Classify_points.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 258,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##  Loading the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 259,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = pd.read_csv(\"Logistic_X_Train.csv\").values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 260,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y_train = pd.read_csv(\"Logistic_Y_Train.csv\").values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 261,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test = pd.read_csv(\"Logistic_X_Test.csv\").values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 262,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2250, 2)"
      ]
     },
     "execution_count": 262,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 263,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2250, 1)"
      ]
     },
     "execution_count": 263,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Y_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 264,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(750, 2)"
      ]
     },
     "execution_count": 264,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 265,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.style.use(\"seaborn\")\n",
    "plt.scatter(X_train[:,0],X_train[:,1],c=Y_train,cmap=plt.cm.Accent)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 266,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras import models\n",
    "from keras.layers import Dense"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 267,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = models.Sequential()\n",
    "model.add(Dense(16,activation='relu',input_shape = (2,)))\n",
    "model.add(Dense(16,activation='relu'))\n",
    "model.add(Dense(1,activation='sigmoid'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 268,
   "metadata": {},
   "outputs": [],
   "source": [
    "#model compile\n",
    "model.compile(optimizer='rmsprop',loss='binary_crossentropy',metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 269,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_5\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_13 (Dense)             (None, 16)                48        \n",
      "_________________________________________________________________\n",
      "dense_14 (Dense)             (None, 16)                272       \n",
      "_________________________________________________________________\n",
      "dense_15 (Dense)             (None, 1)                 17        \n",
      "=================================================================\n",
      "Total params: 337\n",
      "Trainable params: 337\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 270,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_val = X_train[:1125]\n",
    "X_train_new = X_train[1125:]\n",
    "\n",
    "Y_val = Y_train[:1125]\n",
    "Y_train_new = Y_train[1125:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 271,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 1125 samples, validate on 1125 samples\n",
      "Epoch 1/700\n",
      "1125/1125 [==============================] - 0s 310us/step - loss: 0.7154 - accuracy: 0.4889 - val_loss: 0.6878 - val_accuracy: 0.5218\n",
      "Epoch 2/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.6898 - accuracy: 0.4747 - val_loss: 0.6714 - val_accuracy: 0.5280\n",
      "Epoch 3/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.6736 - accuracy: 0.4880 - val_loss: 0.6574 - val_accuracy: 0.5076\n",
      "Epoch 4/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.6596 - accuracy: 0.4667 - val_loss: 0.6452 - val_accuracy: 0.5093\n",
      "Epoch 5/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.6473 - accuracy: 0.4640 - val_loss: 0.6335 - val_accuracy: 0.4844\n",
      "Epoch 6/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.6355 - accuracy: 0.4658 - val_loss: 0.6236 - val_accuracy: 0.4747\n",
      "Epoch 7/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.6252 - accuracy: 0.4764 - val_loss: 0.6132 - val_accuracy: 0.5138\n",
      "Epoch 8/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.6147 - accuracy: 0.5467 - val_loss: 0.6029 - val_accuracy: 0.6898\n",
      "Epoch 9/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.6043 - accuracy: 0.7413 - val_loss: 0.5934 - val_accuracy: 0.7840\n",
      "Epoch 10/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.5943 - accuracy: 0.8036 - val_loss: 0.5839 - val_accuracy: 0.7973\n",
      "Epoch 11/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.5845 - accuracy: 0.8107 - val_loss: 0.5743 - val_accuracy: 0.8018\n",
      "Epoch 12/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.5748 - accuracy: 0.8116 - val_loss: 0.5655 - val_accuracy: 0.8000\n",
      "Epoch 13/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.5657 - accuracy: 0.8116 - val_loss: 0.5566 - val_accuracy: 0.8044\n",
      "Epoch 14/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.5571 - accuracy: 0.8133 - val_loss: 0.5484 - val_accuracy: 0.8044\n",
      "Epoch 15/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.5487 - accuracy: 0.8142 - val_loss: 0.5401 - val_accuracy: 0.8036\n",
      "Epoch 16/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.5403 - accuracy: 0.8124 - val_loss: 0.5322 - val_accuracy: 0.8036\n",
      "Epoch 17/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.5323 - accuracy: 0.8124 - val_loss: 0.5245 - val_accuracy: 0.8036\n",
      "Epoch 18/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.5245 - accuracy: 0.8133 - val_loss: 0.5171 - val_accuracy: 0.8027\n",
      "Epoch 19/700\n",
      "1125/1125 [==============================] - 0s 22us/step - loss: 0.5168 - accuracy: 0.8133 - val_loss: 0.5096 - val_accuracy: 0.8062\n",
      "Epoch 20/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.5093 - accuracy: 0.8169 - val_loss: 0.5024 - val_accuracy: 0.8071\n",
      "Epoch 21/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.5019 - accuracy: 0.8169 - val_loss: 0.4961 - val_accuracy: 0.8089\n",
      "Epoch 22/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.4953 - accuracy: 0.8169 - val_loss: 0.4894 - val_accuracy: 0.8080\n",
      "Epoch 23/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.4885 - accuracy: 0.8169 - val_loss: 0.4833 - val_accuracy: 0.8098\n",
      "Epoch 24/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.4821 - accuracy: 0.8169 - val_loss: 0.4771 - val_accuracy: 0.8107\n",
      "Epoch 25/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.4756 - accuracy: 0.8169 - val_loss: 0.4710 - val_accuracy: 0.8116\n",
      "Epoch 26/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.4693 - accuracy: 0.8187 - val_loss: 0.4647 - val_accuracy: 0.8124\n",
      "Epoch 27/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.4630 - accuracy: 0.8222 - val_loss: 0.4587 - val_accuracy: 0.8124\n",
      "Epoch 28/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.4569 - accuracy: 0.8231 - val_loss: 0.4531 - val_accuracy: 0.8133\n",
      "Epoch 29/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.4510 - accuracy: 0.8231 - val_loss: 0.4476 - val_accuracy: 0.8133\n",
      "Epoch 30/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.4453 - accuracy: 0.8267 - val_loss: 0.4421 - val_accuracy: 0.8133\n",
      "Epoch 31/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.4394 - accuracy: 0.8267 - val_loss: 0.4361 - val_accuracy: 0.8124\n",
      "Epoch 32/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.4336 - accuracy: 0.8302 - val_loss: 0.4315 - val_accuracy: 0.8133\n",
      "Epoch 33/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.4286 - accuracy: 0.8311 - val_loss: 0.4264 - val_accuracy: 0.8133\n",
      "Epoch 34/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.4232 - accuracy: 0.8311 - val_loss: 0.4214 - val_accuracy: 0.8151\n",
      "Epoch 35/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.4181 - accuracy: 0.8320 - val_loss: 0.4169 - val_accuracy: 0.8151\n",
      "Epoch 36/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.4130 - accuracy: 0.8320 - val_loss: 0.4123 - val_accuracy: 0.8160\n",
      "Epoch 37/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.4082 - accuracy: 0.8320 - val_loss: 0.4077 - val_accuracy: 0.8169\n",
      "Epoch 38/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.4036 - accuracy: 0.8338 - val_loss: 0.4036 - val_accuracy: 0.8196\n",
      "Epoch 39/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.3995 - accuracy: 0.8356 - val_loss: 0.4000 - val_accuracy: 0.8196\n",
      "Epoch 40/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.3953 - accuracy: 0.8356 - val_loss: 0.3964 - val_accuracy: 0.8204\n",
      "Epoch 41/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.3912 - accuracy: 0.8373 - val_loss: 0.3921 - val_accuracy: 0.8231\n",
      "Epoch 42/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.3869 - accuracy: 0.8409 - val_loss: 0.3884 - val_accuracy: 0.8249\n",
      "Epoch 43/700\n",
      "1125/1125 [==============================] - ETA: 0s - loss: 0.3827 - accuracy: 0.83 - 0s 19us/step - loss: 0.3828 - accuracy: 0.8409 - val_loss: 0.3845 - val_accuracy: 0.8258\n",
      "Epoch 44/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.3788 - accuracy: 0.8418 - val_loss: 0.3813 - val_accuracy: 0.8258\n",
      "Epoch 45/700\n",
      "1125/1125 [==============================] - 0s 22us/step - loss: 0.3754 - accuracy: 0.8418 - val_loss: 0.3780 - val_accuracy: 0.8276\n",
      "Epoch 46/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.3715 - accuracy: 0.8418 - val_loss: 0.3747 - val_accuracy: 0.8293\n",
      "Epoch 47/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.3677 - accuracy: 0.8418 - val_loss: 0.3712 - val_accuracy: 0.8302\n",
      "Epoch 48/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.3642 - accuracy: 0.8453 - val_loss: 0.3680 - val_accuracy: 0.8338\n",
      "Epoch 49/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.3605 - accuracy: 0.8462 - val_loss: 0.3644 - val_accuracy: 0.8338\n",
      "Epoch 50/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.3567 - accuracy: 0.8489 - val_loss: 0.3616 - val_accuracy: 0.8364\n",
      "Epoch 51/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.3535 - accuracy: 0.8471 - val_loss: 0.3586 - val_accuracy: 0.8373\n",
      "Epoch 52/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.3500 - accuracy: 0.8480 - val_loss: 0.3551 - val_accuracy: 0.8391\n",
      "Epoch 53/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.3470 - accuracy: 0.8516 - val_loss: 0.3523 - val_accuracy: 0.8409\n",
      "Epoch 54/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.3438 - accuracy: 0.8516 - val_loss: 0.3498 - val_accuracy: 0.8418\n",
      "Epoch 55/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.3406 - accuracy: 0.8524 - val_loss: 0.3466 - val_accuracy: 0.8391\n",
      "Epoch 56/700\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.3378 - accuracy: 0.8524 - val_loss: 0.3438 - val_accuracy: 0.8409\n",
      "Epoch 57/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.3350 - accuracy: 0.8524 - val_loss: 0.3417 - val_accuracy: 0.8436\n",
      "Epoch 58/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.3323 - accuracy: 0.8542 - val_loss: 0.3393 - val_accuracy: 0.8436\n",
      "Epoch 59/700\n",
      "1125/1125 [==============================] - 0s 22us/step - loss: 0.3294 - accuracy: 0.8560 - val_loss: 0.3367 - val_accuracy: 0.8444\n",
      "Epoch 60/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.3270 - accuracy: 0.8604 - val_loss: 0.3342 - val_accuracy: 0.8453\n",
      "Epoch 61/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.3245 - accuracy: 0.8596 - val_loss: 0.3325 - val_accuracy: 0.8436\n",
      "Epoch 62/700\n",
      "1125/1125 [==============================] - 0s 25us/step - loss: 0.3220 - accuracy: 0.8604 - val_loss: 0.3296 - val_accuracy: 0.8489\n",
      "Epoch 63/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.3197 - accuracy: 0.8604 - val_loss: 0.3269 - val_accuracy: 0.8533\n",
      "Epoch 64/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.3176 - accuracy: 0.8658 - val_loss: 0.3249 - val_accuracy: 0.8507\n",
      "Epoch 65/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.3152 - accuracy: 0.8649 - val_loss: 0.3232 - val_accuracy: 0.8516\n",
      "Epoch 66/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.3130 - accuracy: 0.8631 - val_loss: 0.3215 - val_accuracy: 0.8524\n",
      "Epoch 67/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.3110 - accuracy: 0.8631 - val_loss: 0.3197 - val_accuracy: 0.8542\n",
      "Epoch 68/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.3091 - accuracy: 0.8631 - val_loss: 0.3186 - val_accuracy: 0.8524\n",
      "Epoch 69/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.3074 - accuracy: 0.8631 - val_loss: 0.3160 - val_accuracy: 0.8587\n",
      "Epoch 70/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.3055 - accuracy: 0.8631 - val_loss: 0.3139 - val_accuracy: 0.8596\n",
