETA: 1:08 - loss: 0.7334 - acc: 0.745 - ETA: 1:00 - loss: 0.7387 - acc: 0.743 - ETA: 52s - loss: 0.7433 - acc: 0.741 - ETA: 44s - loss: 0.7429 - acc: 0.74 - ETA: 36s - loss: 0.7485 - acc: 0.73 - ETA: 28s - loss: 0.7488 - acc: 0.73 - ETA: 20s - loss: 0.7478 - acc: 0.73 - ETA: 12s - loss: 0.7525 - acc: 0.73 - ETA: 4s - loss: 0.7525 - acc: 0.7374 - 332s 125ms/step - loss: 0.7563 - acc: 0.7357\n",
      "Epoch 11/50\n",
      "2656/2656 [==============================] - ETA: 5:03 - loss: 0.5628 - acc: 0.828 - ETA: 4:56 - loss: 0.5992 - acc: 0.812 - ETA: 4:52 - loss: 0.6233 - acc: 0.781 - ETA: 4:45 - loss: 0.6550 - acc: 0.785 - ETA: 4:38 - loss: 0.6456 - acc: 0.778 - ETA: 4:31 - loss: 0.6716 - acc: 0.760 - ETA: 4:24 - loss: 0.6795 - acc: 0.754 - ETA: 4:17 - loss: 0.6994 - acc: 0.750 - ETA: 4:10 - loss: 0.7024 - acc: 0.746 - ETA: 4:02 - loss: 0.7062 - acc: 0.746 - ETA: 3:55 - loss: 0.6959 - acc: 0.755 - ETA: 3:47 - loss: 0.6814 - acc: 0.757 - ETA: 3:39 - loss: 0.6813 - acc: 0.756 - ETA: 3:32 - loss: 0.6733 - acc: 0.761 - ETA: 3:25 - loss: 0.6819 - acc: 0.758 - ETA: 3:17 - loss: 0.6809 - acc: 0.758 - ETA: 3:10 - loss: 0.6737 - acc: 0.761 - ETA: 3:02 - loss: 0.6729 - acc: 0.763 - ETA: 2:55 - loss: 0.6779 - acc: 0.759 - ETA: 2:47 - loss: 0.6762 - acc: 0.760 - ETA: 2:39 - loss: 0.6860 - acc: 0.758 - ETA: 2:32 - loss: 0.6929 - acc: 0.757 - ETA: 2:24 - loss: 0.6864 - acc: 0.760 - ETA: 2:16 - loss: 0.6906 - acc: 0.758 - ETA: 2:08 - loss: 0.6988 - acc: 0.754 - ETA: 2:01 - loss: 0.6957 - acc: 0.756 - ETA: 1:53 - loss: 0.6936 - acc: 0.755 - ETA: 1:45 - loss: 0.7032 - acc: 0.751 - ETA: 1:37 - loss: 0.7001 - acc: 0.750 - ETA: 1:29 - loss: 0.7010 - acc: 0.751 - ETA: 1:22 - loss: 0.6985 - acc: 0.752 - ETA: 1:14 - loss: 0.7081 - acc: 0.749 - ETA: 1:06 - loss: 0.7105 - acc: 0.750 - ETA: 58s - loss: 0.7100 - acc: 0.750 - ETA: 51s - loss: 0.7077 - acc: 0.75 - ETA: 43s - loss: 0.7058 - acc: 0.75 - ETA: 35s - loss: 0.7060 - acc: 0.75 - ETA: 27s - loss: 0.7082 - acc: 0.75 - ETA: 19s - loss: 0.7054 - acc: 0.75 - ETA: 11s - loss: 0.7061 - acc: 0.75 - ETA: 3s - loss: 0.7044 - acc: 0.7569 - 327s 123ms/step - loss: 0.7078 - acc: 0.7549\n",
      "Epoch 12/50\n",
      "2656/2656 [==============================] - ETA: 5:22 - loss: 0.5642 - acc: 0.843 - ETA: 5:17 - loss: 0.5990 - acc: 0.781 - ETA: 5:10 - loss: 0.6444 - acc: 0.776 - ETA: 5:03 - loss: 0.6589 - acc: 0.761 - ETA: 4:56 - loss: 0.6617 - acc: 0.756 - ETA: 4:48 - loss: 0.6519 - acc: 0.763 - ETA: 4:40 - loss: 0.6243 - acc: 0.774 - ETA: 4:32 - loss: 0.6186 - acc: 0.779 - ETA: 4:24 - loss: 0.6513 - acc: 0.772 - ETA: 4:15 - loss: 0.6407 - acc: 0.779 - ETA: 4:07 - loss: 0.6385 - acc: 0.779 - ETA: 3:59 - loss: 0.6323 - acc: 0.779 - ETA: 3:51 - loss: 0.6409 - acc: 0.777 - ETA: 3:44 - loss: 0.6353 - acc: 0.776 - ETA: 3:36 - loss: 0.6421 - acc: 0.777 - ETA: 3:28 - loss: 0.6272 - acc: 0.781 - ETA: 3:20 - loss: 0.6263 - acc: 0.784 - ETA: 3:11 - loss: 0.6333 - acc: 0.784 - ETA: 3:03 - loss: 0.6221 - acc: 0.791 - ETA: 2:55 - loss: 0.6232 - acc: 0.792 - ETA: 2:47 - loss: 0.6235 - acc: 0.791 - ETA: 2:39 - loss: 0.6240 - acc: 0.791 - ETA: 2:31 - loss: 0.6248 - acc: 0.794 - ETA: 2:23 - loss: 0.6252 - acc: 0.794 - ETA: 2:15 - loss: 0.6257 - acc: 0.793 - ETA: 2:07 - loss: 0.6319 - acc: 0.789 - ETA: 1:59 - loss: 0.6352 - acc: 0.788 - ETA: 1:51 - loss: 0.6399 - acc: 0.786 - ETA: 1:42 - loss: 0.6431 - acc: 0.783 - ETA: 1:34 - loss: 0.6448 - acc: 0.781 - ETA: 1:26 - loss: 0.6476 - acc: 0.779 - ETA: 1:18 - loss: 0.6512 - acc: 0.777 - ETA: 1:09 - loss: 0.6506 - acc: 0.778 - ETA: 1:01 - loss: 0.6519 - acc: 0.777 - ETA: 53s - loss: 0.6524 - acc: 0.775 - ETA: 45s - loss: 0.6514 - acc: 0.77 - ETA: 36s - loss: 0.6530 - acc: 0.77 - ETA: 28s - loss: 0.6564 - acc: 0.77 - ETA: 20s - loss: 0.6602 - acc: 0.77 - ETA: 12s - loss: 0.6610 - acc: 0.77 - ETA: 4s - loss: 0.6563 - acc: 0.7725 - 339s 128ms/step - loss: 0.6558 - acc: 0.7733\n",
      "Epoch 13/50\n",
      "2656/2656 [==============================] - ETA: 5:30 - loss: 0.7143 - acc: 0.750 - ETA: 5:21 - loss: 0.6978 - acc: 0.742 - ETA: 5:12 - loss: 0.6966 - acc: 0.739 - ETA: 5:06 - loss: 0.6374 - acc: 0.765 - ETA: 4:57 - loss: 0.6397 - acc: 0.775 - ETA: 4:50 - loss: 0.6202 - acc: 0.778 - ETA: 4:42 - loss: 0.6203 - acc: 0.785 - ETA: 4:33 - loss: 0.6330 - acc: 0.785 - ETA: 4:26 - loss: 0.6219 - acc: 0.789 - ETA: 4:18 - loss: 0.6141 - acc: 0.792 - ETA: 4:10 - loss: 0.6162 - acc: 0.784 - ETA: 4:02 - loss: 0.6044 - acc: 0.791 - ETA: 3:54 - loss: 0.6061 - acc: 0.789 - ETA: 3:46 - loss: 0.6128 - acc: 0.782 - ETA: 3:38 - loss: 0.6140 - acc: 0.778 - ETA: 3:30 - loss: 0.6113 - acc: 0.779 - ETA: 3:21 - loss: 0.6011 - acc: 0.785 - ETA: 3:13 - loss: 0.6003 - acc: 0.785 - ETA: 3:05 - loss: 0.6036 - acc: 0.782 - ETA: 2:56 - loss: 0.5988 - acc: 0.782 - ETA: 2:48 - loss: 0.5984 - acc: 0.784 - ETA: 2:40 - loss: 0.6026 - acc: 0.783 - ETA: 2:32 - loss: 0.6035 - acc: 0.781 - ETA: 2:24 - loss: 0.6064 - acc: 0.783 - ETA: 2:16 - loss: 0.6150 - acc: 0.778 - ETA: 2:07 - loss: 0.6125 - acc: 0.780 - ETA: 1:59 - loss: 0.6107 - acc: 0.778 - ETA: 1:51 - loss: 0.6115 - acc: 0.780 - ETA: 1:42 - loss: 0.6048 - acc: 0.783 - ETA: 1:34 - loss: 0.6088 - acc: 0.782 - ETA: 1:26 - loss: 0.6125 - acc: 0.781 - ETA: 1:18 - loss: 0.6161 - acc: 0.780 - ETA: 1:10 - loss: 0.6226 - acc: 0.776 - ETA: 1:01 - loss: 0.6308 - acc: 0.774 - ETA: 53s - loss: 0.6335 - acc: 0.772 - ETA: 45s - loss: 0.6362 - acc: 0.77 - ETA: 37s - loss: 0.6329 - acc: 0.77 - ETA: 29s - loss: 0.6327 - acc: 0.77 - ETA: 20s - loss: 0.6337 - acc: 0.77 - ETA: 12s - loss: 0.6368 - acc: 0.77 - ETA: 4s - loss: 0.6341 - acc: 0.7729 - 344s 130ms/step - loss: 0.6346 - acc: 0.7730\n",
      "Epoch 14/50\n",
      "2656/2656 [==============================] - ETA: 5:34 - loss: 0.4859 - acc: 0.828 - ETA: 5:34 - loss: 0.5446 - acc: 0.781 - ETA: 5:28 - loss: 0.5659 - acc: 0.791 - ETA: 5:21 - loss: 0.5360 - acc: 0.804 - ETA: 5:12 - loss: 0.5420 - acc: 0.790 - ETA: 5:03 - loss: 0.5309 - acc: 0.804 - ETA: 4:55 - loss: 0.5427 - acc: 0.805 - ETA: 4:46 - loss: 0.5596 - acc: 0.804 - ETA: 4:38 - loss: 0.5615 - acc: 0.798 - ETA: 4:28 - loss: 0.5521 - acc: 0.804 - ETA: 4:19 - loss: 0.5452 - acc: 0.806 - ETA: 4:10 - loss: 0.5542 - acc: 0.804 - ETA: 4:02 - loss: 0.5721 - acc: 0.800 - ETA: 3:53 - loss: 0.5855 - acc: 0.791 - ETA: 3:44 - loss: 0.5910 - acc: 0.786 - ETA: 3:35 - loss: 0.5896 - acc: 0.785 - ETA: 3:26 - loss: 0.5950 - acc: 0.782 - ETA: 3:17 - loss: 0.5883 - acc: 0.781 - ETA: 3:09 - loss: 0.5835 - acc: 0.786 - ETA: 3:00 - loss: 0.5800 - acc: 0.786 - ETA: 2:52 - loss: 0.5842 - acc: 0.783 - ETA: 2:43 - loss: 0.5869 - acc: 0.784 - ETA: 2:35 - loss: 0.5865 - acc: 0.786 - ETA: 2:26 - loss: 0.5849 - acc: 0.786 - ETA: 2:18 - loss: 0.5798 - acc: 0.788 - ETA: 2:10 - loss: 0.5770 - acc: 0.789 - ETA: 2:01 - loss: 0.5831 - acc: 0.787 - ETA: 1:53 - loss: 0.5875 - acc: 0.784 - ETA: 1:45 - loss: 0.5924 - acc: 0.783 - ETA: 1:36 - loss: 0.5924 - acc: 0.783 - ETA: 1:28 - loss: 0.5932 - acc: 0.782 - ETA: 1:19 - loss: 0.5869 - acc: 0.784 - ETA: 1:11 - loss: 0.5909 - acc: 0.784 - ETA: 1:03 - loss: 0.5873 - acc: 0.785 - ETA: 54s - loss: 0.5888 - acc: 0.785 - ETA: 46s - loss: 0.5879 - acc: 0.78 - ETA: 37s - loss: 0.5917 - acc: 0.78 - ETA: 29s - loss: 0.5939 - acc: 0.78 - ETA: 21s - loss: 0.5935 - acc: 0.78 - ETA: 12s - loss: 0.5926 - acc: 0.78 - ETA: 4s - loss: 0.5952 - acc: 0.7801 - 348s 131ms/step - loss: 0.5992 - acc: 0.7782\n",
      "Epoch 15/50\n",
      "2656/2656 [==============================] - ETA: 5:30 - loss: 0.7022 - acc: 0.750 - ETA: 5:21 - loss: 0.7037 - acc: 0.742 - ETA: 5:16 - loss: 0.6551 - acc: 0.755 - ETA: 5:08 - loss: 0.6250 - acc: 0.765 - ETA: 5:01 - loss: 0.6296 - acc: 0.759 - ETA: 4:53 - loss: 0.5953 - acc: 0.778 - ETA: 4:46 - loss: 0.5999 - acc: 0.776 - ETA: 4:39 - loss: 0.5992 - acc: 0.779 - ETA: 4:31 - loss: 0.5872 - acc: 0.788 - ETA: 4:22 - loss: 0.5754 - acc: 0.789 - ETA: 4:14 - loss: 0.5713 - acc: 0.792 - ETA: 4:05 - loss: 0.5772 - acc: 0.791 - ETA: 3:57 - loss: 0.5810 - acc: 0.787 - ETA: 3:49 - loss: 0.5947 - acc: 0.783 - ETA: 3:41 - loss: 0.5984 - acc: 0.782 - ETA: 3:32 - loss: 0.6099 - acc: 0.780 - ETA: 3:24 - loss: 0.6115 - acc: 0.774 - ETA: 3:16 - loss: 0.6162 - acc: 0.772 - ETA: 3:07 - loss: 0.6102 - acc: 0.775 - ETA: 2:59 - loss: 0.6065 - acc: 0.776 - ETA: 2:51 - loss: 0.5941 - acc: 0.782 - ETA: 2:42 - loss: 0.5845 - acc: 0.786 - ETA: 2:34 - loss: 0.5875 - acc: 0.785 - ETA: 2:26 - loss: 0.5847 - acc: 0.788 - ETA: 2:17 - loss: 0.5855 - acc: 0.786 - ETA: 2:09 - loss: 0.5806 - acc: 0.787 - ETA: 2:01 - loss: 0.5842 - acc: 0.785 - ETA: 1:52 - loss: 0.5813 - acc: 0.788 - ETA: 1:44 - loss: 0.5775 - acc: 0.788 - ETA: 1:36 - loss: 0.5781 - acc: 0.789 - ETA: 1:28 - loss: 0.5766 - acc: 0.788 - ETA: 1:19 - loss: 0.5728 - acc: 0.790 - ETA: 1:11 - loss: 0.5729 - acc: 0.791 - ETA: 1:03 - loss: 0.5736 - acc: 0.790 - ETA: 54s - loss: 0.5748 - acc: 0.787 - ETA: 46s - loss: 0.5784 - acc: 0.78 - ETA: 38s - loss: 0.5799 - acc: 0.78 - ETA: 29s - loss: 0.5789 - acc: 0.78 - ETA: 21s - loss: 0.5791 - acc: 0.78 - ETA: 12s - loss: 0.5788 - acc: 0.78 - ETA: 4s - loss: 0.5794 - acc: 0.7847 - 351s 132ms/step - loss: 0.5784 - acc: 0.7846\n",
      "Epoch 16/50\n",
      "2656/2656 [==============================] - ETA: 5:39 - loss: 0.5192 - acc: 0.812 - ETA: 5:37 - loss: 0.6039 - acc: 0.789 - ETA: 5:31 - loss: 0.5702 - acc: 0.786 - ETA: 5:21 - loss: 0.5845 - acc: 0.789 - ETA: 5:14 - loss: 0.5967 - acc: 0.778 - ETA: 5:05 - loss: 0.5813 - acc: 0.789 - ETA: 4:57 - loss: 0.5800 - acc: 0.783 - ETA: 4:47 - loss: 0.5623 - acc: 0.789 - ETA: 4:39 - loss: 0.5689 - acc: 0.786 - ETA: 4:30 - loss: 0.5665 - acc: 0.789 - ETA: 4:21 - loss: 0.5514 - acc: 0.795 - ETA: 4:13 - loss: 0.5465 - acc: 0.796 - ETA: 4:05 - loss: 0.5342 - acc: 0.802 - ETA: 3:57 - loss: 0.5256 - acc: 0.806 - ETA: 3:49 - loss: 0.5310 - acc: 0.807 - ETA: 3:40 - loss: 0.5334 - acc: 0.803 - ETA: 3:32 - loss: 0.5402 - acc: 0.797 - ETA: 3:24 - loss: 0.5356 - acc: 0.797 - ETA: 3:15 - loss: 0.5330 - acc: 0.802 - ETA: 3:07 - loss: 0.5378 - acc: 0.800 - ETA: 2:59 - loss: 0.5384 - acc: 0.799 - ETA: 2:50 - loss: 0.5341 - acc: 0.801 - ETA: 2:42 - loss: 0.5422 - acc: 0.798 - ETA: 2:33 - loss: 0.5390 - acc: 0.800 - ETA: 2:24 - loss: 0.5412 - acc: 0.800 - ETA: 2:16 - loss: 0.5434 - acc: 0.800 - ETA: 2:07 - loss: 0.5411 - acc: 0.802 - ETA: 1:58 - loss: 0.5412 - acc: 0.803 - ETA: 1:49 - loss: 0.5472 - acc: 0.800 - ETA: 1:41 - loss: 0.5533 - acc: 0.799 - ETA: 1:32 - loss: 0.5619 - acc: 0.796 - ETA: 1:23 - loss: 0.5632 - acc: 0.793 - ETA: 1:15 - loss: 0.5611 - acc: 0.794 - ETA: 1:06 - loss: 0.5630 - acc: 0.793 - ETA: 57s - loss: 0.5641 - acc: 0.793 - ETA: 48s - loss: 0.5642 - acc: 0.79 - ETA: 39s - loss: 0.5606 - acc: 0.79 - ETA: 31s - loss: 0.5642 - acc: 0.79 - ETA: 22s - loss: 0.5681 - acc: 0.79 - ETA: 13s - loss: 0.5719 - acc: 0.78 - ETA: 4s - loss: 0.5683 - acc: 0.7908 - 367s 138ms/step - loss: 0.5686 - acc: 0.7910\n",
      "Epoch 17/50\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2656/2656 [==============================] - ETA: 5:45 - loss: 0.3306 - acc: 0.890 - ETA: 5:39 - loss: 0.3931 - acc: 0.867 - ETA: 5:35 - loss: 0.4436 - acc: 0.854 - ETA: 5:23 - loss: 0.4978 - acc: 0.832 - ETA: 5:14 - loss: 0.5006 - acc: 0.825 - ETA: 5:05 - loss: 0.4870 - acc: 0.828 - ETA: 4:57 - loss: 0.4833 - acc: 0.825 - ETA: 4:48 - loss: 0.4999 - acc: 0.808 - ETA: 4:40 - loss: 0.5213 - acc: 0.795 - ETA: 4:31 - loss: 0.5122 - acc: 0.804 - ETA: 4:23 - loss: 0.5142 - acc: 0.805 - ETA: 4:14 - loss: 0.5150 - acc: 0.806 - ETA: 4:06 - loss: 0.5155 - acc: 0.807 - ETA: 3:58 - loss: 0.5114 - acc: 0.810 - ETA: 3:49 - loss: 0.5070 - acc: 0.810 - ETA: 3:41 - loss: 0.5079 - acc: 0.809 - ETA: 3:33 - loss: 0.5147 - acc: 0.807 - ETA: 3:24 - loss: 0.5119 - acc: 0.809 - ETA: 3:16 - loss: 0.5225 - acc: 0.802 - ETA: 3:07 - loss: 0.5245 - acc: 0.803 - ETA: 2:58 - loss: 0.5295 - acc: 0.800 - ETA: 2:50 - loss: 0.5347 - acc: 0.799 - ETA: 2:41 - loss: 0.5374 - acc: 0.800 - ETA: 2:32 - loss: 0.5390 - acc: 0.800 - ETA: 2:24 - loss: 0.5394 - acc: 0.800 - ETA: 2:15 - loss: 0.5395 - acc: 0.801 - ETA: 2:06 - loss: 0.5410 - acc: 0.799 - ETA: 1:58 - loss: 0.5433 - acc: 0.797 - ETA: 1:49 - loss: 0.5414 - acc: 0.798 - ETA: 1:40 - loss: 0.5372 - acc: 0.800 - ETA: 1:32 - loss: 0.5328 - acc: 0.802 - ETA: 1:23 - loss: 0.5372 - acc: 0.799 - ETA: 1:14 - loss: 0.5354 - acc: 0.800 - ETA: 1:05 - loss: 0.5355 - acc: 0.798 - ETA: 57s - loss: 0.5314 - acc: 0.800 - ETA: 48s - loss: 0.5311 - acc: 0.79 - ETA: 39s - loss: 0.5272 - acc: 0.80 - ETA: 30s - loss: 0.5258 - acc: 0.80 - ETA: 22s - loss: 0.5275 - acc: 0.80 - ETA: 13s - loss: 0.5328 - acc: 0.80 - ETA: 4s - loss: 0.5349 - acc: 0.8014 - 365s 137ms/step - loss: 0.5361 - acc: 0.8005\n",
      "Epoch 18/50\n",
      "2656/2656 [==============================] - ETA: 5:38 - loss: 0.5103 - acc: 0.812 - ETA: 5:43 - loss: 0.5455 - acc: 0.796 - ETA: 5:38 - loss: 0.5002 - acc: 0.828 - ETA: 5:32 - loss: 0.4782 - acc: 0.843 - ETA: 5:24 - loss: 0.5140 - acc: 0.806 - ETA: 5:15 - loss: 0.5141 - acc: 0.804 - ETA: 5:06 - loss: 0.5039 - acc: 0.805 - ETA: 4:58 - loss: 0.4970 - acc: 0.804 - ETA: 4:49 - loss: 0.5059 - acc: 0.802 - ETA: 4:41 - loss: 0.5020 - acc: 0.804 - ETA: 4:33 - loss: 0.5023 - acc: 0.806 - ETA: 4:24 - loss: 0.5012 - acc: 0.806 - ETA: 4:15 - loss: 0.5053 - acc: 0.806 - ETA: 4:05 - loss: 0.5130 - acc: 0.804 - ETA: 3:56 - loss: 0.5188 - acc: 0.803 - ETA: 3:47 - loss: 0.5102 - acc: 0.805 - ETA: 3:38 - loss: 0.5093 - acc: 0.805 - ETA: 3:29 - loss: 0.5067 - acc: 0.809 - ETA: 3:20 - loss: 0.5056 - acc: 0.809 - ETA: 3:11 - loss: 0.5118 - acc: 0.806 - ETA: 3:02 - loss: 0.5160 - acc: 0.803 - ETA: 2:53 - loss: 0.5092 - acc: 0.806 - ETA: 2:44 - loss: 0.5144 - acc: 0.806 - ETA: 2:35 - loss: 0.5170 - acc: 0.805 - ETA: 2:26 - loss: 0.5104 - acc: 0.808 - ETA: 2:17 - loss: 0.5120 - acc: 0.806 - ETA: 2:08 - loss: 0.5046 - acc: 0.808 - ETA: 2:00 - loss: 0.5037 - acc: 0.807 - ETA: 1:51 - loss: 0.5079 - acc: 0.804 - ETA: 1:42 - loss: 0.5081 - acc: 0.804 - ETA: 1:33 - loss: 0.5071 - acc: 0.806 - ETA: 1:24 - loss: 0.5087 - acc: 0.803 - ETA: 1:15 - loss: 0.5054 - acc: 0.805 - ETA: 1:06 - loss: 0.5100 - acc: 0.803 - ETA: 57s - loss: 0.5124 - acc: 0.803 - ETA: 49s - loss: 0.5124 - acc: 0.80 - ETA: 40s - loss: 0.5095 - acc: 0.80 - ETA: 31s - loss: 0.5136 - acc: 0.80 - ETA: 22s - loss: 0.5161 - acc: 0.80 - ETA: 13s - loss: 0.5155 - acc: 0.80 - ETA: 4s - loss: 0.5133 - acc: 0.8022 - 369s 139ms/step - loss: 0.5143 - acc: 0.8023\n",
      "Epoch 19/50\n",
      "2656/2656 [==============================] - ETA: 5:41 - loss: 0.5739 - acc: 0.812 - ETA: 5:40 - loss: 0.5655 - acc: 0.804 - ETA: 5:32 - loss: 0.5794 - acc: 0.786 - ETA: 5:25 - loss: 0.5979 - acc: 0.769 - ETA: 5:15 - loss: 0.5836 - acc: 0.771 - ETA: 5:07 - loss: 0.5504 - acc: 0.786 - ETA: 5:00 - loss: 0.5485 - acc: 0.783 - ETA: 4:50 - loss: 0.5499 - acc: 0.781 - ETA: 4:42 - loss: 0.5352 - acc: 0.788 - ETA: 4:35 - loss: 0.5312 - acc: 0.792 - ETA: 4:26 - loss: 0.5153 - acc: 0.799 - ETA: 4:18 - loss: 0.5109 - acc: 0.794 - ETA: 4:08 - loss: 0.5032 - acc: 0.801 - ETA: 3:59 - loss: 0.4880 - acc: 0.809 - ETA: 3:51 - loss: 0.4825 - acc: 0.810 - ETA: 3:42 - loss: 0.4911 - acc: 0.811 - ETA: 3:33 - loss: 0.4879 - acc: 0.814 - ETA: 3:24 - loss: 0.4803 - acc: 0.818 - ETA: 3:15 - loss: 0.4789 - acc: 0.817 - ETA: 3:07 - loss: 0.4810 - acc: 0.818 - ETA: 2:58 - loss: 0.4793 - acc: 0.818 - ETA: 2:49 - loss: 0.4911 - acc: 0.810 - ETA: 2:41 - loss: 0.4901 - acc: 0.811 - ETA: 2:32 - loss: 0.4847 - acc: 0.813 - ETA: 2:24 - loss: 0.4906 - acc: 0.811 - ETA: 2:15 - loss: 0.4990 - acc: 0.808 - ETA: 2:07 - loss: 0.4992 - acc: 0.809 - ETA: 1:58 - loss: 0.4991 - acc: 0.810 - ETA: 1:49 - loss: 0.5004 - acc: 0.810 - ETA: 1:40 - loss: 0.4985 - acc: 0.812 - ETA: 1:32 - loss: 0.5044 - acc: 0.810 - ETA: 1:23 - loss: 0.5099 - acc: 0.809 - ETA: 1:14 - loss: 0.5143 - acc: 0.805 - ETA: 1:05 - loss: 0.5151 - acc: 0.805 - ETA: 57s - loss: 0.5204 - acc: 0.803 - ETA: 48s - loss: 0.5197 - acc: 0.80 - ETA: 39s - loss: 0.5169 - acc: 0.80 - ETA: 30s - loss: 0.5180 - acc: 0.80 - ETA: 21s - loss: 0.5225 - acc: 0.80 - ETA: 13s - loss: 0.5200 - acc: 0.80 - ETA: 4s - loss: 0.5178 - acc: 0.8022 - 363s 137ms/step - loss: 0.5159 - acc: 0.8023\n",
      "Epoch 20/50\n",
      "2656/2656 [==============================] - ETA: 5:34 - loss: 0.4906 - acc: 0.781 - ETA: 5:30 - loss: 0.4585 - acc: 0.820 - ETA: 5:28 - loss: 0.4605 - acc: 0.812 - ETA: 5:20 - loss: 0.4835 - acc: 0.800 - ETA: 5:12 - loss: 0.4814 - acc: 0.793 - ETA: 5:03 - loss: 0.4707 - acc: 0.807 - ETA: 4:56 - loss: 0.4698 - acc: 0.808 - ETA: 4:48 - loss: 0.4513 - acc: 0.818 - ETA: 4:39 - loss: 0.4355 - acc: 0.822 - ETA: 4:31 - loss: 0.4406 - acc: 0.818 - ETA: 4:22 - loss: 0.4366 - acc: 0.822 - ETA: 4:14 - loss: 0.4434 - acc: 0.817 - ETA: 4:05 - loss: 0.4579 - acc: 0.816 - ETA: 3:56 - loss: 0.4615 - acc: 0.812 - ETA: 3:47 - loss: 0.4583 - acc: 0.814 - ETA: 3:39 - loss: 0.4632 - acc: 0.813 - ETA: 3:30 - loss: 0.4765 - acc: 0.808 - ETA: 3:22 - loss: 0.4722 - acc: 0.808 - ETA: 3:13 - loss: 0.4702 - acc: 0.810 - ETA: 3:05 - loss: 0.4653 - acc: 0.810 - ETA: 2:56 - loss: 0.4663 - acc: 0.811 - ETA: 2:48 - loss: 0.4636 - acc: 0.811 - ETA: 2:39 - loss: 0.4688 - acc: 0.808 - ETA: 2:31 - loss: 0.4681 - acc: 0.809 - ETA: 2:22 - loss: 0.4762 - acc: 0.806 - ETA: 2:13 - loss: 0.4740 - acc: 0.807 - ETA: 2:05 - loss: 0.4743 - acc: 0.809 - ETA: 1:56 - loss: 0.4719 - acc: 0.810 - ETA: 1:48 - loss: 0.4750 - acc: 0.810 - ETA: 1:39 - loss: 0.4728 - acc: 0.812 - ETA: 1:30 - loss: 0.4737 - acc: 0.811 - ETA: 1:22 - loss: 0.4775 - acc: 0.809 - ETA: 1:13 - loss: 0.4799 - acc: 0.807 - ETA: 1:04 - loss: 0.4848 - acc: 0.806 - ETA: 56s - loss: 0.4862 - acc: 0.805 - ETA: 47s - loss: 0.4853 - acc: 0.80 - ETA: 39s - loss: 0.4837 - acc: 0.80 - ETA: 30s - loss: 0.4824 - acc: 0.80 - ETA: 21s - loss: 0.4914 - acc: 0.80 - ETA: 13s - loss: 0.4946 - acc: 0.80 - ETA: 4s - loss: 0.4974 - acc: 0.8030 - 361s 136ms/step - loss: 0.4992 - acc: 0.8023\n",