      "Epoch 71/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.3037 - accuracy: 0.8658 - val_loss: 0.3116 - val_accuracy: 0.8596\n",
      "Epoch 72/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.3019 - accuracy: 0.8684 - val_loss: 0.3095 - val_accuracy: 0.8613\n",
      "Epoch 73/700\n",
      "1125/1125 [==============================] - 0s 24us/step - loss: 0.3006 - accuracy: 0.8667 - val_loss: 0.3084 - val_accuracy: 0.8596\n",
      "Epoch 74/700\n",
      "1125/1125 [==============================] - 0s 22us/step - loss: 0.2987 - accuracy: 0.8684 - val_loss: 0.3069 - val_accuracy: 0.8596\n",
      "Epoch 75/700\n",
      "1125/1125 [==============================] - 0s 24us/step - loss: 0.2974 - accuracy: 0.8684 - val_loss: 0.3049 - val_accuracy: 0.8649\n",
      "Epoch 76/700\n",
      "1125/1125 [==============================] - 0s 27us/step - loss: 0.2958 - accuracy: 0.8693 - val_loss: 0.3032 - val_accuracy: 0.8649\n",
      "Epoch 77/700\n",
      "1125/1125 [==============================] - 0s 23us/step - loss: 0.2943 - accuracy: 0.8720 - val_loss: 0.3019 - val_accuracy: 0.8640\n",
      "Epoch 78/700\n",
      "1125/1125 [==============================] - 0s 26us/step - loss: 0.2926 - accuracy: 0.8711 - val_loss: 0.3012 - val_accuracy: 0.8604\n",
      "Epoch 79/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.2910 - accuracy: 0.8693 - val_loss: 0.2990 - val_accuracy: 0.8658\n",
      "Epoch 80/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2897 - accuracy: 0.8702 - val_loss: 0.2972 - val_accuracy: 0.8676\n",
      "Epoch 81/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2885 - accuracy: 0.8720 - val_loss: 0.2958 - val_accuracy: 0.8684\n",
      "Epoch 82/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.2870 - accuracy: 0.8702 - val_loss: 0.2952 - val_accuracy: 0.8667\n",
      "Epoch 83/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.2857 - accuracy: 0.8720 - val_loss: 0.2933 - val_accuracy: 0.8684\n",
      "Epoch 84/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2843 - accuracy: 0.8711 - val_loss: 0.2916 - val_accuracy: 0.8684\n",
      "Epoch 85/700\n",
      "1125/1125 [==============================] - ETA: 0s - loss: 0.2945 - accuracy: 0.86 - 0s 17us/step - loss: 0.2829 - accuracy: 0.8729 - val_loss: 0.2906 - val_accuracy: 0.8684\n",
      "Epoch 86/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.2816 - accuracy: 0.8729 - val_loss: 0.2901 - val_accuracy: 0.8684\n",
      "Epoch 87/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2805 - accuracy: 0.8747 - val_loss: 0.2888 - val_accuracy: 0.8693\n",
      "Epoch 88/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2790 - accuracy: 0.8756 - val_loss: 0.2876 - val_accuracy: 0.8702\n",
      "Epoch 89/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2780 - accuracy: 0.8756 - val_loss: 0.2880 - val_accuracy: 0.8676\n",
      "Epoch 90/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2774 - accuracy: 0.8756 - val_loss: 0.2852 - val_accuracy: 0.8702\n",
      "Epoch 91/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.2758 - accuracy: 0.8782 - val_loss: 0.2839 - val_accuracy: 0.8702\n",
      "Epoch 92/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2747 - accuracy: 0.8782 - val_loss: 0.2832 - val_accuracy: 0.8711\n",
      "Epoch 93/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.2738 - accuracy: 0.8782 - val_loss: 0.2816 - val_accuracy: 0.8711\n",
      "Epoch 94/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.2725 - accuracy: 0.8791 - val_loss: 0.2793 - val_accuracy: 0.8720\n",
      "Epoch 95/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2714 - accuracy: 0.8773 - val_loss: 0.2796 - val_accuracy: 0.8720\n",
      "Epoch 96/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.2705 - accuracy: 0.8773 - val_loss: 0.2794 - val_accuracy: 0.8756\n",
      "Epoch 97/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2698 - accuracy: 0.8800 - val_loss: 0.2777 - val_accuracy: 0.8747\n",
      "Epoch 98/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2685 - accuracy: 0.8800 - val_loss: 0.2756 - val_accuracy: 0.8738\n",
      "Epoch 99/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.2675 - accuracy: 0.8773 - val_loss: 0.2757 - val_accuracy: 0.8756\n",
      "Epoch 100/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.2668 - accuracy: 0.8791 - val_loss: 0.2742 - val_accuracy: 0.8773\n",
      "Epoch 101/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.2656 - accuracy: 0.8818 - val_loss: 0.2716 - val_accuracy: 0.8773\n",
      "Epoch 102/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.2649 - accuracy: 0.8782 - val_loss: 0.2704 - val_accuracy: 0.8782\n",
      "Epoch 103/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2637 - accuracy: 0.8791 - val_loss: 0.2689 - val_accuracy: 0.8773\n",
      "Epoch 104/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2628 - accuracy: 0.8809 - val_loss: 0.2688 - val_accuracy: 0.8800\n",
      "Epoch 105/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2618 - accuracy: 0.8827 - val_loss: 0.2676 - val_accuracy: 0.8818\n",
      "Epoch 106/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.2611 - accuracy: 0.8827 - val_loss: 0.2689 - val_accuracy: 0.8791\n",
      "Epoch 107/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2604 - accuracy: 0.8809 - val_loss: 0.2671 - val_accuracy: 0.8800\n",
      "Epoch 108/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2595 - accuracy: 0.8809 - val_loss: 0.2655 - val_accuracy: 0.8818\n",
      "Epoch 109/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.2587 - accuracy: 0.8827 - val_loss: 0.2640 - val_accuracy: 0.8809\n",
      "Epoch 110/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2580 - accuracy: 0.8862 - val_loss: 0.2649 - val_accuracy: 0.8818\n",
      "Epoch 111/700\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2579 - accuracy: 0.8827 - val_loss: 0.2643 - val_accuracy: 0.8827\n",
      "Epoch 112/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.2566 - accuracy: 0.8836 - val_loss: 0.2632 - val_accuracy: 0.8827\n",
      "Epoch 113/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.2561 - accuracy: 0.8836 - val_loss: 0.2621 - val_accuracy: 0.8836\n",
      "Epoch 114/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.2550 - accuracy: 0.8853 - val_loss: 0.2607 - val_accuracy: 0.8853\n",
      "Epoch 115/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2544 - accuracy: 0.8853 - val_loss: 0.2613 - val_accuracy: 0.8836\n",
      "Epoch 116/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.2539 - accuracy: 0.8844 - val_loss: 0.2585 - val_accuracy: 0.8827\n",
      "Epoch 117/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2533 - accuracy: 0.8880 - val_loss: 0.2584 - val_accuracy: 0.8844\n",
      "Epoch 118/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.2524 - accuracy: 0.8880 - val_loss: 0.2582 - val_accuracy: 0.8844\n",
      "Epoch 119/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2522 - accuracy: 0.8880 - val_loss: 0.2596 - val_accuracy: 0.8827\n",
      "Epoch 120/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2516 - accuracy: 0.8889 - val_loss: 0.2574 - val_accuracy: 0.8836\n",
      "Epoch 121/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.2506 - accuracy: 0.8871 - val_loss: 0.2561 - val_accuracy: 0.8844\n",
      "Epoch 122/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.2503 - accuracy: 0.8889 - val_loss: 0.2549 - val_accuracy: 0.8844\n",
      "Epoch 123/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.2493 - accuracy: 0.8907 - val_loss: 0.2553 - val_accuracy: 0.8836\n",
      "Epoch 124/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2491 - accuracy: 0.8898 - val_loss: 0.2553 - val_accuracy: 0.8844\n",
      "Epoch 125/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.2485 - accuracy: 0.8907 - val_loss: 0.2546 - val_accuracy: 0.8844\n",
      "Epoch 126/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2475 - accuracy: 0.8916 - val_loss: 0.2518 - val_accuracy: 0.8844\n",
      "Epoch 127/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2466 - accuracy: 0.8933 - val_loss: 0.2518 - val_accuracy: 0.8844\n",
      "Epoch 128/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.2461 - accuracy: 0.8924 - val_loss: 0.2507 - val_accuracy: 0.8853\n",
      "Epoch 129/700\n",
      "1125/1125 [==============================] - 0s 23us/step - loss: 0.2454 - accuracy: 0.8933 - val_loss: 0.2493 - val_accuracy: 0.8853\n",
      "Epoch 130/700\n",
      "1125/1125 [==============================] - 0s 24us/step - loss: 0.2447 - accuracy: 0.8942 - val_loss: 0.2511 - val_accuracy: 0.8844\n",
      "Epoch 131/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2442 - accuracy: 0.8916 - val_loss: 0.2490 - val_accuracy: 0.8844\n",
      "Epoch 132/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.2433 - accuracy: 0.8969 - val_loss: 0.2487 - val_accuracy: 0.8853\n",
      "Epoch 133/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2428 - accuracy: 0.8951 - val_loss: 0.2475 - val_accuracy: 0.8853\n",
      "Epoch 134/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2421 - accuracy: 0.8969 - val_loss: 0.2450 - val_accuracy: 0.8898\n",
      "Epoch 135/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.2413 - accuracy: 0.8960 - val_loss: 0.2442 - val_accuracy: 0.8916\n",
      "Epoch 136/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.2412 - accuracy: 0.8978 - val_loss: 0.2452 - val_accuracy: 0.8853\n",
      "Epoch 137/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2400 - accuracy: 0.8987 - val_loss: 0.2441 - val_accuracy: 0.8853\n",
      "Epoch 138/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2393 - accuracy: 0.8969 - val_loss: 0.2445 - val_accuracy: 0.8853\n",
      "Epoch 139/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2388 - accuracy: 0.8987 - val_loss: 0.2425 - val_accuracy: 0.8871\n",
      "Epoch 140/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.2379 - accuracy: 0.8969 - val_loss: 0.2418 - val_accuracy: 0.8880\n",
      "Epoch 141/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2374 - accuracy: 0.8978 - val_loss: 0.2415 - val_accuracy: 0.8862\n",
      "Epoch 142/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2366 - accuracy: 0.8978 - val_loss: 0.2397 - val_accuracy: 0.8889\n",
      "Epoch 143/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.2356 - accuracy: 0.8987 - val_loss: 0.2378 - val_accuracy: 0.8942\n",
      "Epoch 144/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.2349 - accuracy: 0.9004 - val_loss: 0.2366 - val_accuracy: 0.8933\n",
      "Epoch 145/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2343 - accuracy: 0.9022 - val_loss: 0.2384 - val_accuracy: 0.8898\n",
      "Epoch 146/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2336 - accuracy: 0.9022 - val_loss: 0.2384 - val_accuracy: 0.8898\n",
      "Epoch 147/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.2333 - accuracy: 0.9040 - val_loss: 0.2377 - val_accuracy: 0.8898\n",
      "Epoch 148/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.2325 - accuracy: 0.9031 - val_loss: 0.2366 - val_accuracy: 0.8898\n",
      "Epoch 149/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.2320 - accuracy: 0.9067 - val_loss: 0.2338 - val_accuracy: 0.8978\n",
      "Epoch 150/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.2308 - accuracy: 0.9040 - val_loss: 0.2335 - val_accuracy: 0.8951\n",
      "Epoch 151/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2301 - accuracy: 0.9049 - val_loss: 0.2322 - val_accuracy: 0.8969\n",
      "Epoch 152/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.2295 - accuracy: 0.9058 - val_loss: 0.2329 - val_accuracy: 0.8942\n",
      "Epoch 153/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2290 - accuracy: 0.9040 - val_loss: 0.2321 - val_accuracy: 0.8960\n",
      "Epoch 154/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.2282 - accuracy: 0.9049 - val_loss: 0.2290 - val_accuracy: 0.8978\n",
      "Epoch 155/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2277 - accuracy: 0.9067 - val_loss: 0.2277 - val_accuracy: 0.8996\n",
      "Epoch 156/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2272 - accuracy: 0.9067 - val_loss: 0.2278 - val_accuracy: 0.8987\n",
      "Epoch 157/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2262 - accuracy: 0.9067 - val_loss: 0.2275 - val_accuracy: 0.8996\n",
      "Epoch 158/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2255 - accuracy: 0.9102 - val_loss: 0.2278 - val_accuracy: 0.8978\n",
      "Epoch 159/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2250 - accuracy: 0.9084 - val_loss: 0.2269 - val_accuracy: 0.9004\n",
      "Epoch 160/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2243 - accuracy: 0.9093 - val_loss: 0.2269 - val_accuracy: 0.8996\n",
      "Epoch 161/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.2236 - accuracy: 0.9093 - val_loss: 0.2253 - val_accuracy: 0.8996\n",
      "Epoch 162/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2229 - accuracy: 0.9093 - val_loss: 0.2268 - val_accuracy: 0.8987\n",
      "Epoch 163/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2236 - accuracy: 0.9084 - val_loss: 0.2245 - val_accuracy: 0.9004\n",
      "Epoch 164/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.2217 - accuracy: 0.9093 - val_loss: 0.2239 - val_accuracy: 0.9013\n",
      "Epoch 165/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2211 - accuracy: 0.9102 - val_loss: 0.2214 - val_accuracy: 0.9013\n",
      "Epoch 166/700\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2204 - accuracy: 0.9102 - val_loss: 0.2203 - val_accuracy: 0.9013\n",
      "Epoch 167/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2198 - accuracy: 0.9102 - val_loss: 0.2203 - val_accuracy: 0.9022\n",
      "Epoch 168/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2193 - accuracy: 0.9120 - val_loss: 0.2193 - val_accuracy: 0.9031\n",
      "Epoch 169/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.2184 - accuracy: 0.9120 - val_loss: 0.2179 - val_accuracy: 0.9022\n",
      "Epoch 170/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2179 - accuracy: 0.9120 - val_loss: 0.2181 - val_accuracy: 0.9031\n",
      "Epoch 171/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2169 - accuracy: 0.9120 - val_loss: 0.2169 - val_accuracy: 0.9040\n",
      "Epoch 172/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.2169 - accuracy: 0.9120 - val_loss: 0.2178 - val_accuracy: 0.9031\n",
      "Epoch 173/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2156 - accuracy: 0.9129 - val_loss: 0.2182 - val_accuracy: 0.9031\n",
      "Epoch 174/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2151 - accuracy: 0.9129 - val_loss: 0.2171 - val_accuracy: 0.9031\n",
      "Epoch 175/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2143 - accuracy: 0.9129 - val_loss: 0.2145 - val_accuracy: 0.9040\n",
      "Epoch 176/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2134 - accuracy: 0.9129 - val_loss: 0.2145 - val_accuracy: 0.9040\n",
      "Epoch 177/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2128 - accuracy: 0.9147 - val_loss: 0.2142 - val_accuracy: 0.9049\n",
      "Epoch 178/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.2120 - accuracy: 0.9138 - val_loss: 0.2137 - val_accuracy: 0.9049\n",
      "Epoch 179/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2115 - accuracy: 0.9173 - val_loss: 0.2139 - val_accuracy: 0.9058\n",
      "Epoch 180/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2111 - accuracy: 0.9173 - val_loss: 0.2115 - val_accuracy: 0.9058\n",
      "Epoch 181/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2102 - accuracy: 0.9164 - val_loss: 0.2087 - val_accuracy: 0.9076\n",
      "Epoch 182/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2091 - accuracy: 0.9164 - val_loss: 0.2087 - val_accuracy: 0.9067\n",
      "Epoch 183/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2081 - accuracy: 0.9173 - val_loss: 0.2078 - val_accuracy: 0.9067\n",
      "Epoch 184/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2075 - accuracy: 0.9182 - val_loss: 0.2062 - val_accuracy: 0.9076\n",
      "Epoch 185/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.2069 - accuracy: 0.9173 - val_loss: 0.2063 - val_accuracy: 0.9084\n",
      "Epoch 186/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2060 - accuracy: 0.9191 - val_loss: 0.2045 - val_accuracy: 0.9084\n",
      "Epoch 187/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.2055 - accuracy: 0.9191 - val_loss: 0.2044 - val_accuracy: 0.9084\n",
      "Epoch 188/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.2045 - accuracy: 0.9182 - val_loss: 0.2032 - val_accuracy: 0.9102\n",
      "Epoch 189/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.2036 - accuracy: 0.9191 - val_loss: 0.2029 - val_accuracy: 0.9102\n",
      "Epoch 190/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2031 - accuracy: 0.9218 - val_loss: 0.2022 - val_accuracy: 0.9102\n",
      "Epoch 191/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2025 - accuracy: 0.9200 - val_loss: 0.2022 - val_accuracy: 0.9102\n",
      "Epoch 192/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.2015 - accuracy: 0.9200 - val_loss: 0.2031 - val_accuracy: 0.9129\n",
      "Epoch 193/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.2010 - accuracy: 0.9209 - val_loss: 0.2008 - val_accuracy: 0.9120\n",
      "Epoch 194/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1999 - accuracy: 0.9218 - val_loss: 0.2002 - val_accuracy: 0.9129\n",
      "Epoch 195/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1991 - accuracy: 0.9218 - val_loss: 0.1972 - val_accuracy: 0.9138\n",
      "Epoch 196/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1981 - accuracy: 0.9200 - val_loss: 0.1954 - val_accuracy: 0.9173\n",
      "Epoch 197/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1979 - accuracy: 0.9209 - val_loss: 0.1948 - val_accuracy: 0.9182\n",
      "Epoch 198/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1967 - accuracy: 0.9218 - val_loss: 0.1943 - val_accuracy: 0.9164\n",
      "Epoch 199/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1960 - accuracy: 0.9209 - val_loss: 0.1955 - val_accuracy: 0.9164\n",
      "Epoch 200/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1955 - accuracy: 0.9236 - val_loss: 0.1943 - val_accuracy: 0.9173\n",
      "Epoch 201/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1947 - accuracy: 0.9218 - val_loss: 0.1931 - val_accuracy: 0.9191\n",
      "Epoch 202/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1938 - accuracy: 0.9227 - val_loss: 0.1914 - val_accuracy: 0.9209\n",
      "Epoch 203/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1931 - accuracy: 0.9209 - val_loss: 0.1907 - val_accuracy: 0.9191\n",
      "Epoch 204/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1927 - accuracy: 0.9209 - val_loss: 0.1924 - val_accuracy: 0.9200\n",
      "Epoch 205/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1920 - accuracy: 0.9244 - val_loss: 0.1916 - val_accuracy: 0.9191\n",
      "Epoch 206/700\n",
      "1125/1125 [==============================] - 0s 35us/step - loss: 0.1909 - accuracy: 0.9253 - val_loss: 0.1884 - val_accuracy: 0.9209\n",
      "Epoch 207/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1899 - accuracy: 0.9209 - val_loss: 0.1890 - val_accuracy: 0.9227\n",
      "Epoch 208/700\n",
      "1125/1125 [==============================] - 0s 23us/step - loss: 0.1894 - accuracy: 0.9244 - val_loss: 0.1872 - val_accuracy: 0.9218\n",
      "Epoch 209/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.1889 - accuracy: 0.9227 - val_loss: 0.1892 - val_accuracy: 0.9236\n",
      "Epoch 210/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1883 - accuracy: 0.9307 - val_loss: 0.1864 - val_accuracy: 0.9218\n",
      "Epoch 211/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1871 - accuracy: 0.9253 - val_loss: 0.1846 - val_accuracy: 0.9244\n",
      "Epoch 212/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1863 - accuracy: 0.9253 - val_loss: 0.1842 - val_accuracy: 0.9253\n",
      "Epoch 213/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1857 - accuracy: 0.9244 - val_loss: 0.1843 - val_accuracy: 0.9236\n",
      "Epoch 214/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1848 - accuracy: 0.9280 - val_loss: 0.1817 - val_accuracy: 0.9253\n",
      "Epoch 215/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1841 - accuracy: 0.9236 - val_loss: 0.1825 - val_accuracy: 0.9271\n",
      "Epoch 216/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1833 - accuracy: 0.9289 - val_loss: 0.1819 - val_accuracy: 0.9271\n",
      "Epoch 217/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1827 - accuracy: 0.9298 - val_loss: 0.1806 - val_accuracy: 0.9289\n",
      "Epoch 218/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1817 - accuracy: 0.9280 - val_loss: 0.1814 - val_accuracy: 0.9280\n",
      "Epoch 219/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1812 - accuracy: 0.9289 - val_loss: 0.1783 - val_accuracy: 0.9280\n",
      "Epoch 220/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1806 - accuracy: 0.9289 - val_loss: 0.1782 - val_accuracy: 0.9316\n",
      "Epoch 221/700\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1796 - accuracy: 0.9307 - val_loss: 0.1792 - val_accuracy: 0.9289\n",
      "Epoch 222/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1793 - accuracy: 0.9307 - val_loss: 0.1781 - val_accuracy: 0.9316\n",
      "Epoch 223/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1787 - accuracy: 0.9316 - val_loss: 0.1751 - val_accuracy: 0.9298\n",
      "Epoch 224/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1774 - accuracy: 0.9324 - val_loss: 0.1762 - val_accuracy: 0.9333\n",
      "Epoch 225/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1770 - accuracy: 0.9316 - val_loss: 0.1745 - val_accuracy: 0.9307\n",
      "Epoch 226/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1763 - accuracy: 0.9316 - val_loss: 0.1757 - val_accuracy: 0.9333\n",
      "Epoch 227/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1761 - accuracy: 0.9324 - val_loss: 0.1744 - val_accuracy: 0.9342\n",
      "Epoch 228/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1751 - accuracy: 0.9342 - val_loss: 0.1720 - val_accuracy: 0.9307\n",
      "Epoch 229/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1741 - accuracy: 0.9324 - val_loss: 0.1710 - val_accuracy: 0.9324\n",
      "Epoch 230/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1733 - accuracy: 0.9324 - val_loss: 0.1703 - val_accuracy: 0.9324\n",
      "Epoch 231/700\n",
      "1125/1125 [==============================] - 0s 23us/step - loss: 0.1726 - accuracy: 0.9324 - val_loss: 0.1691 - val_accuracy: 0.9324\n",
      "Epoch 232/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1721 - accuracy: 0.9316 - val_loss: 0.1674 - val_accuracy: 0.9342\n",
      "Epoch 233/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.1714 - accuracy: 0.9342 - val_loss: 0.1670 - val_accuracy: 0.9342\n",
      "Epoch 234/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1704 - accuracy: 0.9342 - val_loss: 0.1677 - val_accuracy: 0.9333\n",
      "Epoch 235/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1699 - accuracy: 0.9342 - val_loss: 0.1660 - val_accuracy: 0.9351\n",
      "Epoch 236/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.1689 - accuracy: 0.9333 - val_loss: 0.1660 - val_accuracy: 0.9351\n",
      "Epoch 237/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1696 - accuracy: 0.9351 - val_loss: 0.1660 - val_accuracy: 0.9369\n",
      "Epoch 238/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1677 - accuracy: 0.9360 - val_loss: 0.1638 - val_accuracy: 0.9351\n",
      "Epoch 239/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1667 - accuracy: 0.9360 - val_loss: 0.1640 - val_accuracy: 0.9360\n",
      "Epoch 240/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.1661 - accuracy: 0.9360 - val_loss: 0.1635 - val_accuracy: 0.9378\n",
      "Epoch 241/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1657 - accuracy: 0.9342 - val_loss: 0.1648 - val_accuracy: 0.9378\n",
      "Epoch 242/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1655 - accuracy: 0.9360 - val_loss: 0.1634 - val_accuracy: 0.9378\n",
      "Epoch 243/700\n",
      "1125/1125 [==============================] - 0s 22us/step - loss: 0.1649 - accuracy: 0.9369 - val_loss: 0.1625 - val_accuracy: 0.9378\n",
      "Epoch 244/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.1638 - accuracy: 0.9369 - val_loss: 0.1602 - val_accuracy: 0.9378\n",
      "Epoch 245/700\n",
      "1125/1125 [==============================] - 0s 25us/step - loss: 0.1627 - accuracy: 0.9378 - val_loss: 0.1593 - val_accuracy: 0.9378\n",
      "Epoch 246/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1620 - accuracy: 0.9378 - val_loss: 0.1586 - val_accuracy: 0.9369\n",
      "Epoch 247/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1613 - accuracy: 0.9387 - val_loss: 0.1576 - val_accuracy: 0.9396\n",
      "Epoch 248/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1605 - accuracy: 0.9378 - val_loss: 0.1557 - val_accuracy: 0.9378\n",
      "Epoch 249/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1599 - accuracy: 0.9378 - val_loss: 0.1554 - val_accuracy: 0.9378\n",
      "Epoch 250/700\n",
      "1125/1125 [==============================] - 0s 22us/step - loss: 0.1590 - accuracy: 0.9369 - val_loss: 0.1550 - val_accuracy: 0.9404\n",
      "Epoch 251/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1583 - accuracy: 0.9378 - val_loss: 0.1540 - val_accuracy: 0.9413\n",
      "Epoch 252/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1583 - accuracy: 0.9369 - val_loss: 0.1559 - val_accuracy: 0.9422\n",
      "Epoch 253/700\n",
      "1125/1125 [==============================] - 0s 23us/step - loss: 0.1575 - accuracy: 0.9378 - val_loss: 0.1533 - val_accuracy: 0.9440\n",