      "Epoch 21/50\n",
      "2656/2656 [==============================] - ETA: 5:49 - loss: 0.4612 - acc: 0.812 - ETA: 5:43 - loss: 0.4574 - acc: 0.820 - ETA: 5:32 - loss: 0.5226 - acc: 0.781 - ETA: 5:25 - loss: 0.5186 - acc: 0.789 - ETA: 5:18 - loss: 0.4969 - acc: 0.803 - ETA: 5:12 - loss: 0.4836 - acc: 0.809 - ETA: 5:04 - loss: 0.4903 - acc: 0.803 - ETA: 4:54 - loss: 0.4835 - acc: 0.806 - ETA: 4:45 - loss: 0.4740 - acc: 0.807 - ETA: 4:35 - loss: 0.4834 - acc: 0.801 - ETA: 4:26 - loss: 0.4828 - acc: 0.802 - ETA: 4:18 - loss: 0.4908 - acc: 0.798 - ETA: 4:09 - loss: 0.4909 - acc: 0.795 - ETA: 4:01 - loss: 0.4910 - acc: 0.795 - ETA: 3:52 - loss: 0.4768 - acc: 0.805 - ETA: 3:44 - loss: 0.4844 - acc: 0.805 - ETA: 3:35 - loss: 0.4849 - acc: 0.807 - ETA: 3:27 - loss: 0.4900 - acc: 0.803 - ETA: 3:18 - loss: 0.4908 - acc: 0.801 - ETA: 3:10 - loss: 0.5050 - acc: 0.798 - ETA: 3:01 - loss: 0.5080 - acc: 0.796 - ETA: 2:53 - loss: 0.5047 - acc: 0.796 - ETA: 2:44 - loss: 0.5094 - acc: 0.794 - ETA: 2:36 - loss: 0.5063 - acc: 0.796 - ETA: 2:27 - loss: 0.5036 - acc: 0.796 - ETA: 2:18 - loss: 0.4993 - acc: 0.801 - ETA: 2:09 - loss: 0.5023 - acc: 0.799 - ETA: 2:01 - loss: 0.5003 - acc: 0.799 - ETA: 1:52 - loss: 0.4967 - acc: 0.802 - ETA: 1:43 - loss: 0.4977 - acc: 0.801 - ETA: 1:34 - loss: 0.4984 - acc: 0.798 - ETA: 1:25 - loss: 0.5008 - acc: 0.798 - ETA: 1:16 - loss: 0.5081 - acc: 0.796 - ETA: 1:07 - loss: 0.5075 - acc: 0.796 - ETA: 58s - loss: 0.5050 - acc: 0.798 - ETA: 49s - loss: 0.5092 - acc: 0.79 - ETA: 40s - loss: 0.5072 - acc: 0.79 - ETA: 31s - loss: 0.5103 - acc: 0.79 - ETA: 22s - loss: 0.5109 - acc: 0.79 - ETA: 13s - loss: 0.5098 - acc: 0.79 - ETA: 4s - loss: 0.5069 - acc: 0.7976 - 375s 141ms/step - loss: 0.5080 - acc: 0.7978\n",
      "Epoch 22/50\n",
      "2656/2656 [==============================] - ETA: 5:52 - loss: 0.3662 - acc: 0.890 - ETA: 5:46 - loss: 0.4338 - acc: 0.835 - ETA: 5:40 - loss: 0.4553 - acc: 0.849 - ETA: 5:33 - loss: 0.4451 - acc: 0.847 - ETA: 5:25 - loss: 0.4539 - acc: 0.846 - ETA: 5:16 - loss: 0.4321 - acc: 0.846 - ETA: 5:08 - loss: 0.4613 - acc: 0.825 - ETA: 4:59 - loss: 0.4769 - acc: 0.814 - ETA: 4:52 - loss: 0.4795 - acc: 0.814 - ETA: 4:44 - loss: 0.4739 - acc: 0.817 - ETA: 4:37 - loss: 0.4813 - acc: 0.816 - ETA: 4:28 - loss: 0.4841 - acc: 0.812 - ETA: 4:19 - loss: 0.4870 - acc: 0.812 - ETA: 4:11 - loss: 0.4873 - acc: 0.811 - ETA: 4:02 - loss: 0.4855 - acc: 0.812 - ETA: 3:53 - loss: 0.4773 - acc: 0.814 - ETA: 3:44 - loss: 0.4872 - acc: 0.810 - ETA: 3:35 - loss: 0.4911 - acc: 0.807 - ETA: 3:26 - loss: 0.4909 - acc: 0.807 - ETA: 3:17 - loss: 0.4873 - acc: 0.807 - ETA: 3:07 - loss: 0.4842 - acc: 0.807 - ETA: 2:58 - loss: 0.4837 - acc: 0.808 - ETA: 2:49 - loss: 0.4840 - acc: 0.806 - ETA: 2:40 - loss: 0.4842 - acc: 0.809 - ETA: 2:31 - loss: 0.4856 - acc: 0.808 - ETA: 2:22 - loss: 0.4907 - acc: 0.805 - ETA: 2:12 - loss: 0.4899 - acc: 0.805 - ETA: 2:03 - loss: 0.4947 - acc: 0.800 - ETA: 1:54 - loss: 0.4937 - acc: 0.800 - ETA: 1:45 - loss: 0.4888 - acc: 0.803 - ETA: 1:36 - loss: 0.4890 - acc: 0.802 - ETA: 1:27 - loss: 0.4911 - acc: 0.802 - ETA: 1:17 - loss: 0.4934 - acc: 0.800 - ETA: 1:08 - loss: 0.4955 - acc: 0.799 - ETA: 59s - loss: 0.4932 - acc: 0.800 - ETA: 50s - loss: 0.4970 - acc: 0.80 - ETA: 41s - loss: 0.4955 - acc: 0.79 - ETA: 32s - loss: 0.4978 - acc: 0.79 - ETA: 22s - loss: 0.4942 - acc: 0.80 - ETA: 13s - loss: 0.4957 - acc: 0.80 - ETA: 4s - loss: 0.4940 - acc: 0.8007 - 379s 143ms/step - loss: 0.4955 - acc: 0.8005\n",
      "Epoch 23/50\n",
      "2656/2656 [==============================] - ETA: 6:07 - loss: 0.3929 - acc: 0.890 - ETA: 5:54 - loss: 0.4333 - acc: 0.867 - ETA: 5:46 - loss: 0.4336 - acc: 0.854 - ETA: 5:37 - loss: 0.4572 - acc: 0.824 - ETA: 5:28 - loss: 0.4541 - acc: 0.809 - ETA: 5:20 - loss: 0.4651 - acc: 0.807 - ETA: 5:12 - loss: 0.4478 - acc: 0.817 - ETA: 5:04 - loss: 0.4475 - acc: 0.814 - ETA: 4:57 - loss: 0.4452 - acc: 0.814 - ETA: 4:48 - loss: 0.4371 - acc: 0.820 - ETA: 4:39 - loss: 0.4247 - acc: 0.825 - ETA: 4:30 - loss: 0.4230 - acc: 0.825 - ETA: 4:21 - loss: 0.4206 - acc: 0.830 - ETA: 4:11 - loss: 0.4168 - acc: 0.832 - ETA: 4:02 - loss: 0.4151 - acc: 0.834 - ETA: 3:53 - loss: 0.4112 - acc: 0.838 - ETA: 3:44 - loss: 0.4231 - acc: 0.831 - ETA: 3:35 - loss: 0.4264 - acc: 0.830 - ETA: 3:26 - loss: 0.4289 - acc: 0.831 - ETA: 3:17 - loss: 0.4291 - acc: 0.829 - ETA: 3:08 - loss: 0.4319 - acc: 0.828 - ETA: 2:59 - loss: 0.4373 - acc: 0.828 - ETA: 2:49 - loss: 0.4470 - acc: 0.823 - ETA: 2:40 - loss: 0.4479 - acc: 0.823 - ETA: 2:31 - loss: 0.4415 - acc: 0.825 - ETA: 2:22 - loss: 0.4450 - acc: 0.823 - ETA: 2:12 - loss: 0.4437 - acc: 0.822 - ETA: 2:03 - loss: 0.4383 - acc: 0.824 - ETA: 1:54 - loss: 0.4362 - acc: 0.825 - ETA: 1:45 - loss: 0.4375 - acc: 0.822 - ETA: 1:36 - loss: 0.4488 - acc: 0.818 - ETA: 1:27 - loss: 0.4499 - acc: 0.817 - ETA: 1:17 - loss: 0.4535 - acc: 0.817 - ETA: 1:08 - loss: 0.4548 - acc: 0.817 - ETA: 59s - loss: 0.4552 - acc: 0.816 - ETA: 50s - loss: 0.4576 - acc: 0.81 - ETA: 41s - loss: 0.4614 - acc: 0.81 - ETA: 32s - loss: 0.4602 - acc: 0.81 - ETA: 22s - loss: 0.4614 - acc: 0.81 - ETA: 13s - loss: 0.4687 - acc: 0.81 - ETA: 4s - loss: 0.4710 - acc: 0.8125 - 381s 143ms/step - loss: 0.4714 - acc: 0.8125\n",
      "Epoch 24/50\n",
      "2656/2656 [==============================] - ETA: 6:01 - loss: 0.3824 - acc: 0.859 - ETA: 5:52 - loss: 0.4059 - acc: 0.835 - ETA: 5:48 - loss: 0.4682 - acc: 0.802 - ETA: 5:39 - loss: 0.4588 - acc: 0.804 - ETA: 5:33 - loss: 0.4694 - acc: 0.815 - ETA: 5:25 - loss: 0.4707 - acc: 0.817 - ETA: 5:18 - loss: 0.4573 - acc: 0.823 - ETA: 5:08 - loss: 0.4597 - acc: 0.824 - ETA: 5:00 - loss: 0.4520 - acc: 0.824 - ETA: 4:51 - loss: 0.4602 - acc: 0.820 - ETA: 4:41 - loss: 0.4582 - acc: 0.821 - ETA: 4:32 - loss: 0.4531 - acc: 0.826 - ETA: 4:23 - loss: 0.4514 - acc: 0.825 - ETA: 4:13 - loss: 0.4471 - acc: 0.824 - ETA: 4:04 - loss: 0.4522 - acc: 0.820 - ETA: 3:54 - loss: 0.4575 - acc: 0.815 - ETA: 3:45 - loss: 0.4573 - acc: 0.816 - ETA: 3:36 - loss: 0.4555 - acc: 0.819 - ETA: 3:27 - loss: 0.4563 - acc: 0.816 - ETA: 3:18 - loss: 0.4596 - acc: 0.818 - ETA: 3:09 - loss: 0.4608 - acc: 0.814 - ETA: 3:00 - loss: 0.4674 - acc: 0.811 - ETA: 2:50 - loss: 0.4607 - acc: 0.813 - ETA: 2:41 - loss: 0.4591 - acc: 0.814 - ETA: 2:32 - loss: 0.4566 - acc: 0.814 - ETA: 2:23 - loss: 0.4574 - acc: 0.812 - ETA: 2:13 - loss: 0.4583 - acc: 0.811 - ETA: 2:04 - loss: 0.4570 - acc: 0.811 - ETA: 1:55 - loss: 0.4607 - acc: 0.810 - ETA: 1:45 - loss: 0.4615 - acc: 0.809 - ETA: 1:36 - loss: 0.4580 - acc: 0.812 - ETA: 1:27 - loss: 0.4576 - acc: 0.811 - ETA: 1:18 - loss: 0.4622 - acc: 0.811 - ETA: 1:08 - loss: 0.4631 - acc: 0.809 - ETA: 59s - loss: 0.4674 - acc: 0.808 - ETA: 50s - loss: 0.4676 - acc: 0.80 - ETA: 41s - loss: 0.4673 - acc: 0.80 - ETA: 32s - loss: 0.4695 - acc: 0.80 - ETA: 23s - loss: 0.4702 - acc: 0.80 - ETA: 13s - loss: 0.4700 - acc: 0.80 - ETA: 4s - loss: 0.4667 - acc: 0.8053 - 381s 143ms/step - loss: 0.4655 - acc: 0.8065\n",
      "Epoch 25/50\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2656/2656 [==============================] - ETA: 5:56 - loss: 0.4772 - acc: 0.765 - ETA: 5:49 - loss: 0.4409 - acc: 0.796 - ETA: 5:40 - loss: 0.3908 - acc: 0.833 - ETA: 5:31 - loss: 0.3964 - acc: 0.835 - ETA: 5:22 - loss: 0.4147 - acc: 0.828 - ETA: 5:12 - loss: 0.4158 - acc: 0.822 - ETA: 5:02 - loss: 0.3951 - acc: 0.837 - ETA: 4:53 - loss: 0.4126 - acc: 0.828 - ETA: 4:45 - loss: 0.4200 - acc: 0.826 - ETA: 4:35 - loss: 0.4084 - acc: 0.834 - ETA: 4:26 - loss: 0.4096 - acc: 0.835 - ETA: 4:18 - loss: 0.4101 - acc: 0.837 - ETA: 4:10 - loss: 0.4022 - acc: 0.837 - ETA: 4:01 - loss: 0.4207 - acc: 0.835 - ETA: 3:53 - loss: 0.4161 - acc: 0.836 - ETA: 3:44 - loss: 0.4170 - acc: 0.836 - ETA: 3:36 - loss: 0.4282 - acc: 0.831 - ETA: 3:27 - loss: 0.4238 - acc: 0.831 - ETA: 3:18 - loss: 0.4212 - acc: 0.833 - ETA: 3:09 - loss: 0.4272 - acc: 0.832 - ETA: 3:01 - loss: 0.4261 - acc: 0.830 - ETA: 2:52 - loss: 0.4277 - acc: 0.828 - ETA: 2:43 - loss: 0.4331 - acc: 0.826 - ETA: 2:35 - loss: 0.4355 - acc: 0.824 - ETA: 2:26 - loss: 0.4356 - acc: 0.824 - ETA: 2:17 - loss: 0.4358 - acc: 0.825 - ETA: 2:08 - loss: 0.4409 - acc: 0.823 - ETA: 1:59 - loss: 0.4422 - acc: 0.823 - ETA: 1:51 - loss: 0.4434 - acc: 0.822 - ETA: 1:42 - loss: 0.4429 - acc: 0.820 - ETA: 1:33 - loss: 0.4438 - acc: 0.818 - ETA: 1:24 - loss: 0.4511 - acc: 0.814 - ETA: 1:15 - loss: 0.4526 - acc: 0.813 - ETA: 1:06 - loss: 0.4551 - acc: 0.813 - ETA: 57s - loss: 0.4572 - acc: 0.811 - ETA: 49s - loss: 0.4603 - acc: 0.81 - ETA: 40s - loss: 0.4589 - acc: 0.81 - ETA: 31s - loss: 0.4579 - acc: 0.81 - ETA: 22s - loss: 0.4553 - acc: 0.81 - ETA: 13s - loss: 0.4598 - acc: 0.80 - ETA: 4s - loss: 0.4614 - acc: 0.8091 - 369s 139ms/step - loss: 0.4640 - acc: 0.8095\n",