      "Epoch 254/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1568 - accuracy: 0.9413 - val_loss: 0.1528 - val_accuracy: 0.9440\n",
      "Epoch 255/700\n",
      "1125/1125 [==============================] - 0s 25us/step - loss: 0.1555 - accuracy: 0.9404 - val_loss: 0.1511 - val_accuracy: 0.9413\n",
      "Epoch 256/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1547 - accuracy: 0.9378 - val_loss: 0.1509 - val_accuracy: 0.9440\n",
      "Epoch 257/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1540 - accuracy: 0.9404 - val_loss: 0.1493 - val_accuracy: 0.9440\n",
      "Epoch 258/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1535 - accuracy: 0.9404 - val_loss: 0.1496 - val_accuracy: 0.9440\n",
      "Epoch 259/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1527 - accuracy: 0.9413 - val_loss: 0.1488 - val_accuracy: 0.9440\n",
      "Epoch 260/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1518 - accuracy: 0.9413 - val_loss: 0.1481 - val_accuracy: 0.9449\n",
      "Epoch 261/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1512 - accuracy: 0.9422 - val_loss: 0.1462 - val_accuracy: 0.9458\n",
      "Epoch 262/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1507 - accuracy: 0.9404 - val_loss: 0.1462 - val_accuracy: 0.9458\n",
      "Epoch 263/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.1498 - accuracy: 0.9404 - val_loss: 0.1463 - val_accuracy: 0.9449\n",
      "Epoch 264/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1495 - accuracy: 0.9431 - val_loss: 0.1457 - val_accuracy: 0.9449\n",
      "Epoch 265/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.1493 - accuracy: 0.9422 - val_loss: 0.1454 - val_accuracy: 0.9449\n",
      "Epoch 266/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1482 - accuracy: 0.9422 - val_loss: 0.1447 - val_accuracy: 0.9449\n",
      "Epoch 267/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1475 - accuracy: 0.9440 - val_loss: 0.1427 - val_accuracy: 0.9476\n",
      "Epoch 268/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1469 - accuracy: 0.9413 - val_loss: 0.1433 - val_accuracy: 0.9449\n",
      "Epoch 269/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1464 - accuracy: 0.9422 - val_loss: 0.1423 - val_accuracy: 0.9458\n",
      "Epoch 270/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1456 - accuracy: 0.9422 - val_loss: 0.1410 - val_accuracy: 0.9484\n",
      "Epoch 271/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1447 - accuracy: 0.9431 - val_loss: 0.1398 - val_accuracy: 0.9511\n",
      "Epoch 272/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1445 - accuracy: 0.9431 - val_loss: 0.1388 - val_accuracy: 0.9529\n",
      "Epoch 273/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1434 - accuracy: 0.9440 - val_loss: 0.1386 - val_accuracy: 0.9502\n",
      "Epoch 274/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1431 - accuracy: 0.9440 - val_loss: 0.1385 - val_accuracy: 0.9484\n",
      "Epoch 275/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1423 - accuracy: 0.9476 - val_loss: 0.1367 - val_accuracy: 0.9538\n",
      "Epoch 276/700\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1416 - accuracy: 0.9440 - val_loss: 0.1360 - val_accuracy: 0.9556\n",
      "Epoch 277/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1412 - accuracy: 0.9467 - val_loss: 0.1355 - val_accuracy: 0.9538\n",
      "Epoch 278/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1407 - accuracy: 0.9476 - val_loss: 0.1356 - val_accuracy: 0.9547\n",
      "Epoch 279/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1398 - accuracy: 0.9467 - val_loss: 0.1352 - val_accuracy: 0.9529\n",
      "Epoch 280/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1392 - accuracy: 0.9484 - val_loss: 0.1339 - val_accuracy: 0.9547\n",
      "Epoch 281/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1389 - accuracy: 0.9467 - val_loss: 0.1332 - val_accuracy: 0.9556\n",
      "Epoch 282/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1379 - accuracy: 0.9467 - val_loss: 0.1334 - val_accuracy: 0.9538\n",
      "Epoch 283/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1376 - accuracy: 0.9502 - val_loss: 0.1341 - val_accuracy: 0.9502\n",
      "Epoch 284/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.1373 - accuracy: 0.9493 - val_loss: 0.1338 - val_accuracy: 0.9511\n",
      "Epoch 285/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1373 - accuracy: 0.9493 - val_loss: 0.1321 - val_accuracy: 0.9538\n",
      "Epoch 286/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1361 - accuracy: 0.9484 - val_loss: 0.1326 - val_accuracy: 0.9511\n",
      "Epoch 287/700\n",
      "1125/1125 [==============================] - 0s 22us/step - loss: 0.1359 - accuracy: 0.9520 - val_loss: 0.1307 - val_accuracy: 0.9564\n",
      "Epoch 288/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.1350 - accuracy: 0.9493 - val_loss: 0.1307 - val_accuracy: 0.9538\n",
      "Epoch 289/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1345 - accuracy: 0.9529 - val_loss: 0.1284 - val_accuracy: 0.9573\n",
      "Epoch 290/700\n",
      "1125/1125 [==============================] - 0s 27us/step - loss: 0.1338 - accuracy: 0.9511 - val_loss: 0.1281 - val_accuracy: 0.9556\n",
      "Epoch 291/700\n",
      "1125/1125 [==============================] - 0s 34us/step - loss: 0.1334 - accuracy: 0.9547 - val_loss: 0.1273 - val_accuracy: 0.9573\n",
      "Epoch 292/700\n",
      "1125/1125 [==============================] - 0s 22us/step - loss: 0.1331 - accuracy: 0.9493 - val_loss: 0.1266 - val_accuracy: 0.9573\n",
      "Epoch 293/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.1325 - accuracy: 0.9511 - val_loss: 0.1275 - val_accuracy: 0.9547\n",
      "Epoch 294/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1323 - accuracy: 0.9520 - val_loss: 0.1277 - val_accuracy: 0.9538\n",
      "Epoch 295/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1319 - accuracy: 0.9520 - val_loss: 0.1255 - val_accuracy: 0.9582\n",
      "Epoch 296/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1311 - accuracy: 0.9547 - val_loss: 0.1254 - val_accuracy: 0.9573\n",
      "Epoch 297/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1306 - accuracy: 0.9547 - val_loss: 0.1246 - val_accuracy: 0.9582\n",
      "Epoch 298/700\n",
      "1125/1125 [==============================] - 0s 22us/step - loss: 0.1305 - accuracy: 0.9529 - val_loss: 0.1245 - val_accuracy: 0.9573\n",
      "Epoch 299/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1297 - accuracy: 0.9529 - val_loss: 0.1248 - val_accuracy: 0.9564\n",
      "Epoch 300/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1294 - accuracy: 0.9573 - val_loss: 0.1252 - val_accuracy: 0.9547\n",
      "Epoch 301/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1292 - accuracy: 0.9538 - val_loss: 0.1233 - val_accuracy: 0.9582\n",
      "Epoch 302/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1283 - accuracy: 0.9564 - val_loss: 0.1224 - val_accuracy: 0.9582\n",
      "Epoch 303/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1279 - accuracy: 0.9538 - val_loss: 0.1218 - val_accuracy: 0.9591\n",
      "Epoch 304/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1274 - accuracy: 0.9556 - val_loss: 0.1219 - val_accuracy: 0.9573\n",
      "Epoch 305/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1268 - accuracy: 0.9547 - val_loss: 0.1213 - val_accuracy: 0.9573\n",
      "Epoch 306/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1263 - accuracy: 0.9556 - val_loss: 0.1210 - val_accuracy: 0.9573\n",
      "Epoch 307/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1258 - accuracy: 0.9547 - val_loss: 0.1212 - val_accuracy: 0.9564\n",
      "Epoch 308/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1269 - accuracy: 0.9556 - val_loss: 0.1216 - val_accuracy: 0.9573\n",
      "Epoch 309/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.1254 - accuracy: 0.9573 - val_loss: 0.1204 - val_accuracy: 0.9564\n",
      "Epoch 310/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1246 - accuracy: 0.9573 - val_loss: 0.1199 - val_accuracy: 0.9573\n",
      "Epoch 311/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1241 - accuracy: 0.9538 - val_loss: 0.1189 - val_accuracy: 0.9564\n",
      "Epoch 312/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1238 - accuracy: 0.9573 - val_loss: 0.1179 - val_accuracy: 0.9609\n",
      "Epoch 313/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1235 - accuracy: 0.9556 - val_loss: 0.1174 - val_accuracy: 0.9591\n",
      "Epoch 314/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1228 - accuracy: 0.9582 - val_loss: 0.1171 - val_accuracy: 0.9582\n",
      "Epoch 315/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1225 - accuracy: 0.9600 - val_loss: 0.1172 - val_accuracy: 0.9573\n",
      "Epoch 316/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1221 - accuracy: 0.9564 - val_loss: 0.1165 - val_accuracy: 0.9582\n",
      "Epoch 317/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1213 - accuracy: 0.9582 - val_loss: 0.1173 - val_accuracy: 0.9582\n",
      "Epoch 318/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1211 - accuracy: 0.9591 - val_loss: 0.1158 - val_accuracy: 0.9618\n",
      "Epoch 319/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1211 - accuracy: 0.9591 - val_loss: 0.1150 - val_accuracy: 0.9618\n",
      "Epoch 320/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1201 - accuracy: 0.9600 - val_loss: 0.1144 - val_accuracy: 0.9591\n",
      "Epoch 321/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.1199 - accuracy: 0.9609 - val_loss: 0.1142 - val_accuracy: 0.9618\n",
      "Epoch 322/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1193 - accuracy: 0.9591 - val_loss: 0.1134 - val_accuracy: 0.9618\n",
      "Epoch 323/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1190 - accuracy: 0.9600 - val_loss: 0.1134 - val_accuracy: 0.9618\n",
      "Epoch 324/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1184 - accuracy: 0.9600 - val_loss: 0.1135 - val_accuracy: 0.9609\n",
      "Epoch 325/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1182 - accuracy: 0.9582 - val_loss: 0.1128 - val_accuracy: 0.9618\n",
      "Epoch 326/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1179 - accuracy: 0.9662 - val_loss: 0.1126 - val_accuracy: 0.9618\n",
      "Epoch 327/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1174 - accuracy: 0.9627 - val_loss: 0.1122 - val_accuracy: 0.9627\n",
      "Epoch 328/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.1169 - accuracy: 0.9636 - val_loss: 0.1116 - val_accuracy: 0.9609\n",
      "Epoch 329/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.1164 - accuracy: 0.9618 - val_loss: 0.1113 - val_accuracy: 0.9618\n",
      "Epoch 330/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1160 - accuracy: 0.9644 - val_loss: 0.1112 - val_accuracy: 0.9618\n",
      "Epoch 331/700\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1160 - accuracy: 0.9600 - val_loss: 0.1112 - val_accuracy: 0.9618\n",
      "Epoch 332/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1162 - accuracy: 0.9600 - val_loss: 0.1112 - val_accuracy: 0.9627\n",
      "Epoch 333/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1153 - accuracy: 0.9627 - val_loss: 0.1114 - val_accuracy: 0.9618\n",
      "Epoch 334/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1150 - accuracy: 0.9618 - val_loss: 0.1118 - val_accuracy: 0.9618\n",
      "Epoch 335/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1149 - accuracy: 0.9618 - val_loss: 0.1102 - val_accuracy: 0.9618\n",
      "Epoch 336/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1143 - accuracy: 0.9636 - val_loss: 0.1093 - val_accuracy: 0.9636\n",
      "Epoch 337/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1138 - accuracy: 0.9662 - val_loss: 0.1088 - val_accuracy: 0.9609\n",
      "Epoch 338/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1134 - accuracy: 0.9662 - val_loss: 0.1079 - val_accuracy: 0.9609\n",
      "Epoch 339/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1133 - accuracy: 0.9671 - val_loss: 0.1085 - val_accuracy: 0.9591\n",
      "Epoch 340/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1132 - accuracy: 0.9653 - val_loss: 0.1079 - val_accuracy: 0.9618\n",
      "Epoch 341/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1126 - accuracy: 0.9671 - val_loss: 0.1086 - val_accuracy: 0.9591\n",
      "Epoch 342/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1128 - accuracy: 0.9636 - val_loss: 0.1081 - val_accuracy: 0.9627\n",
      "Epoch 343/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.1121 - accuracy: 0.9653 - val_loss: 0.1071 - val_accuracy: 0.9627\n",
      "Epoch 344/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1124 - accuracy: 0.9671 - val_loss: 0.1068 - val_accuracy: 0.9627\n",
      "Epoch 345/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1116 - accuracy: 0.9653 - val_loss: 0.1067 - val_accuracy: 0.9627\n",
      "Epoch 346/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1113 - accuracy: 0.9662 - val_loss: 0.1074 - val_accuracy: 0.9636\n",
      "Epoch 347/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1114 - accuracy: 0.9644 - val_loss: 0.1068 - val_accuracy: 0.9636\n",
      "Epoch 348/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1108 - accuracy: 0.9662 - val_loss: 0.1066 - val_accuracy: 0.9627\n",
      "Epoch 349/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1105 - accuracy: 0.9680 - val_loss: 0.1065 - val_accuracy: 0.9609\n",
      "Epoch 350/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1106 - accuracy: 0.9671 - val_loss: 0.1053 - val_accuracy: 0.9600\n",
      "Epoch 351/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1100 - accuracy: 0.9671 - val_loss: 0.1050 - val_accuracy: 0.9636\n",
      "Epoch 352/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1097 - accuracy: 0.9662 - val_loss: 0.1049 - val_accuracy: 0.9627\n",
      "Epoch 353/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1094 - accuracy: 0.9680 - val_loss: 0.1043 - val_accuracy: 0.9609\n",
      "Epoch 354/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1090 - accuracy: 0.9689 - val_loss: 0.1050 - val_accuracy: 0.9618\n",
      "Epoch 355/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1092 - accuracy: 0.9680 - val_loss: 0.1051 - val_accuracy: 0.9618\n",
      "Epoch 356/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1088 - accuracy: 0.9680 - val_loss: 0.1045 - val_accuracy: 0.9636\n",
      "Epoch 357/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1086 - accuracy: 0.9662 - val_loss: 0.1042 - val_accuracy: 0.9609\n",
      "Epoch 358/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1087 - accuracy: 0.9680 - val_loss: 0.1037 - val_accuracy: 0.9600\n",
      "Epoch 359/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1081 - accuracy: 0.9689 - val_loss: 0.1034 - val_accuracy: 0.9618\n",
      "Epoch 360/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.1079 - accuracy: 0.9689 - val_loss: 0.1034 - val_accuracy: 0.9636\n",
      "Epoch 361/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1077 - accuracy: 0.9680 - val_loss: 0.1035 - val_accuracy: 0.9644\n",
      "Epoch 362/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1076 - accuracy: 0.9636 - val_loss: 0.1026 - val_accuracy: 0.9618\n",
      "Epoch 363/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1070 - accuracy: 0.9689 - val_loss: 0.1026 - val_accuracy: 0.9618\n",
      "Epoch 364/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1067 - accuracy: 0.9680 - val_loss: 0.1019 - val_accuracy: 0.9618\n",
      "Epoch 365/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1065 - accuracy: 0.9680 - val_loss: 0.1024 - val_accuracy: 0.9609\n",
      "Epoch 366/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1068 - accuracy: 0.9671 - val_loss: 0.1018 - val_accuracy: 0.9618\n",
      "Epoch 367/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1062 - accuracy: 0.9671 - val_loss: 0.1020 - val_accuracy: 0.9609\n",
      "Epoch 368/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1063 - accuracy: 0.9671 - val_loss: 0.1013 - val_accuracy: 0.9618\n",
      "Epoch 369/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1057 - accuracy: 0.9689 - val_loss: 0.1012 - val_accuracy: 0.9627\n",
      "Epoch 370/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1055 - accuracy: 0.9680 - val_loss: 0.1011 - val_accuracy: 0.9627\n",
      "Epoch 371/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1056 - accuracy: 0.9671 - val_loss: 0.1006 - val_accuracy: 0.9618\n",
      "Epoch 372/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1049 - accuracy: 0.9680 - val_loss: 0.1009 - val_accuracy: 0.9627\n",
      "Epoch 373/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.1047 - accuracy: 0.9689 - val_loss: 0.1003 - val_accuracy: 0.9627\n",
      "Epoch 374/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.1045 - accuracy: 0.9671 - val_loss: 0.1006 - val_accuracy: 0.9636\n",
      "Epoch 375/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.1049 - accuracy: 0.9671 - val_loss: 0.1010 - val_accuracy: 0.9653\n",
      "Epoch 376/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1049 - accuracy: 0.9662 - val_loss: 0.1018 - val_accuracy: 0.9644\n",
      "Epoch 377/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.1048 - accuracy: 0.9653 - val_loss: 0.1008 - val_accuracy: 0.9653\n",
      "Epoch 378/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.1042 - accuracy: 0.9653 - val_loss: 0.1012 - val_accuracy: 0.9618\n",
      "Epoch 379/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1040 - accuracy: 0.9689 - val_loss: 0.1001 - val_accuracy: 0.9627\n",
      "Epoch 380/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.1034 - accuracy: 0.9671 - val_loss: 0.1001 - val_accuracy: 0.9636\n",
      "Epoch 381/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.1035 - accuracy: 0.9689 - val_loss: 0.0999 - val_accuracy: 0.9627\n",
      "Epoch 382/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.1034 - accuracy: 0.9671 - val_loss: 0.0990 - val_accuracy: 0.9636\n",
      "Epoch 383/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.1030 - accuracy: 0.9680 - val_loss: 0.0994 - val_accuracy: 0.9627\n",
      "Epoch 384/700\n",
      "1125/1125 [==============================] - 0s 13us/step - loss: 0.1027 - accuracy: 0.9671 - val_loss: 0.0999 - val_accuracy: 0.9627\n",
      "Epoch 385/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.1030 - accuracy: 0.9680 - val_loss: 0.0995 - val_accuracy: 0.9627\n",
      "Epoch 386/700\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1026 - accuracy: 0.9680 - val_loss: 0.0999 - val_accuracy: 0.9627\n",
      "Epoch 387/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.1033 - accuracy: 0.9680 - val_loss: 0.0990 - val_accuracy: 0.9636\n",
      "Epoch 388/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.1022 - accuracy: 0.9671 - val_loss: 0.0992 - val_accuracy: 0.9636\n",
      "Epoch 389/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1022 - accuracy: 0.9689 - val_loss: 0.0992 - val_accuracy: 0.9627\n",
      "Epoch 390/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.1021 - accuracy: 0.9680 - val_loss: 0.0993 - val_accuracy: 0.9618\n",
      "Epoch 391/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1020 - accuracy: 0.9671 - val_loss: 0.0985 - val_accuracy: 0.9636\n",
      "Epoch 392/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.1017 - accuracy: 0.9671 - val_loss: 0.0981 - val_accuracy: 0.9609\n",
      "Epoch 393/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1018 - accuracy: 0.9671 - val_loss: 0.0986 - val_accuracy: 0.9618\n",
      "Epoch 394/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1014 - accuracy: 0.9671 - val_loss: 0.0991 - val_accuracy: 0.9653\n",
      "Epoch 395/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.1023 - accuracy: 0.9653 - val_loss: 0.0986 - val_accuracy: 0.9636\n",
      "Epoch 396/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.1011 - accuracy: 0.9671 - val_loss: 0.0989 - val_accuracy: 0.9644\n",
      "Epoch 397/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.1012 - accuracy: 0.9662 - val_loss: 0.0990 - val_accuracy: 0.9627\n",
      "Epoch 398/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.1012 - accuracy: 0.9662 - val_loss: 0.0980 - val_accuracy: 0.9644\n",
      "Epoch 399/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1011 - accuracy: 0.9671 - val_loss: 0.0975 - val_accuracy: 0.9636\n",
      "Epoch 400/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1005 - accuracy: 0.9671 - val_loss: 0.0976 - val_accuracy: 0.9618\n",
      "Epoch 401/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.1005 - accuracy: 0.9680 - val_loss: 0.0977 - val_accuracy: 0.9636\n",
      "Epoch 402/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1005 - accuracy: 0.9671 - val_loss: 0.0973 - val_accuracy: 0.9636\n",
      "Epoch 403/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1009 - accuracy: 0.9671 - val_loss: 0.0972 - val_accuracy: 0.9618\n",
      "Epoch 404/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.1002 - accuracy: 0.9680 - val_loss: 0.0979 - val_accuracy: 0.9627\n",
      "Epoch 405/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.1005 - accuracy: 0.9671 - val_loss: 0.0973 - val_accuracy: 0.9636\n",
      "Epoch 406/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.1003 - accuracy: 0.9671 - val_loss: 0.0964 - val_accuracy: 0.9627\n",
      "Epoch 407/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0999 - accuracy: 0.9671 - val_loss: 0.0963 - val_accuracy: 0.9636\n",
      "Epoch 408/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0999 - accuracy: 0.9680 - val_loss: 0.0969 - val_accuracy: 0.9627\n",
      "Epoch 409/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.0997 - accuracy: 0.9680 - val_loss: 0.0968 - val_accuracy: 0.9618\n",
      "Epoch 410/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0996 - accuracy: 0.9680 - val_loss: 0.0964 - val_accuracy: 0.9627\n",
      "Epoch 411/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0991 - accuracy: 0.9671 - val_loss: 0.0967 - val_accuracy: 0.9627\n",
      "Epoch 412/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0993 - accuracy: 0.9680 - val_loss: 0.0966 - val_accuracy: 0.9627\n",
      "Epoch 413/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0989 - accuracy: 0.9680 - val_loss: 0.0966 - val_accuracy: 0.9636\n",
      "Epoch 414/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0992 - accuracy: 0.9671 - val_loss: 0.0970 - val_accuracy: 0.9636\n",
      "Epoch 415/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0989 - accuracy: 0.9671 - val_loss: 0.0964 - val_accuracy: 0.9627\n",
      "Epoch 416/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0991 - accuracy: 0.9680 - val_loss: 0.0954 - val_accuracy: 0.9627\n",
      "Epoch 417/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0989 - accuracy: 0.9680 - val_loss: 0.0955 - val_accuracy: 0.9636\n",
      "Epoch 418/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0985 - accuracy: 0.9662 - val_loss: 0.0963 - val_accuracy: 0.9627\n",
      "Epoch 419/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0987 - accuracy: 0.9680 - val_loss: 0.0953 - val_accuracy: 0.9627\n",
      "Epoch 420/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.0982 - accuracy: 0.9671 - val_loss: 0.0948 - val_accuracy: 0.9627\n",
      "Epoch 421/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.0981 - accuracy: 0.9671 - val_loss: 0.0952 - val_accuracy: 0.9627\n",
      "Epoch 422/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.0984 - accuracy: 0.9671 - val_loss: 0.0950 - val_accuracy: 0.9644\n",
      "Epoch 423/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0983 - accuracy: 0.9671 - val_loss: 0.0952 - val_accuracy: 0.9636\n",
      "Epoch 424/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0980 - accuracy: 0.9671 - val_loss: 0.0955 - val_accuracy: 0.9644\n",
      "Epoch 425/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0977 - accuracy: 0.9671 - val_loss: 0.0956 - val_accuracy: 0.9627\n",
      "Epoch 426/700\n",
      "1125/1125 [==============================] - ETA: 0s - loss: 0.1070 - accuracy: 0.96 - 0s 19us/step - loss: 0.0985 - accuracy: 0.9680 - val_loss: 0.0958 - val_accuracy: 0.9627\n",
      "Epoch 427/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0981 - accuracy: 0.9680 - val_loss: 0.0956 - val_accuracy: 0.9636\n",
      "Epoch 428/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0977 - accuracy: 0.9671 - val_loss: 0.0947 - val_accuracy: 0.9636\n",
      "Epoch 429/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0973 - accuracy: 0.9662 - val_loss: 0.0953 - val_accuracy: 0.9644\n",
      "Epoch 430/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0975 - accuracy: 0.9680 - val_loss: 0.0954 - val_accuracy: 0.9644\n",
      "Epoch 431/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0974 - accuracy: 0.9671 - val_loss: 0.0955 - val_accuracy: 0.9644\n",
      "Epoch 432/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0971 - accuracy: 0.9671 - val_loss: 0.0953 - val_accuracy: 0.9627\n",
      "Epoch 433/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0970 - accuracy: 0.9689 - val_loss: 0.0944 - val_accuracy: 0.9644\n",
      "Epoch 434/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0968 - accuracy: 0.9671 - val_loss: 0.0950 - val_accuracy: 0.9636\n",
      "Epoch 435/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0967 - accuracy: 0.9671 - val_loss: 0.0949 - val_accuracy: 0.9644\n",
      "Epoch 436/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0969 - accuracy: 0.9671 - val_loss: 0.0943 - val_accuracy: 0.9644\n",
      "Epoch 437/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0966 - accuracy: 0.9671 - val_loss: 0.0937 - val_accuracy: 0.9662\n",
      "Epoch 438/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0969 - accuracy: 0.9662 - val_loss: 0.0932 - val_accuracy: 0.9671\n",
      "Epoch 439/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0967 - accuracy: 0.9662 - val_loss: 0.0929 - val_accuracy: 0.9671\n",