      "Epoch 26/50\n",
      "2656/2656 [==============================] - ETA: 5:42 - loss: 0.3018 - acc: 0.859 - ETA: 5:39 - loss: 0.3107 - acc: 0.867 - ETA: 5:31 - loss: 0.3351 - acc: 0.859 - ETA: 5:23 - loss: 0.3780 - acc: 0.847 - ETA: 5:15 - loss: 0.4131 - acc: 0.837 - ETA: 5:07 - loss: 0.4238 - acc: 0.830 - ETA: 5:00 - loss: 0.4135 - acc: 0.830 - ETA: 4:53 - loss: 0.4086 - acc: 0.832 - ETA: 4:45 - loss: 0.4267 - acc: 0.816 - ETA: 4:37 - loss: 0.4226 - acc: 0.815 - ETA: 4:29 - loss: 0.4170 - acc: 0.821 - ETA: 4:20 - loss: 0.4142 - acc: 0.824 - ETA: 4:12 - loss: 0.4151 - acc: 0.828 - ETA: 4:03 - loss: 0.4282 - acc: 0.823 - ETA: 3:55 - loss: 0.4307 - acc: 0.821 - ETA: 3:46 - loss: 0.4311 - acc: 0.819 - ETA: 3:37 - loss: 0.4279 - acc: 0.821 - ETA: 3:29 - loss: 0.4219 - acc: 0.822 - ETA: 3:20 - loss: 0.4330 - acc: 0.818 - ETA: 3:11 - loss: 0.4359 - acc: 0.815 - ETA: 3:02 - loss: 0.4379 - acc: 0.814 - ETA: 2:54 - loss: 0.4369 - acc: 0.815 - ETA: 2:45 - loss: 0.4389 - acc: 0.814 - ETA: 2:36 - loss: 0.4366 - acc: 0.815 - ETA: 2:27 - loss: 0.4431 - acc: 0.811 - ETA: 2:18 - loss: 0.4466 - acc: 0.810 - ETA: 2:09 - loss: 0.4486 - acc: 0.808 - ETA: 2:00 - loss: 0.4478 - acc: 0.807 - ETA: 1:52 - loss: 0.4449 - acc: 0.809 - ETA: 1:43 - loss: 0.4450 - acc: 0.808 - ETA: 1:34 - loss: 0.4460 - acc: 0.807 - ETA: 1:25 - loss: 0.4495 - acc: 0.807 - ETA: 1:16 - loss: 0.4505 - acc: 0.806 - ETA: 1:07 - loss: 0.4565 - acc: 0.805 - ETA: 58s - loss: 0.4559 - acc: 0.806 - ETA: 49s - loss: 0.4592 - acc: 0.80 - ETA: 40s - loss: 0.4657 - acc: 0.80 - ETA: 31s - loss: 0.4673 - acc: 0.80 - ETA: 22s - loss: 0.4700 - acc: 0.80 - ETA: 13s - loss: 0.4696 - acc: 0.80 - ETA: 4s - loss: 0.4688 - acc: 0.8030 - 373s 140ms/step - loss: 0.4680 - acc: 0.8027\n",
      "Epoch 27/50\n",
      "2656/2656 [==============================] - ETA: 5:54 - loss: 0.4995 - acc: 0.812 - ETA: 5:45 - loss: 0.3988 - acc: 0.835 - ETA: 5:36 - loss: 0.3865 - acc: 0.859 - ETA: 5:27 - loss: 0.4181 - acc: 0.832 - ETA: 5:18 - loss: 0.4028 - acc: 0.840 - ETA: 5:12 - loss: 0.4190 - acc: 0.825 - ETA: 5:04 - loss: 0.4405 - acc: 0.814 - ETA: 4:55 - loss: 0.4308 - acc: 0.822 - ETA: 4:46 - loss: 0.4347 - acc: 0.824 - ETA: 4:37 - loss: 0.4362 - acc: 0.825 - ETA: 4:29 - loss: 0.4317 - acc: 0.822 - ETA: 4:21 - loss: 0.4204 - acc: 0.828 - ETA: 4:13 - loss: 0.4258 - acc: 0.825 - ETA: 4:05 - loss: 0.4169 - acc: 0.831 - ETA: 3:56 - loss: 0.4211 - acc: 0.830 - ETA: 3:46 - loss: 0.4241 - acc: 0.827 - ETA: 3:37 - loss: 0.4377 - acc: 0.819 - ETA: 3:28 - loss: 0.4330 - acc: 0.822 - ETA: 3:19 - loss: 0.4294 - acc: 0.823 - ETA: 3:11 - loss: 0.4345 - acc: 0.817 - ETA: 3:02 - loss: 0.4401 - acc: 0.815 - ETA: 2:53 - loss: 0.4389 - acc: 0.813 - ETA: 2:44 - loss: 0.4390 - acc: 0.812 - ETA: 2:35 - loss: 0.4368 - acc: 0.813 - ETA: 2:27 - loss: 0.4405 - acc: 0.811 - ETA: 2:18 - loss: 0.4424 - acc: 0.810 - ETA: 2:09 - loss: 0.4414 - acc: 0.811 - ETA: 2:00 - loss: 0.4373 - acc: 0.813 - ETA: 1:51 - loss: 0.4339 - acc: 0.814 - ETA: 1:42 - loss: 0.4375 - acc: 0.813 - ETA: 1:33 - loss: 0.4422 - acc: 0.810 - ETA: 1:24 - loss: 0.4414 - acc: 0.811 - ETA: 1:16 - loss: 0.4443 - acc: 0.811 - ETA: 1:07 - loss: 0.4467 - acc: 0.810 - ETA: 58s - loss: 0.4470 - acc: 0.810 - ETA: 49s - loss: 0.4454 - acc: 0.81 - ETA: 40s - loss: 0.4475 - acc: 0.81 - ETA: 31s - loss: 0.4470 - acc: 0.81 - ETA: 22s - loss: 0.4509 - acc: 0.80 - ETA: 13s - loss: 0.4531 - acc: 0.80 - ETA: 4s - loss: 0.4526 - acc: 0.8079 - 372s 140ms/step - loss: 0.4544 - acc: 0.8072\n",
      "Epoch 28/50\n",
      "2656/2656 [==============================] - ETA: 5:59 - loss: 0.5629 - acc: 0.750 - ETA: 5:49 - loss: 0.5040 - acc: 0.773 - ETA: 5:45 - loss: 0.4761 - acc: 0.776 - ETA: 5:39 - loss: 0.4384 - acc: 0.808 - ETA: 5:31 - loss: 0.4310 - acc: 0.821 - ETA: 5:24 - loss: 0.4304 - acc: 0.817 - ETA: 5:16 - loss: 0.4357 - acc: 0.817 - ETA: 5:08 - loss: 0.4335 - acc: 0.814 - ETA: 4:59 - loss: 0.4330 - acc: 0.812 - ETA: 4:50 - loss: 0.4581 - acc: 0.798 - ETA: 4:40 - loss: 0.4498 - acc: 0.801 - ETA: 4:32 - loss: 0.4539 - acc: 0.796 - ETA: 4:23 - loss: 0.4563 - acc: 0.798 - ETA: 4:14 - loss: 0.4610 - acc: 0.793 - ETA: 4:05 - loss: 0.4539 - acc: 0.799 - ETA: 3:56 - loss: 0.4567 - acc: 0.797 - ETA: 3:46 - loss: 0.4568 - acc: 0.798 - ETA: 3:37 - loss: 0.4499 - acc: 0.801 - ETA: 3:27 - loss: 0.4431 - acc: 0.805 - ETA: 3:18 - loss: 0.4523 - acc: 0.801 - ETA: 3:09 - loss: 0.4463 - acc: 0.802 - ETA: 3:00 - loss: 0.4468 - acc: 0.803 - ETA: 2:50 - loss: 0.4438 - acc: 0.804 - ETA: 2:41 - loss: 0.4463 - acc: 0.802 - ETA: 2:32 - loss: 0.4505 - acc: 0.800 - ETA: 2:23 - loss: 0.4511 - acc: 0.800 - ETA: 2:14 - loss: 0.4495 - acc: 0.800 - ETA: 2:04 - loss: 0.4472 - acc: 0.801 - ETA: 1:55 - loss: 0.4454 - acc: 0.802 - ETA: 1:46 - loss: 0.4429 - acc: 0.803 - ETA: 1:37 - loss: 0.4392 - acc: 0.805 - ETA: 1:28 - loss: 0.4370 - acc: 0.806 - ETA: 1:19 - loss: 0.4354 - acc: 0.808 - ETA: 1:09 - loss: 0.4362 - acc: 0.808 - ETA: 1:00 - loss: 0.4413 - acc: 0.807 - ETA: 51s - loss: 0.4424 - acc: 0.808 - ETA: 42s - loss: 0.4435 - acc: 0.80 - ETA: 32s - loss: 0.4457 - acc: 0.80 - ETA: 23s - loss: 0.4436 - acc: 0.80 - ETA: 14s - loss: 0.4452 - acc: 0.80 - ETA: 4s - loss: 0.4434 - acc: 0.8079 - 388s 146ms/step - loss: 0.4414 - acc: 0.8087\n",
      "Epoch 29/50\n",
      "2656/2656 [==============================] - ETA: 6:08 - loss: 0.2779 - acc: 0.890 - ETA: 6:00 - loss: 0.3295 - acc: 0.859 - ETA: 5:50 - loss: 0.3716 - acc: 0.854 - ETA: 5:41 - loss: 0.3923 - acc: 0.851 - ETA: 5:31 - loss: 0.4353 - acc: 0.825 - ETA: 5:23 - loss: 0.4304 - acc: 0.820 - ETA: 5:14 - loss: 0.4562 - acc: 0.814 - ETA: 5:05 - loss: 0.4567 - acc: 0.818 - ETA: 4:56 - loss: 0.4548 - acc: 0.814 - ETA: 4:47 - loss: 0.4538 - acc: 0.810 - ETA: 4:38 - loss: 0.4413 - acc: 0.811 - ETA: 4:30 - loss: 0.4512 - acc: 0.808 - ETA: 4:21 - loss: 0.4593 - acc: 0.805 - ETA: 4:11 - loss: 0.4559 - acc: 0.808 - ETA: 4:02 - loss: 0.4726 - acc: 0.795 - ETA: 3:53 - loss: 0.4736 - acc: 0.799 - ETA: 3:43 - loss: 0.4667 - acc: 0.804 - ETA: 3:34 - loss: 0.4639 - acc: 0.803 - ETA: 3:25 - loss: 0.4568 - acc: 0.805 - ETA: 3:16 - loss: 0.4522 - acc: 0.807 - ETA: 3:07 - loss: 0.4502 - acc: 0.806 - ETA: 2:58 - loss: 0.4480 - acc: 0.809 - ETA: 2:49 - loss: 0.4374 - acc: 0.814 - ETA: 2:40 - loss: 0.4389 - acc: 0.813 - ETA: 2:31 - loss: 0.4364 - acc: 0.815 - ETA: 2:21 - loss: 0.4289 - acc: 0.819 - ETA: 2:12 - loss: 0.4361 - acc: 0.814 - ETA: 2:03 - loss: 0.4456 - acc: 0.811 - ETA: 1:54 - loss: 0.4443 - acc: 0.812 - ETA: 1:45 - loss: 0.4429 - acc: 0.811 - ETA: 1:36 - loss: 0.4474 - acc: 0.809 - ETA: 1:27 - loss: 0.4483 - acc: 0.809 - ETA: 1:18 - loss: 0.4492 - acc: 0.807 - ETA: 1:09 - loss: 0.4475 - acc: 0.808 - ETA: 59s - loss: 0.4439 - acc: 0.810 - ETA: 50s - loss: 0.4427 - acc: 0.81 - ETA: 41s - loss: 0.4441 - acc: 0.81 - ETA: 32s - loss: 0.4459 - acc: 0.80 - ETA: 23s - loss: 0.4502 - acc: 0.80 - ETA: 13s - loss: 0.4496 - acc: 0.80 - ETA: 4s - loss: 0.4544 - acc: 0.8041 - 382s 144ms/step - loss: 0.4533 - acc: 0.8046\n",
      "Epoch 30/50\n",
      "2656/2656 [==============================] - ETA: 6:00 - loss: 0.3891 - acc: 0.828 - ETA: 5:53 - loss: 0.4487 - acc: 0.820 - ETA: 5:44 - loss: 0.3948 - acc: 0.854 - ETA: 5:33 - loss: 0.3923 - acc: 0.843 - ETA: 5:26 - loss: 0.3916 - acc: 0.846 - ETA: 5:19 - loss: 0.4055 - acc: 0.838 - ETA: 5:11 - loss: 0.4165 - acc: 0.830 - ETA: 5:01 - loss: 0.4293 - acc: 0.822 - ETA: 4:52 - loss: 0.4249 - acc: 0.822 - ETA: 4:43 - loss: 0.4059 - acc: 0.832 - ETA: 4:34 - loss: 0.4180 - acc: 0.822 - ETA: 4:25 - loss: 0.4108 - acc: 0.825 - ETA: 4:16 - loss: 0.4097 - acc: 0.826 - ETA: 4:07 - loss: 0.4135 - acc: 0.827 - ETA: 3:58 - loss: 0.4125 - acc: 0.825 - ETA: 3:50 - loss: 0.4161 - acc: 0.824 - ETA: 3:40 - loss: 0.4111 - acc: 0.826 - ETA: 3:31 - loss: 0.4113 - acc: 0.824 - ETA: 3:22 - loss: 0.4152 - acc: 0.823 - ETA: 3:13 - loss: 0.4171 - acc: 0.821 - ETA: 3:04 - loss: 0.4225 - acc: 0.819 - ETA: 2:55 - loss: 0.4263 - acc: 0.816 - ETA: 2:46 - loss: 0.4288 - acc: 0.813 - ETA: 2:37 - loss: 0.4300 - acc: 0.813 - ETA: 2:28 - loss: 0.4303 - acc: 0.813 - ETA: 2:19 - loss: 0.4344 - acc: 0.811 - ETA: 2:10 - loss: 0.4317 - acc: 0.811 - ETA: 2:01 - loss: 0.4349 - acc: 0.809 - ETA: 1:52 - loss: 0.4368 - acc: 0.809 - ETA: 1:43 - loss: 0.4415 - acc: 0.806 - ETA: 1:34 - loss: 0.4427 - acc: 0.807 - ETA: 1:25 - loss: 0.4411 - acc: 0.808 - ETA: 1:16 - loss: 0.4412 - acc: 0.807 - ETA: 1:07 - loss: 0.4464 - acc: 0.806 - ETA: 58s - loss: 0.4483 - acc: 0.806 - ETA: 49s - loss: 0.4489 - acc: 0.80 - ETA: 40s - loss: 0.4496 - acc: 0.80 - ETA: 31s - loss: 0.4508 - acc: 0.80 - ETA: 22s - loss: 0.4506 - acc: 0.80 - ETA: 13s - loss: 0.4521 - acc: 0.80 - ETA: 4s - loss: 0.4526 - acc: 0.8060 - 373s 141ms/step - loss: 0.4504 - acc: 0.8069\n",
      "Epoch 31/50\n",