      "Epoch 440/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0964 - accuracy: 0.9671 - val_loss: 0.0933 - val_accuracy: 0.9662\n",
      "Epoch 441/700\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0964 - accuracy: 0.9671 - val_loss: 0.0944 - val_accuracy: 0.9644\n",
      "Epoch 442/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0963 - accuracy: 0.9662 - val_loss: 0.0948 - val_accuracy: 0.9644\n",
      "Epoch 443/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0961 - accuracy: 0.9671 - val_loss: 0.0941 - val_accuracy: 0.9636\n",
      "Epoch 444/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0961 - accuracy: 0.9671 - val_loss: 0.0936 - val_accuracy: 0.9644\n",
      "Epoch 445/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0961 - accuracy: 0.9671 - val_loss: 0.0930 - val_accuracy: 0.9662\n",
      "Epoch 446/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0959 - accuracy: 0.9671 - val_loss: 0.0928 - val_accuracy: 0.9671\n",
      "Epoch 447/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0960 - accuracy: 0.9680 - val_loss: 0.0927 - val_accuracy: 0.9671\n",
      "Epoch 448/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0960 - accuracy: 0.9671 - val_loss: 0.0927 - val_accuracy: 0.9671\n",
      "Epoch 449/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0960 - accuracy: 0.9680 - val_loss: 0.0924 - val_accuracy: 0.9671\n",
      "Epoch 450/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0955 - accuracy: 0.9671 - val_loss: 0.0922 - val_accuracy: 0.9680\n",
      "Epoch 451/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0957 - accuracy: 0.9671 - val_loss: 0.0930 - val_accuracy: 0.9653\n",
      "Epoch 452/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0955 - accuracy: 0.9671 - val_loss: 0.0938 - val_accuracy: 0.9644\n",
      "Epoch 453/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.0957 - accuracy: 0.9671 - val_loss: 0.0938 - val_accuracy: 0.9627\n",
      "Epoch 454/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0955 - accuracy: 0.9680 - val_loss: 0.0941 - val_accuracy: 0.9644\n",
      "Epoch 455/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0952 - accuracy: 0.9671 - val_loss: 0.0939 - val_accuracy: 0.9644\n",
      "Epoch 456/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0951 - accuracy: 0.9671 - val_loss: 0.0928 - val_accuracy: 0.9636\n",
      "Epoch 457/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0951 - accuracy: 0.9671 - val_loss: 0.0928 - val_accuracy: 0.9644\n",
      "Epoch 458/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0947 - accuracy: 0.9671 - val_loss: 0.0926 - val_accuracy: 0.9662\n",
      "Epoch 459/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0948 - accuracy: 0.9680 - val_loss: 0.0927 - val_accuracy: 0.9644\n",
      "Epoch 460/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0948 - accuracy: 0.9662 - val_loss: 0.0928 - val_accuracy: 0.9671\n",
      "Epoch 461/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0954 - accuracy: 0.9680 - val_loss: 0.0926 - val_accuracy: 0.9644\n",
      "Epoch 462/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0947 - accuracy: 0.9680 - val_loss: 0.0937 - val_accuracy: 0.9662\n",
      "Epoch 463/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.0950 - accuracy: 0.9671 - val_loss: 0.0939 - val_accuracy: 0.9671\n",
      "Epoch 464/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0948 - accuracy: 0.9680 - val_loss: 0.0929 - val_accuracy: 0.9636\n",
      "Epoch 465/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0947 - accuracy: 0.9671 - val_loss: 0.0924 - val_accuracy: 0.9671\n",
      "Epoch 466/700\n",
      "1125/1125 [==============================] - 0s 40us/step - loss: 0.0947 - accuracy: 0.9680 - val_loss: 0.0922 - val_accuracy: 0.9636\n",
      "Epoch 467/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0944 - accuracy: 0.9671 - val_loss: 0.0922 - val_accuracy: 0.9662\n",
      "Epoch 468/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0943 - accuracy: 0.9671 - val_loss: 0.0929 - val_accuracy: 0.9662\n",
      "Epoch 469/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.0946 - accuracy: 0.9662 - val_loss: 0.0926 - val_accuracy: 0.9636\n",
      "Epoch 470/700\n",
      "1125/1125 [==============================] - 0s 13us/step - loss: 0.0942 - accuracy: 0.9662 - val_loss: 0.0931 - val_accuracy: 0.9636\n",
      "Epoch 471/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.0944 - accuracy: 0.9671 - val_loss: 0.0929 - val_accuracy: 0.9636\n",
      "Epoch 472/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0944 - accuracy: 0.9662 - val_loss: 0.0923 - val_accuracy: 0.9636\n",
      "Epoch 473/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0941 - accuracy: 0.9671 - val_loss: 0.0927 - val_accuracy: 0.9680\n",
      "Epoch 474/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0946 - accuracy: 0.9689 - val_loss: 0.0927 - val_accuracy: 0.9671\n",
      "Epoch 475/700\n",
      "1125/1125 [==============================] - 0s 13us/step - loss: 0.0941 - accuracy: 0.9680 - val_loss: 0.0937 - val_accuracy: 0.9671\n",
      "Epoch 476/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0943 - accuracy: 0.9680 - val_loss: 0.0932 - val_accuracy: 0.9662\n",
      "Epoch 477/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0941 - accuracy: 0.9680 - val_loss: 0.0929 - val_accuracy: 0.9644\n",
      "Epoch 478/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.0941 - accuracy: 0.9671 - val_loss: 0.0926 - val_accuracy: 0.9644\n",
      "Epoch 479/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.0938 - accuracy: 0.9671 - val_loss: 0.0924 - val_accuracy: 0.9644\n",
      "Epoch 480/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0939 - accuracy: 0.9671 - val_loss: 0.0923 - val_accuracy: 0.9644\n",
      "Epoch 481/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0936 - accuracy: 0.9671 - val_loss: 0.0921 - val_accuracy: 0.9653\n",
      "Epoch 482/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0939 - accuracy: 0.9671 - val_loss: 0.0918 - val_accuracy: 0.9653\n",
      "Epoch 483/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.0938 - accuracy: 0.9680 - val_loss: 0.0929 - val_accuracy: 0.9662\n",
      "Epoch 484/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0938 - accuracy: 0.9680 - val_loss: 0.0920 - val_accuracy: 0.9680\n",
      "Epoch 485/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0936 - accuracy: 0.9680 - val_loss: 0.0914 - val_accuracy: 0.9644\n",
      "Epoch 486/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0937 - accuracy: 0.9680 - val_loss: 0.0920 - val_accuracy: 0.9636\n",
      "Epoch 487/700\n",
      "1125/1125 [==============================] - ETA: 0s - loss: 0.0991 - accuracy: 0.96 - 0s 19us/step - loss: 0.0936 - accuracy: 0.9671 - val_loss: 0.0925 - val_accuracy: 0.9636\n",
      "Epoch 488/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0937 - accuracy: 0.9671 - val_loss: 0.0928 - val_accuracy: 0.9636\n",
      "Epoch 489/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0936 - accuracy: 0.9671 - val_loss: 0.0920 - val_accuracy: 0.9636\n",
      "Epoch 490/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0936 - accuracy: 0.9662 - val_loss: 0.0920 - val_accuracy: 0.9644\n",
      "Epoch 491/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0933 - accuracy: 0.9671 - val_loss: 0.0928 - val_accuracy: 0.9671\n",
      "Epoch 492/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0933 - accuracy: 0.9689 - val_loss: 0.0913 - val_accuracy: 0.9653\n",
      "Epoch 493/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0934 - accuracy: 0.9680 - val_loss: 0.0917 - val_accuracy: 0.9671\n",
      "Epoch 494/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0936 - accuracy: 0.9671 - val_loss: 0.0917 - val_accuracy: 0.9653\n",
      "Epoch 495/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0933 - accuracy: 0.9680 - val_loss: 0.0916 - val_accuracy: 0.9644\n",
      "Epoch 496/700\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0943 - accuracy: 0.9680 - val_loss: 0.0910 - val_accuracy: 0.9653\n",
      "Epoch 497/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.0932 - accuracy: 0.9689 - val_loss: 0.0918 - val_accuracy: 0.9627\n",
      "Epoch 498/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.0929 - accuracy: 0.9671 - val_loss: 0.0920 - val_accuracy: 0.9636\n",
      "Epoch 499/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.0938 - accuracy: 0.9671 - val_loss: 0.0912 - val_accuracy: 0.9680\n",
      "Epoch 500/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.0931 - accuracy: 0.9680 - val_loss: 0.0909 - val_accuracy: 0.9662\n",
      "Epoch 501/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0929 - accuracy: 0.9671 - val_loss: 0.0905 - val_accuracy: 0.9698\n",
      "Epoch 502/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0932 - accuracy: 0.9689 - val_loss: 0.0902 - val_accuracy: 0.9662\n",
      "Epoch 503/700\n",
      "1125/1125 [==============================] - 0s 13us/step - loss: 0.0932 - accuracy: 0.9689 - val_loss: 0.0905 - val_accuracy: 0.9662\n",
      "Epoch 504/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.0935 - accuracy: 0.9689 - val_loss: 0.0903 - val_accuracy: 0.9662\n",
      "Epoch 505/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0929 - accuracy: 0.9689 - val_loss: 0.0915 - val_accuracy: 0.9671\n",
      "Epoch 506/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0931 - accuracy: 0.9689 - val_loss: 0.0922 - val_accuracy: 0.9671\n",
      "Epoch 507/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0929 - accuracy: 0.9680 - val_loss: 0.0919 - val_accuracy: 0.9644\n",
      "Epoch 508/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0927 - accuracy: 0.9662 - val_loss: 0.0913 - val_accuracy: 0.9644\n",
      "Epoch 509/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0928 - accuracy: 0.9680 - val_loss: 0.0916 - val_accuracy: 0.9644\n",
      "Epoch 510/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0928 - accuracy: 0.9671 - val_loss: 0.0920 - val_accuracy: 0.9644\n",
      "Epoch 511/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0929 - accuracy: 0.9671 - val_loss: 0.0935 - val_accuracy: 0.9662\n",
      "Epoch 512/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0932 - accuracy: 0.9671 - val_loss: 0.0931 - val_accuracy: 0.9671\n",
      "Epoch 513/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.0935 - accuracy: 0.9680 - val_loss: 0.0942 - val_accuracy: 0.9671\n",
      "Epoch 514/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.0935 - accuracy: 0.9689 - val_loss: 0.0924 - val_accuracy: 0.9644\n",
      "Epoch 515/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0930 - accuracy: 0.9680 - val_loss: 0.0917 - val_accuracy: 0.9636\n",
      "Epoch 516/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0928 - accuracy: 0.9671 - val_loss: 0.0919 - val_accuracy: 0.9636\n",
      "Epoch 517/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0924 - accuracy: 0.9671 - val_loss: 0.0925 - val_accuracy: 0.9627\n",
      "Epoch 518/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0925 - accuracy: 0.9662 - val_loss: 0.0927 - val_accuracy: 0.9644\n",
      "Epoch 519/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0932 - accuracy: 0.9698 - val_loss: 0.0921 - val_accuracy: 0.9644\n",
      "Epoch 520/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0924 - accuracy: 0.9680 - val_loss: 0.0935 - val_accuracy: 0.9644\n",
      "Epoch 521/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0927 - accuracy: 0.9689 - val_loss: 0.0934 - val_accuracy: 0.9636\n",
      "Epoch 522/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0926 - accuracy: 0.9680 - val_loss: 0.0933 - val_accuracy: 0.9636\n",
      "Epoch 523/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0928 - accuracy: 0.9680 - val_loss: 0.0918 - val_accuracy: 0.9644\n",
      "Epoch 524/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.0926 - accuracy: 0.9671 - val_loss: 0.0912 - val_accuracy: 0.9653\n",
      "Epoch 525/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0921 - accuracy: 0.9671 - val_loss: 0.0911 - val_accuracy: 0.9671\n",
      "Epoch 526/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0922 - accuracy: 0.9671 - val_loss: 0.0920 - val_accuracy: 0.9671\n",
      "Epoch 527/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0928 - accuracy: 0.9698 - val_loss: 0.0935 - val_accuracy: 0.9653\n",
      "Epoch 528/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0932 - accuracy: 0.9689 - val_loss: 0.0922 - val_accuracy: 0.9662\n",
      "Epoch 529/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0922 - accuracy: 0.9698 - val_loss: 0.0915 - val_accuracy: 0.9680\n",
      "Epoch 530/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0923 - accuracy: 0.9689 - val_loss: 0.0914 - val_accuracy: 0.9689\n",
      "Epoch 531/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0922 - accuracy: 0.9689 - val_loss: 0.0917 - val_accuracy: 0.9653\n",