      "2656/2656 [==============================] - ETA: 5:49 - loss: 0.3202 - acc: 0.843 - ETA: 5:43 - loss: 0.3437 - acc: 0.859 - ETA: 5:35 - loss: 0.4230 - acc: 0.833 - ETA: 5:28 - loss: 0.4122 - acc: 0.832 - ETA: 5:21 - loss: 0.4186 - acc: 0.828 - ETA: 5:11 - loss: 0.4164 - acc: 0.822 - ETA: 5:03 - loss: 0.4015 - acc: 0.828 - ETA: 4:55 - loss: 0.3998 - acc: 0.826 - ETA: 4:47 - loss: 0.4086 - acc: 0.821 - ETA: 4:38 - loss: 0.4100 - acc: 0.823 - ETA: 4:30 - loss: 0.4020 - acc: 0.826 - ETA: 4:21 - loss: 0.4021 - acc: 0.824 - ETA: 4:13 - loss: 0.4086 - acc: 0.823 - ETA: 4:05 - loss: 0.4133 - acc: 0.820 - ETA: 3:56 - loss: 0.4208 - acc: 0.817 - ETA: 3:47 - loss: 0.4255 - acc: 0.814 - ETA: 3:39 - loss: 0.4205 - acc: 0.816 - ETA: 3:30 - loss: 0.4324 - acc: 0.812 - ETA: 3:21 - loss: 0.4319 - acc: 0.812 - ETA: 3:12 - loss: 0.4304 - acc: 0.812 - ETA: 3:03 - loss: 0.4316 - acc: 0.813 - ETA: 2:55 - loss: 0.4395 - acc: 0.811 - ETA: 2:46 - loss: 0.4454 - acc: 0.807 - ETA: 2:37 - loss: 0.4430 - acc: 0.809 - ETA: 2:28 - loss: 0.4425 - acc: 0.810 - ETA: 2:19 - loss: 0.4422 - acc: 0.811 - ETA: 2:10 - loss: 0.4449 - acc: 0.810 - ETA: 2:01 - loss: 0.4460 - acc: 0.809 - ETA: 1:52 - loss: 0.4458 - acc: 0.810 - ETA: 1:43 - loss: 0.4429 - acc: 0.813 - ETA: 1:34 - loss: 0.4413 - acc: 0.813 - ETA: 1:25 - loss: 0.4447 - acc: 0.812 - ETA: 1:16 - loss: 0.4469 - acc: 0.810 - ETA: 1:07 - loss: 0.4496 - acc: 0.807 - ETA: 58s - loss: 0.4485 - acc: 0.807 - ETA: 49s - loss: 0.4513 - acc: 0.80 - ETA: 40s - loss: 0.4468 - acc: 0.80 - ETA: 31s - loss: 0.4498 - acc: 0.80 - ETA: 22s - loss: 0.4497 - acc: 0.80 - ETA: 13s - loss: 0.4540 - acc: 0.80 - ETA: 4s - loss: 0.4548 - acc: 0.8041 - 376s 142ms/step - loss: 0.4538 - acc: 0.8042\n",
      "Epoch 32/50\n",
      "2656/2656 [==============================] - ETA: 6:00 - loss: 0.5415 - acc: 0.781 - ETA: 5:54 - loss: 0.4948 - acc: 0.835 - ETA: 5:47 - loss: 0.4916 - acc: 0.817 - ETA: 5:39 - loss: 0.4751 - acc: 0.812 - ETA: 5:29 - loss: 0.4738 - acc: 0.812 - ETA: 5:20 - loss: 0.4545 - acc: 0.820 - ETA: 5:12 - loss: 0.4600 - acc: 0.810 - ETA: 5:03 - loss: 0.4507 - acc: 0.814 - ETA: 4:55 - loss: 0.4468 - acc: 0.817 - ETA: 4:46 - loss: 0.4489 - acc: 0.815 - ETA: 4:37 - loss: 0.4378 - acc: 0.823 - ETA: 4:28 - loss: 0.4269 - acc: 0.828 - ETA: 4:18 - loss: 0.4193 - acc: 0.829 - ETA: 4:10 - loss: 0.4153 - acc: 0.832 - ETA: 4:00 - loss: 0.4172 - acc: 0.829 - ETA: 3:51 - loss: 0.4185 - acc: 0.827 - ETA: 3:43 - loss: 0.4301 - acc: 0.821 - ETA: 3:33 - loss: 0.4298 - acc: 0.822 - ETA: 3:24 - loss: 0.4262 - acc: 0.822 - ETA: 3:15 - loss: 0.4248 - acc: 0.824 - ETA: 3:06 - loss: 0.4223 - acc: 0.824 - ETA: 2:56 - loss: 0.4235 - acc: 0.822 - ETA: 2:47 - loss: 0.4198 - acc: 0.824 - ETA: 2:38 - loss: 0.4166 - acc: 0.823 - ETA: 2:29 - loss: 0.4166 - acc: 0.822 - ETA: 2:20 - loss: 0.4204 - acc: 0.822 - ETA: 2:11 - loss: 0.4217 - acc: 0.822 - ETA: 2:02 - loss: 0.4329 - acc: 0.817 - ETA: 1:53 - loss: 0.4363 - acc: 0.817 - ETA: 1:44 - loss: 0.4357 - acc: 0.816 - ETA: 1:35 - loss: 0.4354 - acc: 0.816 - ETA: 1:26 - loss: 0.4336 - acc: 0.816 - ETA: 1:17 - loss: 0.4319 - acc: 0.818 - ETA: 1:07 - loss: 0.4310 - acc: 0.818 - ETA: 58s - loss: 0.4265 - acc: 0.821 - ETA: 49s - loss: 0.4249 - acc: 0.82 - ETA: 40s - loss: 0.4254 - acc: 0.82 - ETA: 31s - loss: 0.4294 - acc: 0.81 - ETA: 22s - loss: 0.4303 - acc: 0.81 - ETA: 13s - loss: 0.4291 - acc: 0.81 - ETA: 4s - loss: 0.4311 - acc: 0.8182 - 376s 142ms/step - loss: 0.4305 - acc: 0.8181\n",
      "Epoch 33/50\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2656/2656 [==============================] - ETA: 5:55 - loss: 0.4717 - acc: 0.765 - ETA: 5:48 - loss: 0.4527 - acc: 0.796 - ETA: 5:39 - loss: 0.4401 - acc: 0.817 - ETA: 5:30 - loss: 0.4229 - acc: 0.816 - ETA: 5:22 - loss: 0.4035 - acc: 0.831 - ETA: 5:14 - loss: 0.3984 - acc: 0.833 - ETA: 5:05 - loss: 0.4189 - acc: 0.823 - ETA: 4:57 - loss: 0.4294 - acc: 0.816 - ETA: 4:48 - loss: 0.4241 - acc: 0.817 - ETA: 4:40 - loss: 0.4219 - acc: 0.818 - ETA: 4:31 - loss: 0.4278 - acc: 0.813 - ETA: 4:23 - loss: 0.4280 - acc: 0.813 - ETA: 4:14 - loss: 0.4276 - acc: 0.810 - ETA: 4:05 - loss: 0.4322 - acc: 0.805 - ETA: 3:57 - loss: 0.4227 - acc: 0.812 - ETA: 3:48 - loss: 0.4300 - acc: 0.811 - ETA: 3:39 - loss: 0.4273 - acc: 0.813 - ETA: 3:30 - loss: 0.4308 - acc: 0.809 - ETA: 3:21 - loss: 0.4271 - acc: 0.811 - ETA: 3:12 - loss: 0.4364 - acc: 0.809 - ETA: 3:03 - loss: 0.4312 - acc: 0.813 - ETA: 2:54 - loss: 0.4339 - acc: 0.811 - ETA: 2:45 - loss: 0.4303 - acc: 0.813 - ETA: 2:36 - loss: 0.4242 - acc: 0.817 - ETA: 2:28 - loss: 0.4274 - acc: 0.814 - ETA: 2:19 - loss: 0.4300 - acc: 0.814 - ETA: 2:10 - loss: 0.4275 - acc: 0.816 - ETA: 2:01 - loss: 0.4305 - acc: 0.815 - ETA: 1:52 - loss: 0.4292 - acc: 0.815 - ETA: 1:43 - loss: 0.4241 - acc: 0.816 - ETA: 1:34 - loss: 0.4234 - acc: 0.817 - ETA: 1:25 - loss: 0.4213 - acc: 0.819 - ETA: 1:16 - loss: 0.4240 - acc: 0.817 - ETA: 1:07 - loss: 0.4301 - acc: 0.813 - ETA: 58s - loss: 0.4326 - acc: 0.813 - ETA: 49s - loss: 0.4350 - acc: 0.81 - ETA: 40s - loss: 0.4348 - acc: 0.81 - ETA: 31s - loss: 0.4353 - acc: 0.81 - ETA: 22s - loss: 0.4320 - acc: 0.81 - ETA: 13s - loss: 0.4315 - acc: 0.81 - ETA: 4s - loss: 0.4328 - acc: 0.8148 - 377s 142ms/step - loss: 0.4320 - acc: 0.8159\n",
      "Epoch 34/50\n",
      "2656/2656 [==============================] - ETA: 6:08 - loss: 0.3415 - acc: 0.812 - ETA: 5:59 - loss: 0.3340 - acc: 0.804 - ETA: 5:49 - loss: 0.4087 - acc: 0.791 - ETA: 5:46 - loss: 0.4132 - acc: 0.804 - ETA: 5:38 - loss: 0.4336 - acc: 0.806 - ETA: 5:29 - loss: 0.4498 - acc: 0.807 - ETA: 5:19 - loss: 0.4406 - acc: 0.805 - ETA: 5:11 - loss: 0.4270 - acc: 0.808 - ETA: 5:01 - loss: 0.4247 - acc: 0.810 - ETA: 4:52 - loss: 0.4202 - acc: 0.815 - ETA: 4:42 - loss: 0.4364 - acc: 0.812 - ETA: 4:33 - loss: 0.4406 - acc: 0.808 - ETA: 4:24 - loss: 0.4454 - acc: 0.806 - ETA: 4:15 - loss: 0.4382 - acc: 0.810 - ETA: 4:06 - loss: 0.4390 - acc: 0.808 - ETA: 3:57 - loss: 0.4383 - acc: 0.806 - ETA: 3:47 - loss: 0.4374 - acc: 0.806 - ETA: 3:38 - loss: 0.4316 - acc: 0.809 - ETA: 3:29 - loss: 0.4376 - acc: 0.807 - ETA: 3:20 - loss: 0.4379 - acc: 0.807 - ETA: 3:10 - loss: 0.4300 - acc: 0.811 - ETA: 3:01 - loss: 0.4363 - acc: 0.808 - ETA: 2:52 - loss: 0.4389 - acc: 0.805 - ETA: 2:43 - loss: 0.4375 - acc: 0.806 - ETA: 2:34 - loss: 0.4327 - acc: 0.808 - ETA: 2:24 - loss: 0.4349 - acc: 0.806 - ETA: 2:15 - loss: 0.4360 - acc: 0.806 - ETA: 2:06 - loss: 0.4350 - acc: 0.806 - ETA: 1:56 - loss: 0.4357 - acc: 0.805 - ETA: 1:47 - loss: 0.4324 - acc: 0.807 - ETA: 1:38 - loss: 0.4359 - acc: 0.807 - ETA: 1:28 - loss: 0.4388 - acc: 0.804 - ETA: 1:19 - loss: 0.4396 - acc: 0.805 - ETA: 1:10 - loss: 0.4354 - acc: 0.807 - ETA: 1:00 - loss: 0.4337 - acc: 0.809 - ETA: 51s - loss: 0.4345 - acc: 0.808 - ETA: 42s - loss: 0.4346 - acc: 0.80 - ETA: 32s - loss: 0.4360 - acc: 0.80 - ETA: 23s - loss: 0.4350 - acc: 0.80 - ETA: 14s - loss: 0.4355 - acc: 0.80 - ETA: 4s - loss: 0.4342 - acc: 0.8091 - 388s 146ms/step - loss: 0.4339 - acc: 0.8095\n",
      "Epoch 35/50\n",
      "2656/2656 [==============================] - ETA: 6:06 - loss: 0.3485 - acc: 0.859 - ETA: 6:01 - loss: 0.3367 - acc: 0.851 - ETA: 5:53 - loss: 0.3294 - acc: 0.859 - ETA: 5:44 - loss: 0.3594 - acc: 0.847 - ETA: 5:36 - loss: 0.3709 - acc: 0.843 - ETA: 5:27 - loss: 0.3745 - acc: 0.830 - ETA: 5:18 - loss: 0.3859 - acc: 0.828 - ETA: 5:08 - loss: 0.3956 - acc: 0.820 - ETA: 4:58 - loss: 0.3913 - acc: 0.828 - ETA: 4:50 - loss: 0.4045 - acc: 0.829 - ETA: 4:41 - loss: 0.4088 - acc: 0.825 - ETA: 4:32 - loss: 0.4048 - acc: 0.829 - ETA: 4:23 - loss: 0.4048 - acc: 0.830 - ETA: 4:13 - loss: 0.4024 - acc: 0.831 - ETA: 4:04 - loss: 0.4049 - acc: 0.831 - ETA: 3:54 - loss: 0.4081 - acc: 0.826 - ETA: 3:45 - loss: 0.4136 - acc: 0.826 - ETA: 3:36 - loss: 0.4151 - acc: 0.827 - ETA: 3:26 - loss: 0.4185 - acc: 0.826 - ETA: 3:17 - loss: 0.4167 - acc: 0.825 - ETA: 3:08 - loss: 0.4184 - acc: 0.823 - ETA: 2:59 - loss: 0.4163 - acc: 0.826 - ETA: 2:50 - loss: 0.4183 - acc: 0.824 - ETA: 2:41 - loss: 0.4158 - acc: 0.823 - ETA: 2:32 - loss: 0.4152 - acc: 0.822 - ETA: 2:22 - loss: 0.4143 - acc: 0.823 - ETA: 2:13 - loss: 0.4141 - acc: 0.823 - ETA: 2:04 - loss: 0.4172 - acc: 0.822 - ETA: 1:55 - loss: 0.4169 - acc: 0.821 - ETA: 1:46 - loss: 0.4150 - acc: 0.821 - ETA: 1:37 - loss: 0.4221 - acc: 0.817 - ETA: 1:28 - loss: 0.4216 - acc: 0.815 - ETA: 1:18 - loss: 0.4232 - acc: 0.814 - ETA: 1:09 - loss: 0.4269 - acc: 0.813 - ETA: 1:00 - loss: 0.4275 - acc: 0.813 - ETA: 51s - loss: 0.4252 - acc: 0.814 - ETA: 41s - loss: 0.4272 - acc: 0.81 - ETA: 32s - loss: 0.4272 - acc: 0.81 - ETA: 23s - loss: 0.4291 - acc: 0.81 - ETA: 14s - loss: 0.4294 - acc: 0.81 - ETA: 4s - loss: 0.4285 - acc: 0.8144 - 387s 146ms/step - loss: 0.4313 - acc: 0.8136\n",
      "Epoch 36/50\n",