      "Epoch 532/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0922 - accuracy: 0.9671 - val_loss: 0.0934 - val_accuracy: 0.9662\n",
      "Epoch 533/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0926 - accuracy: 0.9707 - val_loss: 0.0934 - val_accuracy: 0.9662\n",
      "Epoch 534/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0925 - accuracy: 0.9707 - val_loss: 0.0928 - val_accuracy: 0.9662\n",
      "Epoch 535/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0924 - accuracy: 0.9689 - val_loss: 0.0933 - val_accuracy: 0.9653\n",
      "Epoch 536/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0925 - accuracy: 0.9707 - val_loss: 0.0928 - val_accuracy: 0.9662\n",
      "Epoch 537/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.0918 - accuracy: 0.9689 - val_loss: 0.0932 - val_accuracy: 0.9636\n",
      "Epoch 538/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0920 - accuracy: 0.9671 - val_loss: 0.0934 - val_accuracy: 0.9636\n",
      "Epoch 539/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0919 - accuracy: 0.9680 - val_loss: 0.0929 - val_accuracy: 0.9636\n",
      "Epoch 540/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0918 - accuracy: 0.9680 - val_loss: 0.0935 - val_accuracy: 0.9644\n",
      "Epoch 541/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0919 - accuracy: 0.9680 - val_loss: 0.0930 - val_accuracy: 0.9618\n",
      "Epoch 542/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0924 - accuracy: 0.9680 - val_loss: 0.0935 - val_accuracy: 0.9662\n",
      "Epoch 543/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0920 - accuracy: 0.9689 - val_loss: 0.0939 - val_accuracy: 0.9662\n",
      "Epoch 544/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0921 - accuracy: 0.9698 - val_loss: 0.0942 - val_accuracy: 0.9662\n",
      "Epoch 545/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0922 - accuracy: 0.9698 - val_loss: 0.0932 - val_accuracy: 0.9662\n",
      "Epoch 546/700\n",
      "1125/1125 [==============================] - 0s 22us/step - loss: 0.0923 - accuracy: 0.9707 - val_loss: 0.0918 - val_accuracy: 0.9644\n",
      "Epoch 547/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0916 - accuracy: 0.9680 - val_loss: 0.0931 - val_accuracy: 0.9662\n",
      "Epoch 548/700\n",
      "1125/1125 [==============================] - 0s 24us/step - loss: 0.0919 - accuracy: 0.9698 - val_loss: 0.0927 - val_accuracy: 0.9636\n",
      "Epoch 549/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0918 - accuracy: 0.9671 - val_loss: 0.0921 - val_accuracy: 0.9627\n",
      "Epoch 550/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0920 - accuracy: 0.9671 - val_loss: 0.0916 - val_accuracy: 0.9644\n",
      "Epoch 551/700\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0918 - accuracy: 0.9680 - val_loss: 0.0909 - val_accuracy: 0.9671\n",
      "Epoch 552/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0915 - accuracy: 0.9671 - val_loss: 0.0915 - val_accuracy: 0.9662\n",
      "Epoch 553/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0918 - accuracy: 0.9689 - val_loss: 0.0917 - val_accuracy: 0.9636\n",
      "Epoch 554/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0914 - accuracy: 0.9680 - val_loss: 0.0918 - val_accuracy: 0.9609\n",
      "Epoch 555/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0921 - accuracy: 0.9680 - val_loss: 0.0924 - val_accuracy: 0.9618\n",
      "Epoch 556/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0916 - accuracy: 0.9671 - val_loss: 0.0923 - val_accuracy: 0.9618\n",
      "Epoch 557/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0917 - accuracy: 0.9680 - val_loss: 0.0918 - val_accuracy: 0.9618\n",
      "Epoch 558/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0922 - accuracy: 0.9671 - val_loss: 0.0921 - val_accuracy: 0.9636\n",
      "Epoch 559/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0918 - accuracy: 0.9671 - val_loss: 0.0914 - val_accuracy: 0.9644\n",
      "Epoch 560/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0913 - accuracy: 0.9680 - val_loss: 0.0905 - val_accuracy: 0.9644\n",
      "Epoch 561/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0914 - accuracy: 0.9689 - val_loss: 0.0911 - val_accuracy: 0.9671\n",
      "Epoch 562/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0926 - accuracy: 0.9689 - val_loss: 0.0899 - val_accuracy: 0.9662\n",
      "Epoch 563/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0917 - accuracy: 0.9698 - val_loss: 0.0898 - val_accuracy: 0.9671\n",
      "Epoch 564/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0916 - accuracy: 0.9707 - val_loss: 0.0905 - val_accuracy: 0.9680\n",
      "Epoch 565/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0914 - accuracy: 0.9689 - val_loss: 0.0907 - val_accuracy: 0.9662\n",
      "Epoch 566/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0912 - accuracy: 0.9680 - val_loss: 0.0907 - val_accuracy: 0.9653\n",
      "Epoch 567/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0914 - accuracy: 0.9689 - val_loss: 0.0915 - val_accuracy: 0.9653\n",
      "Epoch 568/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0913 - accuracy: 0.9671 - val_loss: 0.0917 - val_accuracy: 0.9653\n",
      "Epoch 569/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0912 - accuracy: 0.9689 - val_loss: 0.0917 - val_accuracy: 0.9627\n",
      "Epoch 570/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0922 - accuracy: 0.9680 - val_loss: 0.0916 - val_accuracy: 0.9653\n",
      "Epoch 571/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0917 - accuracy: 0.9680 - val_loss: 0.0913 - val_accuracy: 0.9653\n",
      "Epoch 572/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0913 - accuracy: 0.9680 - val_loss: 0.0924 - val_accuracy: 0.9662\n",
      "Epoch 573/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0913 - accuracy: 0.9689 - val_loss: 0.0918 - val_accuracy: 0.9644\n",
      "Epoch 574/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0910 - accuracy: 0.9689 - val_loss: 0.0925 - val_accuracy: 0.9662\n",
      "Epoch 575/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0913 - accuracy: 0.9689 - val_loss: 0.0930 - val_accuracy: 0.9662\n",
      "Epoch 576/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0913 - accuracy: 0.9698 - val_loss: 0.0919 - val_accuracy: 0.9644\n",
      "Epoch 577/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0911 - accuracy: 0.9689 - val_loss: 0.0917 - val_accuracy: 0.9662\n",
      "Epoch 578/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0910 - accuracy: 0.9689 - val_loss: 0.0905 - val_accuracy: 0.9671\n",
      "Epoch 579/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0912 - accuracy: 0.9689 - val_loss: 0.0914 - val_accuracy: 0.9671\n",
      "Epoch 580/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0914 - accuracy: 0.9689 - val_loss: 0.0913 - val_accuracy: 0.9671\n",
      "Epoch 581/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0910 - accuracy: 0.9698 - val_loss: 0.0910 - val_accuracy: 0.9680\n",
      "Epoch 582/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0909 - accuracy: 0.9698 - val_loss: 0.0909 - val_accuracy: 0.9671\n",
      "Epoch 583/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0911 - accuracy: 0.9698 - val_loss: 0.0912 - val_accuracy: 0.9671\n",
      "Epoch 584/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0909 - accuracy: 0.9680 - val_loss: 0.0918 - val_accuracy: 0.9662\n",
      "Epoch 585/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.0911 - accuracy: 0.9698 - val_loss: 0.0913 - val_accuracy: 0.9680\n",
      "Epoch 586/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0909 - accuracy: 0.9698 - val_loss: 0.0920 - val_accuracy: 0.9653\n",
      "Epoch 587/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0908 - accuracy: 0.9698 - val_loss: 0.0922 - val_accuracy: 0.9671\n",
      "Epoch 588/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0909 - accuracy: 0.9680 - val_loss: 0.0913 - val_accuracy: 0.9653\n",
      "Epoch 589/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0908 - accuracy: 0.9689 - val_loss: 0.0920 - val_accuracy: 0.9653\n",
      "Epoch 590/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0909 - accuracy: 0.9689 - val_loss: 0.0914 - val_accuracy: 0.9662\n",
      "Epoch 591/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0907 - accuracy: 0.9698 - val_loss: 0.0909 - val_accuracy: 0.9653\n",
      "Epoch 592/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0910 - accuracy: 0.9689 - val_loss: 0.0919 - val_accuracy: 0.9662\n",
      "Epoch 593/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0910 - accuracy: 0.9689 - val_loss: 0.0923 - val_accuracy: 0.9662\n",
      "Epoch 594/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0910 - accuracy: 0.9716 - val_loss: 0.0911 - val_accuracy: 0.9653\n",
      "Epoch 595/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0906 - accuracy: 0.9689 - val_loss: 0.0904 - val_accuracy: 0.9662\n",
      "Epoch 596/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0904 - accuracy: 0.9689 - val_loss: 0.0910 - val_accuracy: 0.9644\n",
      "Epoch 597/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0905 - accuracy: 0.9689 - val_loss: 0.0914 - val_accuracy: 0.9653\n",
      "Epoch 598/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0905 - accuracy: 0.9698 - val_loss: 0.0912 - val_accuracy: 0.9644\n",
      "Epoch 599/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0905 - accuracy: 0.9689 - val_loss: 0.0904 - val_accuracy: 0.9680\n",
      "Epoch 600/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0909 - accuracy: 0.9689 - val_loss: 0.0911 - val_accuracy: 0.9653\n",
      "Epoch 601/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0908 - accuracy: 0.9680 - val_loss: 0.0909 - val_accuracy: 0.9662\n",
      "Epoch 602/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0908 - accuracy: 0.9680 - val_loss: 0.0916 - val_accuracy: 0.9644\n",
      "Epoch 603/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0905 - accuracy: 0.9707 - val_loss: 0.0910 - val_accuracy: 0.9653\n",
      "Epoch 604/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0905 - accuracy: 0.9680 - val_loss: 0.0917 - val_accuracy: 0.9644\n",
      "Epoch 605/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0905 - accuracy: 0.9689 - val_loss: 0.0917 - val_accuracy: 0.9644\n",
      "Epoch 606/700\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0905 - accuracy: 0.9689 - val_loss: 0.0926 - val_accuracy: 0.9644\n",
      "Epoch 607/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0907 - accuracy: 0.9680 - val_loss: 0.0933 - val_accuracy: 0.9653\n",
      "Epoch 608/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0909 - accuracy: 0.9716 - val_loss: 0.0935 - val_accuracy: 0.9644\n",
      "Epoch 609/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0910 - accuracy: 0.9689 - val_loss: 0.0915 - val_accuracy: 0.9653\n",
      "Epoch 610/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0903 - accuracy: 0.9689 - val_loss: 0.0921 - val_accuracy: 0.9644\n",
      "Epoch 611/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0906 - accuracy: 0.9707 - val_loss: 0.0924 - val_accuracy: 0.9653\n",
      "Epoch 612/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0904 - accuracy: 0.9698 - val_loss: 0.0930 - val_accuracy: 0.9644\n",
      "Epoch 613/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0910 - accuracy: 0.9707 - val_loss: 0.0926 - val_accuracy: 0.9662\n",
      "Epoch 614/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0907 - accuracy: 0.9698 - val_loss: 0.0913 - val_accuracy: 0.9653\n",
      "Epoch 615/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0901 - accuracy: 0.9689 - val_loss: 0.0911 - val_accuracy: 0.9653\n",
      "Epoch 616/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0903 - accuracy: 0.9698 - val_loss: 0.0911 - val_accuracy: 0.9662\n",
      "Epoch 617/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0908 - accuracy: 0.9689 - val_loss: 0.0915 - val_accuracy: 0.9653\n",
      "Epoch 618/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0901 - accuracy: 0.9689 - val_loss: 0.0907 - val_accuracy: 0.9671\n",
      "Epoch 619/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0901 - accuracy: 0.9689 - val_loss: 0.0914 - val_accuracy: 0.9671\n",
      "Epoch 620/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0906 - accuracy: 0.9707 - val_loss: 0.0912 - val_accuracy: 0.9671\n",
      "Epoch 621/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0901 - accuracy: 0.9707 - val_loss: 0.0917 - val_accuracy: 0.9662\n",
      "Epoch 622/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0902 - accuracy: 0.9698 - val_loss: 0.0918 - val_accuracy: 0.9653\n",
      "Epoch 623/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0903 - accuracy: 0.9689 - val_loss: 0.0922 - val_accuracy: 0.9662\n",
      "Epoch 624/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0907 - accuracy: 0.9724 - val_loss: 0.0912 - val_accuracy: 0.9644\n",