      "2656/2656 [==============================] - ETA: 6:17 - loss: 0.3880 - acc: 0.843 - ETA: 6:09 - loss: 0.3867 - acc: 0.828 - ETA: 6:01 - loss: 0.3728 - acc: 0.822 - ETA: 5:53 - loss: 0.3777 - acc: 0.828 - ETA: 5:43 - loss: 0.3544 - acc: 0.843 - ETA: 5:33 - loss: 0.3699 - acc: 0.838 - ETA: 5:23 - loss: 0.3686 - acc: 0.841 - ETA: 5:13 - loss: 0.3586 - acc: 0.847 - ETA: 5:04 - loss: 0.3631 - acc: 0.845 - ETA: 4:54 - loss: 0.3746 - acc: 0.845 - ETA: 4:45 - loss: 0.3777 - acc: 0.843 - ETA: 4:35 - loss: 0.3806 - acc: 0.843 - ETA: 4:26 - loss: 0.3944 - acc: 0.838 - ETA: 4:17 - loss: 0.3980 - acc: 0.838 - ETA: 4:08 - loss: 0.4016 - acc: 0.834 - ETA: 3:58 - loss: 0.3945 - acc: 0.837 - ETA: 3:49 - loss: 0.3944 - acc: 0.835 - ETA: 3:40 - loss: 0.4019 - acc: 0.832 - ETA: 3:30 - loss: 0.4043 - acc: 0.831 - ETA: 3:21 - loss: 0.4129 - acc: 0.827 - ETA: 3:12 - loss: 0.4129 - acc: 0.825 - ETA: 3:03 - loss: 0.4093 - acc: 0.828 - ETA: 2:54 - loss: 0.4056 - acc: 0.830 - ETA: 2:44 - loss: 0.4056 - acc: 0.830 - ETA: 2:35 - loss: 0.4098 - acc: 0.826 - ETA: 2:25 - loss: 0.4121 - acc: 0.824 - ETA: 2:16 - loss: 0.4078 - acc: 0.825 - ETA: 2:07 - loss: 0.4065 - acc: 0.826 - ETA: 1:57 - loss: 0.4055 - acc: 0.826 - ETA: 1:48 - loss: 0.4109 - acc: 0.823 - ETA: 1:38 - loss: 0.4104 - acc: 0.822 - ETA: 1:29 - loss: 0.4087 - acc: 0.822 - ETA: 1:20 - loss: 0.4070 - acc: 0.822 - ETA: 1:10 - loss: 0.4101 - acc: 0.820 - ETA: 1:01 - loss: 0.4108 - acc: 0.820 - ETA: 51s - loss: 0.4105 - acc: 0.821 - ETA: 42s - loss: 0.4120 - acc: 0.82 - ETA: 33s - loss: 0.4162 - acc: 0.81 - ETA: 23s - loss: 0.4201 - acc: 0.81 - ETA: 14s - loss: 0.4219 - acc: 0.81 - ETA: 4s - loss: 0.4253 - acc: 0.8117 - 391s 147ms/step - loss: 0.4223 - acc: 0.8133\n",
      "Epoch 37/50\n",
      "2656/2656 [==============================] - ETA: 6:11 - loss: 0.5760 - acc: 0.687 - ETA: 6:05 - loss: 0.5004 - acc: 0.734 - ETA: 5:59 - loss: 0.4642 - acc: 0.776 - ETA: 5:50 - loss: 0.4731 - acc: 0.765 - ETA: 5:42 - loss: 0.4458 - acc: 0.781 - ETA: 5:33 - loss: 0.4350 - acc: 0.799 - ETA: 5:23 - loss: 0.4326 - acc: 0.799 - ETA: 5:14 - loss: 0.4191 - acc: 0.804 - ETA: 5:05 - loss: 0.4107 - acc: 0.809 - ETA: 4:56 - loss: 0.4150 - acc: 0.807 - ETA: 4:46 - loss: 0.4130 - acc: 0.809 - ETA: 4:37 - loss: 0.4082 - acc: 0.813 - ETA: 4:28 - loss: 0.4176 - acc: 0.810 - ETA: 4:20 - loss: 0.4249 - acc: 0.802 - ETA: 4:10 - loss: 0.4157 - acc: 0.807 - ETA: 4:01 - loss: 0.4148 - acc: 0.807 - ETA: 3:51 - loss: 0.4197 - acc: 0.806 - ETA: 3:41 - loss: 0.4223 - acc: 0.806 - ETA: 3:32 - loss: 0.4190 - acc: 0.809 - ETA: 3:23 - loss: 0.4148 - acc: 0.809 - ETA: 3:13 - loss: 0.4068 - acc: 0.814 - ETA: 3:04 - loss: 0.4047 - acc: 0.814 - ETA: 2:55 - loss: 0.4087 - acc: 0.813 - ETA: 2:46 - loss: 0.4125 - acc: 0.812 - ETA: 2:36 - loss: 0.4129 - acc: 0.811 - ETA: 2:27 - loss: 0.4142 - acc: 0.810 - ETA: 2:17 - loss: 0.4122 - acc: 0.810 - ETA: 2:08 - loss: 0.4142 - acc: 0.810 - ETA: 1:59 - loss: 0.4128 - acc: 0.812 - ETA: 1:49 - loss: 0.4108 - acc: 0.813 - ETA: 1:40 - loss: 0.4101 - acc: 0.813 - ETA: 1:30 - loss: 0.4075 - acc: 0.814 - ETA: 1:21 - loss: 0.4087 - acc: 0.812 - ETA: 1:11 - loss: 0.4091 - acc: 0.813 - ETA: 1:01 - loss: 0.4108 - acc: 0.813 - ETA: 52s - loss: 0.4100 - acc: 0.815 - ETA: 42s - loss: 0.4157 - acc: 0.81 - ETA: 33s - loss: 0.4205 - acc: 0.81 - ETA: 23s - loss: 0.4202 - acc: 0.81 - ETA: 14s - loss: 0.4243 - acc: 0.80 - ETA: 4s - loss: 0.4274 - acc: 0.8072 - 395s 149ms/step - loss: 0.4254 - acc: 0.8076\n",
      "Epoch 38/50\n",
      "2656/2656 [==============================] - ETA: 6:19 - loss: 0.2926 - acc: 0.859 - ETA: 6:10 - loss: 0.3309 - acc: 0.843 - ETA: 6:02 - loss: 0.3360 - acc: 0.838 - ETA: 5:52 - loss: 0.3462 - acc: 0.843 - ETA: 5:46 - loss: 0.3768 - acc: 0.825 - ETA: 5:36 - loss: 0.3917 - acc: 0.815 - ETA: 5:27 - loss: 0.3725 - acc: 0.823 - ETA: 5:18 - loss: 0.3800 - acc: 0.822 - ETA: 5:10 - loss: 0.3863 - acc: 0.819 - ETA: 5:00 - loss: 0.3898 - acc: 0.815 - ETA: 4:52 - loss: 0.3829 - acc: 0.821 - ETA: 4:43 - loss: 0.3848 - acc: 0.824 - ETA: 4:33 - loss: 0.3772 - acc: 0.828 - ETA: 4:23 - loss: 0.3826 - acc: 0.827 - ETA: 4:13 - loss: 0.3876 - acc: 0.826 - ETA: 4:04 - loss: 0.3960 - acc: 0.820 - ETA: 3:54 - loss: 0.3950 - acc: 0.819 - ETA: 3:44 - loss: 0.3914 - acc: 0.822 - ETA: 3:34 - loss: 0.3880 - acc: 0.823 - ETA: 3:25 - loss: 0.4010 - acc: 0.818 - ETA: 3:15 - loss: 0.4099 - acc: 0.815 - ETA: 3:06 - loss: 0.4136 - acc: 0.813 - ETA: 2:56 - loss: 0.4135 - acc: 0.813 - ETA: 2:46 - loss: 0.4126 - acc: 0.814 - ETA: 2:37 - loss: 0.4101 - acc: 0.816 - ETA: 2:27 - loss: 0.4171 - acc: 0.807 - ETA: 2:18 - loss: 0.4159 - acc: 0.808 - ETA: 2:08 - loss: 0.4195 - acc: 0.808 - ETA: 1:59 - loss: 0.4234 - acc: 0.807 - ETA: 1:49 - loss: 0.4260 - acc: 0.808 - ETA: 1:39 - loss: 0.4267 - acc: 0.807 - ETA: 1:30 - loss: 0.4279 - acc: 0.805 - ETA: 1:20 - loss: 0.4286 - acc: 0.806 - ETA: 1:11 - loss: 0.4287 - acc: 0.807 - ETA: 1:01 - loss: 0.4284 - acc: 0.807 - ETA: 52s - loss: 0.4290 - acc: 0.806 - ETA: 42s - loss: 0.4313 - acc: 0.80 - ETA: 33s - loss: 0.4328 - acc: 0.80 - ETA: 23s - loss: 0.4304 - acc: 0.80 - ETA: 14s - loss: 0.4309 - acc: 0.80 - ETA: 4s - loss: 0.4310 - acc: 0.8060 - 394s 148ms/step - loss: 0.4291 - acc: 0.8072\n",
      "Epoch 39/50\n",
      "2656/2656 [==============================] - ETA: 6:11 - loss: 0.3018 - acc: 0.890 - ETA: 5:56 - loss: 0.3257 - acc: 0.882 - ETA: 5:47 - loss: 0.3254 - acc: 0.875 - ETA: 5:37 - loss: 0.3460 - acc: 0.855 - ETA: 5:28 - loss: 0.3754 - acc: 0.843 - ETA: 5:20 - loss: 0.3913 - acc: 0.830 - ETA: 5:12 - loss: 0.3795 - acc: 0.830 - ETA: 5:03 - loss: 0.3896 - acc: 0.824 - ETA: 4:55 - loss: 0.3970 - acc: 0.822 - ETA: 4:46 - loss: 0.3980 - acc: 0.820 - ETA: 4:38 - loss: 0.3883 - acc: 0.821 - ETA: 4:29 - loss: 0.3936 - acc: 0.820 - ETA: 4:19 - loss: 0.4064 - acc: 0.812 - ETA: 4:10 - loss: 0.4179 - acc: 0.804 - ETA: 4:01 - loss: 0.4209 - acc: 0.803 - ETA: 3:53 - loss: 0.4140 - acc: 0.806 - ETA: 3:44 - loss: 0.4195 - acc: 0.804 - ETA: 3:35 - loss: 0.4222 - acc: 0.799 - ETA: 3:26 - loss: 0.4169 - acc: 0.802 - ETA: 3:16 - loss: 0.4116 - acc: 0.804 - ETA: 3:08 - loss: 0.4066 - acc: 0.807 - ETA: 2:59 - loss: 0.4106 - acc: 0.806 - ETA: 2:50 - loss: 0.4120 - acc: 0.805 - ETA: 2:41 - loss: 0.4145 - acc: 0.804 - ETA: 2:32 - loss: 0.4141 - acc: 0.804 - ETA: 2:23 - loss: 0.4126 - acc: 0.805 - ETA: 2:13 - loss: 0.4132 - acc: 0.806 - ETA: 2:04 - loss: 0.4110 - acc: 0.807 - ETA: 1:55 - loss: 0.4087 - acc: 0.808 - ETA: 1:46 - loss: 0.4052 - acc: 0.811 - ETA: 1:37 - loss: 0.4112 - acc: 0.809 - ETA: 1:27 - loss: 0.4180 - acc: 0.804 - ETA: 1:18 - loss: 0.4168 - acc: 0.806 - ETA: 1:09 - loss: 0.4150 - acc: 0.807 - ETA: 1:00 - loss: 0.4155 - acc: 0.807 - ETA: 50s - loss: 0.4194 - acc: 0.805 - ETA: 41s - loss: 0.4152 - acc: 0.80 - ETA: 32s - loss: 0.4151 - acc: 0.80 - ETA: 23s - loss: 0.4127 - acc: 0.80 - ETA: 13s - loss: 0.4123 - acc: 0.80 - ETA: 4s - loss: 0.4131 - acc: 0.8102 - 384s 145ms/step - loss: 0.4148 - acc: 0.8095\n",
      "Epoch 40/50\n",
      "2656/2656 [==============================] - ETA: 6:05 - loss: 0.4940 - acc: 0.765 - ETA: 5:54 - loss: 0.4275 - acc: 0.804 - ETA: 5:45 - loss: 0.4000 - acc: 0.833 - ETA: 5:38 - loss: 0.4178 - acc: 0.828 - ETA: 5:30 - loss: 0.3960 - acc: 0.834 - ETA: 5:21 - loss: 0.3976 - acc: 0.830 - ETA: 5:14 - loss: 0.3911 - acc: 0.832 - ETA: 5:07 - loss: 0.3935 - acc: 0.828 - ETA: 5:00 - loss: 0.3888 - acc: 0.829 - ETA: 4:51 - loss: 0.3912 - acc: 0.828 - ETA: 4:43 - loss: 0.3837 - acc: 0.835 - ETA: 4:34 - loss: 0.3893 - acc: 0.832 - ETA: 4:25 - loss: 0.3836 - acc: 0.835 - ETA: 4:16 - loss: 0.4030 - acc: 0.827 - ETA: 4:08 - loss: 0.4085 - acc: 0.826 - ETA: 3:59 - loss: 0.4081 - acc: 0.826 - ETA: 3:50 - loss: 0.4043 - acc: 0.827 - ETA: 3:41 - loss: 0.4012 - acc: 0.829 - ETA: 3:32 - loss: 0.4023 - acc: 0.828 - ETA: 3:23 - loss: 0.4019 - acc: 0.828 - ETA: 3:13 - loss: 0.4032 - acc: 0.828 - ETA: 3:04 - loss: 0.4045 - acc: 0.828 - ETA: 2:54 - loss: 0.4108 - acc: 0.824 - ETA: 2:45 - loss: 0.4157 - acc: 0.819 - ETA: 2:36 - loss: 0.4141 - acc: 0.821 - ETA: 2:26 - loss: 0.4148 - acc: 0.819 - ETA: 2:17 - loss: 0.4140 - acc: 0.817 - ETA: 2:07 - loss: 0.4128 - acc: 0.817 - ETA: 1:58 - loss: 0.4164 - acc: 0.814 - ETA: 1:49 - loss: 0.4097 - acc: 0.818 - ETA: 1:39 - loss: 0.4089 - acc: 0.818 - ETA: 1:30 - loss: 0.4067 - acc: 0.819 - ETA: 1:20 - loss: 0.4045 - acc: 0.820 - ETA: 1:11 - loss: 0.4056 - acc: 0.819 - ETA: 1:02 - loss: 0.4062 - acc: 0.819 - ETA: 52s - loss: 0.4051 - acc: 0.819 - ETA: 43s - loss: 0.4051 - acc: 0.81 - ETA: 33s - loss: 0.4140 - acc: 0.81 - ETA: 23s - loss: 0.4164 - acc: 0.81 - ETA: 14s - loss: 0.4170 - acc: 0.81 - ETA: 4s - loss: 0.4162 - acc: 0.8148 - 396s 149ms/step - loss: 0.4158 - acc: 0.8155\n",
      "Epoch 41/50\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2656/2656 [==============================] - ETA: 6:15 - loss: 0.4300 - acc: 0.796 - ETA: 6:10 - loss: 0.3610 - acc: 0.843 - ETA: 5:59 - loss: 0.3734 - acc: 0.838 - ETA: 5:48 - loss: 0.3749 - acc: 0.832 - ETA: 5:40 - loss: 0.3749 - acc: 0.825 - ETA: 5:30 - loss: 0.3555 - acc: 0.835 - ETA: 5:22 - loss: 0.3830 - acc: 0.821 - ETA: 5:13 - loss: 0.3938 - acc: 0.822 - ETA: 5:04 - loss: 0.3942 - acc: 0.824 - ETA: 4:55 - loss: 0.3769 - acc: 0.837 - ETA: 4:47 - loss: 0.3902 - acc: 0.826 - ETA: 4:38 - loss: 0.3974 - acc: 0.824 - ETA: 4:29 - loss: 0.3949 - acc: 0.825 - ETA: 4:20 - loss: 0.3837 - acc: 0.829 - ETA: 4:12 - loss: 0.3809 - acc: 0.831 - ETA: 4:02 - loss: 0.3919 - acc: 0.825 - ETA: 3:53 - loss: 0.3870 - acc: 0.828 - ETA: 3:44 - loss: 0.3857 - acc: 0.829 - ETA: 3:34 - loss: 0.3945 - acc: 0.825 - ETA: 3:25 - loss: 0.3976 - acc: 0.824 - ETA: 3:15 - loss: 0.4009 - acc: 0.822 - ETA: 3:05 - loss: 0.3993 - acc: 0.822 - ETA: 2:56 - loss: 0.4060 - acc: 0.817 - ETA: 2:46 - loss: 0.4043 - acc: 0.818 - ETA: 2:37 - loss: 0.4077 - acc: 0.816 - ETA: 2:27 - loss: 0.4108 - acc: 0.814 - ETA: 2:18 - loss: 0.4098 - acc: 0.816 - ETA: 2:08 - loss: 0.4077 - acc: 0.815 - ETA: 1:59 - loss: 0.4129 - acc: 0.814 - ETA: 1:49 - loss: 0.4115 - acc: 0.816 - ETA: 1:40 - loss: 0.4120 - acc: 0.815 - ETA: 1:30 - loss: 0.4084 - acc: 0.818 - ETA: 1:21 - loss: 0.4045 - acc: 0.821 - ETA: 1:11 - loss: 0.4080 - acc: 0.819 - ETA: 1:02 - loss: 0.4077 - acc: 0.820 - ETA: 52s - loss: 0.4108 - acc: 0.819 - ETA: 43s - loss: 0.4105 - acc: 0.81 - ETA: 33s - loss: 0.4139 - acc: 0.81 - ETA: 23s - loss: 0.4132 - acc: 0.81 - ETA: 14s - loss: 0.4121 - acc: 0.81 - ETA: 4s - loss: 0.4145 - acc: 0.8182 - 397s 149ms/step - loss: 0.4169 - acc: 0.8166\n",
      "Epoch 42/50\n",
      "2656/2656 [==============================] - ETA: 6:13 - loss: 0.4149 - acc: 0.812 - ETA: 6:07 - loss: 0.3837 - acc: 0.835 - ETA: 5:56 - loss: 0.4069 - acc: 0.828 - ETA: 5:46 - loss: 0.3774 - acc: 0.847 - ETA: 5:37 - loss: 0.3637 - acc: 0.853 - ETA: 5:29 - loss: 0.3845 - acc: 0.841 - ETA: 5:21 - loss: 0.3917 - acc: 0.832 - ETA: 5:12 - loss: 0.4116 - acc: 0.826 - ETA: 5:02 - loss: 0.4057 - acc: 0.828 - ETA: 4:54 - loss: 0.4113 - acc: 0.823 - ETA: 4:45 - loss: 0.4028 - acc: 0.826 - ETA: 4:36 - loss: 0.3971 - acc: 0.829 - ETA: 4:27 - loss: 0.3926 - acc: 0.830 - ETA: 4:18 - loss: 0.3931 - acc: 0.830 - ETA: 4:08 - loss: 0.3928 - acc: 0.830 - ETA: 3:58 - loss: 0.4002 - acc: 0.826 - ETA: 3:49 - loss: 0.3966 - acc: 0.824 - ETA: 3:39 - loss: 0.4014 - acc: 0.820 - ETA: 3:29 - loss: 0.4033 - acc: 0.818 - ETA: 3:20 - loss: 0.4046 - acc: 0.814 - ETA: 3:11 - loss: 0.4045 - acc: 0.817 - ETA: 3:01 - loss: 0.4107 - acc: 0.814 - ETA: 2:52 - loss: 0.4158 - acc: 0.812 - ETA: 2:42 - loss: 0.4121 - acc: 0.813 - ETA: 2:33 - loss: 0.4097 - acc: 0.815 - ETA: 2:24 - loss: 0.4086 - acc: 0.814 - ETA: 2:15 - loss: 0.4069 - acc: 0.814 - ETA: 2:05 - loss: 0.4115 - acc: 0.814 - ETA: 1:56 - loss: 0.4128 - acc: 0.813 - ETA: 1:47 - loss: 0.4136 - acc: 0.812 - ETA: 1:38 - loss: 0.4126 - acc: 0.811 - ETA: 1:28 - loss: 0.4140 - acc: 0.810 - ETA: 1:19 - loss: 0.4124 - acc: 0.811 - ETA: 1:10 - loss: 0.4101 - acc: 0.812 - ETA: 1:00 - loss: 0.4111 - acc: 0.810 - ETA: 51s - loss: 0.4090 - acc: 0.811 - ETA: 42s - loss: 0.4104 - acc: 0.81 - ETA: 32s - loss: 0.4102 - acc: 0.81 - ETA: 23s - loss: 0.4099 - acc: 0.81 - ETA: 14s - loss: 0.4072 - acc: 0.81 - ETA: 4s - loss: 0.4064 - acc: 0.8121 - 389s 146ms/step - loss: 0.4094 - acc: 0.8102\n",
      "Epoch 43/50\n",
      "2656/2656 [==============================] - ETA: 6:07 - loss: 0.4238 - acc: 0.796 - ETA: 5:58 - loss: 0.4311 - acc: 0.789 - ETA: 5:53 - loss: 0.4249 - acc: 0.812 - ETA: 5:46 - loss: 0.4083 - acc: 0.824 - ETA: 5:38 - loss: 0.4056 - acc: 0.825 - ETA: 5:29 - loss: 0.4356 - acc: 0.809 - ETA: 5:21 - loss: 0.4403 - acc: 0.812 - ETA: 5:12 - loss: 0.4317 - acc: 0.820 - ETA: 5:04 - loss: 0.4267 - acc: 0.817 - ETA: 4:54 - loss: 0.4204 - acc: 0.820 - ETA: 4:45 - loss: 0.4191 - acc: 0.816 - ETA: 4:36 - loss: 0.4069 - acc: 0.824 - ETA: 4:27 - loss: 0.3997 - acc: 0.824 - ETA: 4:18 - loss: 0.4035 - acc: 0.822 - ETA: 4:08 - loss: 0.4087 - acc: 0.822 - ETA: 3:58 - loss: 0.4117 - acc: 0.820 - ETA: 3:49 - loss: 0.4175 - acc: 0.816 - ETA: 3:40 - loss: 0.4133 - acc: 0.816 - ETA: 3:31 - loss: 0.4111 - acc: 0.817 - ETA: 3:21 - loss: 0.4033 - acc: 0.820 - ETA: 3:12 - loss: 0.4027 - acc: 0.821 - ETA: 3:03 - loss: 0.3988 - acc: 0.823 - ETA: 2:54 - loss: 0.3996 - acc: 0.826 - ETA: 2:44 - loss: 0.4007 - acc: 0.824 - ETA: 2:35 - loss: 0.4019 - acc: 0.823 - ETA: 2:26 - loss: 0.4087 - acc: 0.822 - ETA: 2:16 - loss: 0.4083 - acc: 0.822 - ETA: 2:07 - loss: 0.4048 - acc: 0.825 - ETA: 1:57 - loss: 0.4031 - acc: 0.826 - ETA: 1:48 - loss: 0.4065 - acc: 0.825 - ETA: 1:39 - loss: 0.4119 - acc: 0.822 - ETA: 1:29 - loss: 0.4124 - acc: 0.821 - ETA: 1:20 - loss: 0.4082 - acc: 0.824 - ETA: 1:11 - loss: 0.4074 - acc: 0.824 - ETA: 1:01 - loss: 0.4085 - acc: 0.822 - ETA: 52s - loss: 0.4119 - acc: 0.818 - ETA: 42s - loss: 0.4146 - acc: 0.81 - ETA: 33s - loss: 0.4139 - acc: 0.81 - ETA: 23s - loss: 0.4113 - acc: 0.81 - ETA: 14s - loss: 0.4149 - acc: 0.81 - ETA: 4s - loss: 0.4149 - acc: 0.8175 - 394s 148ms/step - loss: 0.4161 - acc: 0.8155\n",
      "Epoch 44/50\n",
      "2656/2656 [==============================] - ETA: 6:21 - loss: 0.4265 - acc: 0.828 - ETA: 6:08 - loss: 0.4104 - acc: 0.804 - ETA: 5:59 - loss: 0.4066 - acc: 0.822 - ETA: 5:49 - loss: 0.3880 - acc: 0.835 - ETA: 5:42 - loss: 0.3800 - acc: 0.834 - ETA: 5:34 - loss: 0.3706 - acc: 0.838 - ETA: 5:25 - loss: 0.3666 - acc: 0.846 - ETA: 5:15 - loss: 0.3601 - acc: 0.847 - ETA: 5:06 - loss: 0.3659 - acc: 0.845 - ETA: 4:58 - loss: 0.3748 - acc: 0.837 - ETA: 4:49 - loss: 0.3678 - acc: 0.839 - ETA: 4:40 - loss: 0.3752 - acc: 0.837 - ETA: 4:31 - loss: 0.3745 - acc: 0.842 - ETA: 4:22 - loss: 0.3834 - acc: 0.834 - ETA: 4:13 - loss: 0.3833 - acc: 0.833 - ETA: 4:04 - loss: 0.3765 - acc: 0.836 - ETA: 3:55 - loss: 0.3789 - acc: 0.838 - ETA: 3:45 - loss: 0.3791 - acc: 0.836 - ETA: 3:36 - loss: 0.3846 - acc: 0.835 - ETA: 3:26 - loss: 0.3862 - acc: 0.833 - ETA: 3:17 - loss: 0.3919 - acc: 0.830 - ETA: 3:07 - loss: 0.3992 - acc: 0.826 - ETA: 2:58 - loss: 0.4035 - acc: 0.824 - ETA: 2:48 - loss: 0.4002 - acc: 0.826 - ETA: 2:39 - loss: 0.4068 - acc: 0.823 - ETA: 2:29 - loss: 0.4013 - acc: 0.825 - ETA: 2:20 - loss: 0.3989 - acc: 0.825 - ETA: 2:10 - loss: 0.3977 - acc: 0.825 - ETA: 2:01 - loss: 0.3984 - acc: 0.825 - ETA: 1:51 - loss: 0.4021 - acc: 0.823 - ETA: 1:41 - loss: 0.4057 - acc: 0.821 - ETA: 1:32 - loss: 0.4088 - acc: 0.817 - ETA: 1:22 - loss: 0.4088 - acc: 0.818 - ETA: 1:12 - loss: 0.4104 - acc: 0.817 - ETA: 1:03 - loss: 0.4144 - acc: 0.816 - ETA: 53s - loss: 0.4133 - acc: 0.817 - ETA: 43s - loss: 0.4119 - acc: 0.81 - ETA: 34s - loss: 0.4151 - acc: 0.81 - ETA: 24s - loss: 0.4191 - acc: 0.81 - ETA: 14s - loss: 0.4185 - acc: 0.81 - ETA: 4s - loss: 0.4169 - acc: 0.8167 - 404s 152ms/step - loss: 0.4167 - acc: 0.8170\n",
      "Epoch 45/50\n",
      "2656/2656 [==============================] - ETA: 6:30 - loss: 0.3263 - acc: 0.890 - ETA: 6:18 - loss: 0.3580 - acc: 0.867 - ETA: 6:12 - loss: 0.3832 - acc: 0.833 - ETA: 6:04 - loss: 0.4144 - acc: 0.812 - ETA: 5:56 - loss: 0.3961 - acc: 0.818 - ETA: 5:46 - loss: 0.4050 - acc: 0.817 - ETA: 5:36 - loss: 0.3918 - acc: 0.819 - ETA: 5:27 - loss: 0.3855 - acc: 0.828 - ETA: 5:18 - loss: 0.3770 - acc: 0.833 - ETA: 5:09 - loss: 0.3689 - acc: 0.839 - ETA: 4:58 - loss: 0.3772 - acc: 0.832 - ETA: 4:49 - loss: 0.3777 - acc: 0.829 - ETA: 4:40 - loss: 0.3775 - acc: 0.831 - ETA: 4:30 - loss: 0.3776 - acc: 0.832 - ETA: 4:20 - loss: 0.3749 - acc: 0.832 - ETA: 4:09 - loss: 0.3901 - acc: 0.826 - ETA: 4:00 - loss: 0.3906 - acc: 0.825 - ETA: 3:50 - loss: 0.3956 - acc: 0.822 - ETA: 3:40 - loss: 0.3982 - acc: 0.820 - ETA: 3:30 - loss: 0.3976 - acc: 0.819 - ETA: 3:20 - loss: 0.3914 - acc: 0.825 - ETA: 3:10 - loss: 0.3912 - acc: 0.825 - ETA: 3:00 - loss: 0.3953 - acc: 0.823 - ETA: 2:51 - loss: 0.3899 - acc: 0.825 - ETA: 2:41 - loss: 0.3930 - acc: 0.824 - ETA: 2:31 - loss: 0.3914 - acc: 0.825 - ETA: 2:21 - loss: 0.3977 - acc: 0.820 - ETA: 2:12 - loss: 0.3978 - acc: 0.820 - ETA: 2:02 - loss: 0.4053 - acc: 0.818 - ETA: 1:52 - loss: 0.4044 - acc: 0.817 - ETA: 1:42 - loss: 0.4045 - acc: 0.818 - ETA: 1:33 - loss: 0.4033 - acc: 0.820 - ETA: 1:23 - loss: 0.4062 - acc: 0.817 - ETA: 1:13 - loss: 0.4036 - acc: 0.818 - ETA: 1:03 - loss: 0.4048 - acc: 0.818 - ETA: 53s - loss: 0.4072 - acc: 0.818 - ETA: 44s - loss: 0.4110 - acc: 0.81 - ETA: 34s - loss: 0.4160 - acc: 0.81 - ETA: 24s - loss: 0.4138 - acc: 0.81 - ETA: 14s - loss: 0.4132 - acc: 0.81 - ETA: 4s - loss: 0.4169 - acc: 0.8155 - 407s 153ms/step - loss: 0.4164 - acc: 0.8155\n",
      "Epoch 46/50\n",