      "Epoch 625/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0902 - accuracy: 0.9698 - val_loss: 0.0916 - val_accuracy: 0.9653\n",
      "Epoch 626/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0903 - accuracy: 0.9707 - val_loss: 0.0925 - val_accuracy: 0.9644\n",
      "Epoch 627/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0904 - accuracy: 0.9707 - val_loss: 0.0929 - val_accuracy: 0.9644\n",
      "Epoch 628/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0905 - accuracy: 0.9698 - val_loss: 0.0936 - val_accuracy: 0.9653\n",
      "Epoch 629/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0909 - accuracy: 0.9716 - val_loss: 0.0915 - val_accuracy: 0.9644\n",
      "Epoch 630/700\n",
      "1125/1125 [==============================] - 0s 23us/step - loss: 0.0904 - accuracy: 0.9689 - val_loss: 0.0907 - val_accuracy: 0.9653\n",
      "Epoch 631/700\n",
      "1125/1125 [==============================] - 0s 22us/step - loss: 0.0899 - accuracy: 0.9698 - val_loss: 0.0911 - val_accuracy: 0.9644\n",
      "Epoch 632/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0903 - accuracy: 0.9707 - val_loss: 0.0911 - val_accuracy: 0.9653\n",
      "Epoch 633/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.0906 - accuracy: 0.9689 - val_loss: 0.0904 - val_accuracy: 0.9662\n",
      "Epoch 634/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0905 - accuracy: 0.9707 - val_loss: 0.0896 - val_accuracy: 0.9671\n",
      "Epoch 635/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0899 - accuracy: 0.9707 - val_loss: 0.0902 - val_accuracy: 0.9671\n",
      "Epoch 636/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0902 - accuracy: 0.9707 - val_loss: 0.0896 - val_accuracy: 0.9680\n",
      "Epoch 637/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0899 - accuracy: 0.9698 - val_loss: 0.0901 - val_accuracy: 0.9653\n",
      "Epoch 638/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0898 - accuracy: 0.9698 - val_loss: 0.0910 - val_accuracy: 0.9644\n",
      "Epoch 639/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0899 - accuracy: 0.9698 - val_loss: 0.0901 - val_accuracy: 0.9662\n",
      "Epoch 640/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0900 - accuracy: 0.9689 - val_loss: 0.0904 - val_accuracy: 0.9671\n",
      "Epoch 641/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0900 - accuracy: 0.9707 - val_loss: 0.0920 - val_accuracy: 0.9662\n",
      "Epoch 642/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0900 - accuracy: 0.9698 - val_loss: 0.0909 - val_accuracy: 0.9671\n",
      "Epoch 643/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0901 - accuracy: 0.9707 - val_loss: 0.0909 - val_accuracy: 0.9671\n",
      "Epoch 644/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0907 - accuracy: 0.9707 - val_loss: 0.0908 - val_accuracy: 0.9662\n",
      "Epoch 645/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0900 - accuracy: 0.9689 - val_loss: 0.0905 - val_accuracy: 0.9680\n",
      "Epoch 646/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0908 - accuracy: 0.9698 - val_loss: 0.0904 - val_accuracy: 0.9680\n",
      "Epoch 647/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0902 - accuracy: 0.9707 - val_loss: 0.0906 - val_accuracy: 0.9671\n",
      "Epoch 648/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0901 - accuracy: 0.9707 - val_loss: 0.0903 - val_accuracy: 0.9662\n",
      "Epoch 649/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.0899 - accuracy: 0.9689 - val_loss: 0.0906 - val_accuracy: 0.9671\n",
      "Epoch 650/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0904 - accuracy: 0.9707 - val_loss: 0.0908 - val_accuracy: 0.9662\n",
      "Epoch 651/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0901 - accuracy: 0.9689 - val_loss: 0.0920 - val_accuracy: 0.9662\n",
      "Epoch 652/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0901 - accuracy: 0.9698 - val_loss: 0.0914 - val_accuracy: 0.9662\n",
      "Epoch 653/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0900 - accuracy: 0.9716 - val_loss: 0.0903 - val_accuracy: 0.9653\n",
      "Epoch 654/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0897 - accuracy: 0.9689 - val_loss: 0.0915 - val_accuracy: 0.9662\n",
      "Epoch 655/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0900 - accuracy: 0.9707 - val_loss: 0.0924 - val_accuracy: 0.9662\n",
      "Epoch 656/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0902 - accuracy: 0.9716 - val_loss: 0.0915 - val_accuracy: 0.9662\n",
      "Epoch 657/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0901 - accuracy: 0.9698 - val_loss: 0.0919 - val_accuracy: 0.9671\n",
      "Epoch 658/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0899 - accuracy: 0.9724 - val_loss: 0.0914 - val_accuracy: 0.9662\n",
      "Epoch 659/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0901 - accuracy: 0.9716 - val_loss: 0.0918 - val_accuracy: 0.9671\n",
      "Epoch 660/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0904 - accuracy: 0.9707 - val_loss: 0.0912 - val_accuracy: 0.9644\n",
      "Epoch 661/700\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0898 - accuracy: 0.9698 - val_loss: 0.0922 - val_accuracy: 0.9662\n",
      "Epoch 662/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.0902 - accuracy: 0.9707 - val_loss: 0.0923 - val_accuracy: 0.9662\n",
      "Epoch 663/700\n",
      "1125/1125 [==============================] - 0s 15us/step - loss: 0.0899 - accuracy: 0.9724 - val_loss: 0.0911 - val_accuracy: 0.9644\n",
      "Epoch 664/700\n",
      "1125/1125 [==============================] - 0s 14us/step - loss: 0.0897 - accuracy: 0.9698 - val_loss: 0.0910 - val_accuracy: 0.9644\n",
      "Epoch 665/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0900 - accuracy: 0.9707 - val_loss: 0.0908 - val_accuracy: 0.9671\n",
      "Epoch 666/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0907 - accuracy: 0.9698 - val_loss: 0.0903 - val_accuracy: 0.9671\n",
      "Epoch 667/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0903 - accuracy: 0.9707 - val_loss: 0.0898 - val_accuracy: 0.9689\n",
      "Epoch 668/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0901 - accuracy: 0.9698 - val_loss: 0.0894 - val_accuracy: 0.9707\n",
      "Epoch 669/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0901 - accuracy: 0.9707 - val_loss: 0.0900 - val_accuracy: 0.9689\n",
      "Epoch 670/700\n",
      "1125/1125 [==============================] - 0s 16us/step - loss: 0.0897 - accuracy: 0.9716 - val_loss: 0.0897 - val_accuracy: 0.9689\n",
      "Epoch 671/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0899 - accuracy: 0.9716 - val_loss: 0.0899 - val_accuracy: 0.9689\n",
      "Epoch 672/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0896 - accuracy: 0.9716 - val_loss: 0.0900 - val_accuracy: 0.9680\n",
      "Epoch 673/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0899 - accuracy: 0.9716 - val_loss: 0.0902 - val_accuracy: 0.9662\n",
      "Epoch 674/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0897 - accuracy: 0.9689 - val_loss: 0.0907 - val_accuracy: 0.9644\n",
      "Epoch 675/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0896 - accuracy: 0.9698 - val_loss: 0.0904 - val_accuracy: 0.9653\n",
      "Epoch 676/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0898 - accuracy: 0.9698 - val_loss: 0.0908 - val_accuracy: 0.9653\n",
      "Epoch 677/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.0897 - accuracy: 0.9689 - val_loss: 0.0913 - val_accuracy: 0.9636\n",
      "Epoch 678/700\n",
      "1125/1125 [==============================] - 0s 30us/step - loss: 0.0897 - accuracy: 0.9698 - val_loss: 0.0908 - val_accuracy: 0.9653\n",
      "Epoch 679/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0906 - accuracy: 0.9689 - val_loss: 0.0906 - val_accuracy: 0.9671\n",
      "Epoch 680/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0897 - accuracy: 0.9707 - val_loss: 0.0919 - val_accuracy: 0.9653\n",
      "Epoch 681/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0895 - accuracy: 0.9698 - val_loss: 0.0916 - val_accuracy: 0.9644\n",
      "Epoch 682/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0896 - accuracy: 0.9707 - val_loss: 0.0914 - val_accuracy: 0.9653\n",
      "Epoch 683/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0896 - accuracy: 0.9689 - val_loss: 0.0906 - val_accuracy: 0.9671\n",
      "Epoch 684/700\n",
      "1125/1125 [==============================] - 0s 21us/step - loss: 0.0897 - accuracy: 0.9698 - val_loss: 0.0912 - val_accuracy: 0.9671\n",
      "Epoch 685/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0895 - accuracy: 0.9707 - val_loss: 0.0917 - val_accuracy: 0.9662\n",
      "Epoch 686/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0897 - accuracy: 0.9716 - val_loss: 0.0912 - val_accuracy: 0.9662\n",
      "Epoch 687/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0897 - accuracy: 0.9707 - val_loss: 0.0923 - val_accuracy: 0.9662\n",
      "Epoch 688/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0896 - accuracy: 0.9707 - val_loss: 0.0932 - val_accuracy: 0.9662\n",
      "Epoch 689/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0897 - accuracy: 0.9716 - val_loss: 0.0931 - val_accuracy: 0.9644\n",
      "Epoch 690/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0897 - accuracy: 0.9689 - val_loss: 0.0916 - val_accuracy: 0.9662\n",
      "Epoch 691/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0894 - accuracy: 0.9707 - val_loss: 0.0917 - val_accuracy: 0.9644\n",
      "Epoch 692/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0894 - accuracy: 0.9698 - val_loss: 0.0911 - val_accuracy: 0.9680\n",
      "Epoch 693/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0895 - accuracy: 0.9689 - val_loss: 0.0907 - val_accuracy: 0.9671\n",
      "Epoch 694/700\n",
      "1125/1125 [==============================] - 0s 20us/step - loss: 0.0896 - accuracy: 0.9716 - val_loss: 0.0903 - val_accuracy: 0.9671\n",
      "Epoch 695/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0898 - accuracy: 0.9689 - val_loss: 0.0910 - val_accuracy: 0.9671\n",
      "Epoch 696/700\n",
      "1125/1125 [==============================] - 0s 19us/step - loss: 0.0898 - accuracy: 0.9707 - val_loss: 0.0908 - val_accuracy: 0.9680\n",
      "Epoch 697/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0898 - accuracy: 0.9707 - val_loss: 0.0903 - val_accuracy: 0.9698\n",
      "Epoch 698/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0895 - accuracy: 0.9716 - val_loss: 0.0902 - val_accuracy: 0.9698\n",
      "Epoch 699/700\n",
      "1125/1125 [==============================] - 0s 18us/step - loss: 0.0895 - accuracy: 0.9716 - val_loss: 0.0904 - val_accuracy: 0.9671\n",
      "Epoch 700/700\n",
      "1125/1125 [==============================] - 0s 17us/step - loss: 0.0895 - accuracy: 0.9707 - val_loss: 0.0900 - val_accuracy: 0.9698\n"
     ]
    }
   ],
   "source": [
    "hist  = model.fit(X_train_new,Y_train_new,epochs=700,batch_size=512,validation_data=(X_val,Y_val))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 272,
   "metadata": {},
   "outputs": [],
   "source": [
    "h = hist.history"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 273,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(h['val_loss'],label = \"Validation Loss\")\n",
    "plt.plot(h['loss'],label=\"Training Loss\")\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.ylabel(\"Loss\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 274,
   "metadata": {},
   "outputs": [],
   "source": [
    "output = model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 275,
   "metadata": {},
   "outputs": [],
   "source": [
    "m = output.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 276,
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in range(m):\n",
    "    if(output[i]>0.5):\n",
    "        output[i]=int(1)\n",
    "    else:\n",
    "        output[i]=int(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 277,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [1.]\n",
      " [0.]\n",
      " [0.]\n",
      " [0.]]\n"
     ]
    }
   ],
   "source": [
    "print(output)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 278,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.DataFrame(data=output,columns=[\"label\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 279,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(750, 1)"
      ]
     },
     "execution_count": 279,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 280,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.to_csv('done.csv',index=False) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 281,
   "metadata": {},
   "outputs": [],
   "source": [
    "from visualize import plot_decision_boundary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 282,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_decision_boundary(lambda X_test:model.predict(X_test),X_train,Y_train)"
   ]
  }
 ],
 "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": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}