      "2656/2656 [==============================] - ETA: 6:34 - loss: 0.3411 - acc: 0.828 - ETA: 6:27 - loss: 0.4114 - acc: 0.796 - ETA: 6:17 - loss: 0.4067 - acc: 0.817 - ETA: 6:10 - loss: 0.4031 - acc: 0.812 - ETA: 5:59 - loss: 0.3882 - acc: 0.821 - ETA: 5:50 - loss: 0.3827 - acc: 0.822 - ETA: 5:40 - loss: 0.3555 - acc: 0.839 - ETA: 5:32 - loss: 0.3571 - acc: 0.841 - ETA: 5:22 - loss: 0.3510 - acc: 0.843 - ETA: 5:13 - loss: 0.3529 - acc: 0.842 - ETA: 5:03 - loss: 0.3498 - acc: 0.840 - ETA: 4:53 - loss: 0.3520 - acc: 0.837 - ETA: 4:43 - loss: 0.3527 - acc: 0.836 - ETA: 4:34 - loss: 0.3522 - acc: 0.839 - ETA: 4:24 - loss: 0.3738 - acc: 0.828 - ETA: 4:14 - loss: 0.3743 - acc: 0.828 - ETA: 4:04 - loss: 0.3765 - acc: 0.828 - ETA: 3:54 - loss: 0.3789 - acc: 0.826 - ETA: 3:44 - loss: 0.3792 - acc: 0.828 - ETA: 3:34 - loss: 0.3855 - acc: 0.823 - ETA: 3:24 - loss: 0.3876 - acc: 0.822 - ETA: 3:14 - loss: 0.3895 - acc: 0.823 - ETA: 3:04 - loss: 0.4025 - acc: 0.816 - ETA: 2:54 - loss: 0.4012 - acc: 0.816 - ETA: 2:44 - loss: 0.3973 - acc: 0.818 - ETA: 2:34 - loss: 0.3961 - acc: 0.819 - ETA: 2:24 - loss: 0.3977 - acc: 0.817 - ETA: 2:14 - loss: 0.4000 - acc: 0.815 - ETA: 2:04 - loss: 0.3961 - acc: 0.817 - ETA: 1:54 - loss: 0.4012 - acc: 0.816 - ETA: 1:44 - loss: 0.4008 - acc: 0.815 - ETA: 1:34 - loss: 0.4021 - acc: 0.814 - ETA: 1:24 - loss: 0.4000 - acc: 0.816 - ETA: 1:14 - loss: 0.4001 - acc: 0.816 - ETA: 1:04 - loss: 0.4016 - acc: 0.815 - ETA: 54s - loss: 0.4024 - acc: 0.814 - ETA: 44s - loss: 0.4067 - acc: 0.81 - ETA: 34s - loss: 0.4058 - acc: 0.81 - ETA: 24s - loss: 0.4073 - acc: 0.81 - ETA: 14s - loss: 0.4075 - acc: 0.81 - ETA: 4s - loss: 0.4081 - acc: 0.8121 - 411s 155ms/step - loss: 0.4092 - acc: 0.8117\n",
      "Epoch 47/50\n",
      "2656/2656 [==============================] - ETA: 6:34 - loss: 0.4119 - acc: 0.812 - ETA: 6:32 - loss: 0.3702 - acc: 0.851 - ETA: 6:24 - loss: 0.4358 - acc: 0.822 - ETA: 6:12 - loss: 0.4127 - acc: 0.835 - ETA: 6:01 - loss: 0.3969 - acc: 0.834 - ETA: 5:51 - loss: 0.3768 - acc: 0.841 - ETA: 5:41 - loss: 0.3765 - acc: 0.841 - ETA: 5:31 - loss: 0.3663 - acc: 0.847 - ETA: 5:19 - loss: 0.3655 - acc: 0.845 - ETA: 5:08 - loss: 0.3709 - acc: 0.845 - ETA: 4:58 - loss: 0.3814 - acc: 0.845 - ETA: 4:48 - loss: 0.3868 - acc: 0.842 - ETA: 4:40 - loss: 0.3859 - acc: 0.842 - ETA: 4:30 - loss: 0.3806 - acc: 0.844 - ETA: 4:20 - loss: 0.3793 - acc: 0.844 - ETA: 4:10 - loss: 0.3807 - acc: 0.842 - ETA: 4:00 - loss: 0.3792 - acc: 0.842 - ETA: 3:50 - loss: 0.3793 - acc: 0.840 - ETA: 3:40 - loss: 0.3793 - acc: 0.839 - ETA: 3:31 - loss: 0.3795 - acc: 0.835 - ETA: 3:20 - loss: 0.3816 - acc: 0.835 - ETA: 3:10 - loss: 0.3839 - acc: 0.834 - ETA: 3:00 - loss: 0.3868 - acc: 0.832 - ETA: 2:51 - loss: 0.3838 - acc: 0.834 - ETA: 2:41 - loss: 0.3880 - acc: 0.830 - ETA: 2:31 - loss: 0.3869 - acc: 0.829 - ETA: 2:21 - loss: 0.3906 - acc: 0.828 - ETA: 2:12 - loss: 0.3932 - acc: 0.825 - ETA: 2:02 - loss: 0.3954 - acc: 0.826 - ETA: 1:52 - loss: 0.3979 - acc: 0.825 - ETA: 1:42 - loss: 0.3988 - acc: 0.823 - ETA: 1:32 - loss: 0.4003 - acc: 0.822 - ETA: 1:23 - loss: 0.4043 - acc: 0.821 - ETA: 1:13 - loss: 0.4062 - acc: 0.819 - ETA: 1:03 - loss: 0.4043 - acc: 0.820 - ETA: 53s - loss: 0.4061 - acc: 0.818 - ETA: 43s - loss: 0.4092 - acc: 0.81 - ETA: 34s - loss: 0.4108 - acc: 0.81 - ETA: 24s - loss: 0.4073 - acc: 0.81 - ETA: 14s - loss: 0.4096 - acc: 0.81 - ETA: 4s - loss: 0.4124 - acc: 0.8159 - 403s 152ms/step - loss: 0.4106 - acc: 0.8170\n",
      "Epoch 48/50\n",
      "2656/2656 [==============================] - ETA: 6:11 - loss: 0.4300 - acc: 0.781 - ETA: 6:03 - loss: 0.3969 - acc: 0.820 - ETA: 5:52 - loss: 0.4358 - acc: 0.802 - ETA: 5:44 - loss: 0.4220 - acc: 0.812 - ETA: 5:36 - loss: 0.4064 - acc: 0.825 - ETA: 5:27 - loss: 0.4009 - acc: 0.828 - ETA: 5:19 - loss: 0.3957 - acc: 0.830 - ETA: 5:11 - loss: 0.4005 - acc: 0.822 - ETA: 5:02 - loss: 0.3854 - acc: 0.829 - ETA: 4:54 - loss: 0.3768 - acc: 0.835 - ETA: 4:46 - loss: 0.3661 - acc: 0.840 - ETA: 4:38 - loss: 0.3611 - acc: 0.847 - ETA: 4:30 - loss: 0.3692 - acc: 0.842 - ETA: 4:21 - loss: 0.3614 - acc: 0.846 - ETA: 4:11 - loss: 0.3556 - acc: 0.849 - ETA: 4:02 - loss: 0.3606 - acc: 0.843 - ETA: 3:53 - loss: 0.3726 - acc: 0.841 - ETA: 3:44 - loss: 0.3839 - acc: 0.834 - ETA: 3:35 - loss: 0.3812 - acc: 0.835 - ETA: 3:25 - loss: 0.3785 - acc: 0.835 - ETA: 3:16 - loss: 0.3776 - acc: 0.835 - ETA: 3:07 - loss: 0.3801 - acc: 0.832 - ETA: 2:58 - loss: 0.3840 - acc: 0.830 - ETA: 2:48 - loss: 0.3800 - acc: 0.831 - ETA: 2:39 - loss: 0.3831 - acc: 0.829 - ETA: 2:29 - loss: 0.3841 - acc: 0.827 - ETA: 2:19 - loss: 0.3882 - acc: 0.824 - ETA: 2:10 - loss: 0.3905 - acc: 0.823 - ETA: 2:01 - loss: 0.3949 - acc: 0.821 - ETA: 1:51 - loss: 0.4011 - acc: 0.817 - ETA: 1:42 - loss: 0.4000 - acc: 0.818 - ETA: 1:32 - loss: 0.3997 - acc: 0.817 - ETA: 1:22 - loss: 0.4028 - acc: 0.816 - ETA: 1:13 - loss: 0.4048 - acc: 0.815 - ETA: 1:03 - loss: 0.4018 - acc: 0.817 - ETA: 53s - loss: 0.3986 - acc: 0.819 - ETA: 44s - loss: 0.3970 - acc: 0.81 - ETA: 34s - loss: 0.4012 - acc: 0.81 - ETA: 24s - loss: 0.4031 - acc: 0.81 - ETA: 14s - loss: 0.4059 - acc: 0.81 - ETA: 4s - loss: 0.4066 - acc: 0.8125 - 407s 153ms/step - loss: 0.4049 - acc: 0.8133\n",
      "Epoch 49/50\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2656/2656 [==============================] - ETA: 6:44 - loss: 0.3671 - acc: 0.843 - ETA: 6:36 - loss: 0.4031 - acc: 0.828 - ETA: 6:24 - loss: 0.4088 - acc: 0.828 - ETA: 6:16 - loss: 0.3866 - acc: 0.843 - ETA: 6:08 - loss: 0.3901 - acc: 0.840 - ETA: 5:59 - loss: 0.4037 - acc: 0.822 - ETA: 5:49 - loss: 0.4017 - acc: 0.821 - ETA: 5:39 - loss: 0.3943 - acc: 0.824 - ETA: 5:29 - loss: 0.3916 - acc: 0.826 - ETA: 5:19 - loss: 0.4054 - acc: 0.820 - ETA: 5:10 - loss: 0.3990 - acc: 0.823 - ETA: 4:59 - loss: 0.4085 - acc: 0.819 - ETA: 4:49 - loss: 0.4099 - acc: 0.820 - ETA: 4:38 - loss: 0.4050 - acc: 0.822 - ETA: 4:28 - loss: 0.4058 - acc: 0.821 - ETA: 4:17 - loss: 0.4077 - acc: 0.822 - ETA: 4:07 - loss: 0.4048 - acc: 0.823 - ETA: 3:57 - loss: 0.4092 - acc: 0.821 - ETA: 3:48 - loss: 0.4041 - acc: 0.822 - ETA: 3:37 - loss: 0.4070 - acc: 0.820 - ETA: 3:27 - loss: 0.4057 - acc: 0.820 - ETA: 3:17 - loss: 0.4101 - acc: 0.818 - ETA: 3:07 - loss: 0.4119 - acc: 0.820 - ETA: 2:58 - loss: 0.4221 - acc: 0.814 - ETA: 2:48 - loss: 0.4159 - acc: 0.816 - ETA: 2:38 - loss: 0.4157 - acc: 0.815 - ETA: 2:27 - loss: 0.4142 - acc: 0.814 - ETA: 2:17 - loss: 0.4124 - acc: 0.816 - ETA: 2:07 - loss: 0.4116 - acc: 0.816 - ETA: 1:57 - loss: 0.4102 - acc: 0.817 - ETA: 1:47 - loss: 0.4126 - acc: 0.815 - ETA: 1:37 - loss: 0.4130 - acc: 0.815 - ETA: 1:27 - loss: 0.4096 - acc: 0.817 - ETA: 1:16 - loss: 0.4080 - acc: 0.816 - ETA: 1:06 - loss: 0.4081 - acc: 0.815 - ETA: 56s - loss: 0.4100 - acc: 0.814 - ETA: 45s - loss: 0.4089 - acc: 0.81 - ETA: 35s - loss: 0.4072 - acc: 0.81 - ETA: 25s - loss: 0.4057 - acc: 0.81 - ETA: 15s - loss: 0.4046 - acc: 0.81 - ETA: 5s - loss: 0.4028 - acc: 0.8186 - 422s 159ms/step - loss: 0.4034 - acc: 0.8185\n",
      "Epoch 50/50\n",
      "2656/2656 [==============================] - ETA: 6:46 - loss: 0.3619 - acc: 0.859 - ETA: 6:34 - loss: 0.3356 - acc: 0.859 - ETA: 6:24 - loss: 0.3431 - acc: 0.854 - ETA: 6:15 - loss: 0.3792 - acc: 0.832 - ETA: 6:06 - loss: 0.3690 - acc: 0.834 - ETA: 5:55 - loss: 0.3359 - acc: 0.851 - ETA: 5:45 - loss: 0.3366 - acc: 0.852 - ETA: 5:34 - loss: 0.3422 - acc: 0.847 - ETA: 5:24 - loss: 0.3484 - acc: 0.842 - ETA: 5:13 - loss: 0.3420 - acc: 0.845 - ETA: 5:03 - loss: 0.3396 - acc: 0.845 - ETA: 4:53 - loss: 0.3401 - acc: 0.845 - ETA: 4:43 - loss: 0.3434 - acc: 0.841 - ETA: 4:33 - loss: 0.3439 - acc: 0.839 - ETA: 4:23 - loss: 0.3442 - acc: 0.841 - ETA: 4:13 - loss: 0.3546 - acc: 0.836 - ETA: 4:03 - loss: 0.3603 - acc: 0.832 - ETA: 3:53 - loss: 0.3534 - acc: 0.835 - ETA: 3:44 - loss: 0.3594 - acc: 0.831 - ETA: 3:34 - loss: 0.3666 - acc: 0.830 - ETA: 3:24 - loss: 0.3719 - acc: 0.828 - ETA: 3:14 - loss: 0.3740 - acc: 0.826 - ETA: 3:04 - loss: 0.3742 - acc: 0.824 - ETA: 2:54 - loss: 0.3851 - acc: 0.821 - ETA: 2:44 - loss: 0.3857 - acc: 0.819 - ETA: 2:34 - loss: 0.3859 - acc: 0.820 - ETA: 2:25 - loss: 0.3849 - acc: 0.820 - ETA: 2:15 - loss: 0.3882 - acc: 0.818 - ETA: 2:05 - loss: 0.3923 - acc: 0.815 - ETA: 1:55 - loss: 0.3915 - acc: 0.816 - ETA: 1:45 - loss: 0.3932 - acc: 0.816 - ETA: 1:35 - loss: 0.3947 - acc: 0.814 - ETA: 1:25 - loss: 0.3950 - acc: 0.815 - ETA: 1:15 - loss: 0.4016 - acc: 0.812 - ETA: 1:05 - loss: 0.3989 - acc: 0.812 - ETA: 55s - loss: 0.3965 - acc: 0.812 - ETA: 45s - loss: 0.4041 - acc: 0.80 - ETA: 35s - loss: 0.4044 - acc: 0.80 - ETA: 25s - loss: 0.4056 - acc: 0.80 - ETA: 15s - loss: 0.4037 - acc: 0.80 - ETA: 5s - loss: 0.4028 - acc: 0.8075 - 416s 157ms/step - loss: 0.4019 - acc: 0.8080\n",
      "Loss: 1.72%\n",
      "Accuracy: 56.69%\n"
     ]
    }
   ],
   "source": [
    "model.fit(X_train, Y_train, nb_epoch=50, batch_size=64)\n",
    "# Final evaluation of the model\n",
    "scores = model.evaluate(X_test, Y_test, verbose=0)\n",
    "print(\"Loss: %.2f%%\" % (scores[0]))\n",
    "print(\"Accuracy: %.2f%%\" % (scores[1]*100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}