Assignment 1. Pixel regression.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Assignment 1: Pixel Regression.\n",
    "\n",
    "Can we recover an image by learning a deep regression map from pixels $(x,y)$ to colors $(r,g,b)$?\n",
    "\n",
    "Our target image will be Mona Lisa:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fc6cdfa7f98>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "(199, 153, 3)"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import matplotlib.image as mpimg\n",
    "import matplotlib.pylab as plt\n",
    "import numpy as np\n",
    "%matplotlib inline\n",
    "\n",
    "im = mpimg.imread(\"data/monalisa.jpg\")\n",
    "\n",
    "plt.imshow(im)\n",
    "plt.show()\n",
    "im.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Ourt training dataset will be composed of pixels locatiions and input and pixel values as output:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Samples: 30447\n",
      "(x,y): [0. 0.] \n",
      " (r,g,b): [ 85 105 116]\n"
     ]
    }
   ],
   "source": [
    "X_train = []\n",
    "Y_train = []\n",
    "for i in range(im.shape[0]):\n",
    "    for j in range(im.shape[1]):\n",
    "        X_train.append([float(i),float(j)])\n",
    "        Y_train.append(im[i][j])\n",
    "        \n",
    "X_train = np.array(X_train)\n",
    "Y_train = np.array(Y_train)\n",
    "print('Samples:', X_train.shape[0])\n",
    "print('(x,y):', X_train[0],'\\n', '(r,g,b):',Y_train[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Our objective is to train a deep MLP that is able to reconstruct the image:\n",
    "\n",
    "![alt text](images/result.png)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model with 5 hidden layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n",
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_1 (Dense)              (None, 500)               1500      \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 500)               250500    \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 500)               250500    \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 500)               250500    \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 500)               250500    \n",
      "_________________________________________________________________\n",
      "dense_6 (Dense)              (None, 3)                 1503      \n",
      "=================================================================\n",
      "Total params: 1,005,003\n",
      "Trainable params: 1,005,003\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "import keras\n",
    "from keras.models import Sequential\n",
    "from keras.layers.core import Dense, Activation, Dropout\n",
    "from sklearn import preprocessing\n",
    "from keras.optimizers import *\n",
    "from keras.initializers import *\n",
    "\n",
    "# your model here\n",
    "# hints: k*10^2 neurons per layer + good initialization + deep network (>2 layers)\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "model.add(Dense(units=500, input_shape=(2,), \n",
    "                kernel_initializer='normal',\n",
    "                bias_initializer='zeros',\n",
    "                activation='relu'))\n",
    "\n",
    "model.add(Dense(units=500, \n",
    "                kernel_initializer='normal',\n",
    "                bias_initializer='zeros',\n",
    "                activation='relu'))\n",
    "\n",
    "model.add(Dense(units=500,\n",
    "                kernel_initializer='normal',\n",
    "                bias_initializer='zeros',\n",
    "                activation='relu'))\n",
    "\n",
    "model.add(Dense(units=500,\n",
    "                kernel_initializer='normal',\n",
    "                bias_initializer='zeros',\n",
    "                activation='relu'))\n",
    "\n",
    "model.add(Dense(units=500,\n",
    "                kernel_initializer='normal',\n",
    "                bias_initializer='zeros',\n",
    "                activation='relu'))\n",
    "\n",
    "model.add(Dense(units=3, \n",
    "                kernel_initializer='normal', \n",
    "                bias_initializer='zeros'))\n",
    "\n",
    "model.compile(loss='mean_squared_error', \n",
    "              optimizer='adam',\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/utils/validation.py:475: DataConversionWarning: Data with input dtype uint8 was converted to float64 by MinMaxScaler.\n",
      "  warnings.warn(msg, DataConversionWarning)\n"
     ]
    }
   ],
   "source": [
    "# construct two scalers, one for the features and one for the targets \n",
    "# (default range values are 0-1)\n",
    "scaler_features = preprocessing.MinMaxScaler()\n",
    "scaler_target = preprocessing.MinMaxScaler()\n",
    "\n",
    "# scale features and targets between 0 and 1\n",
    "X_train_scaled = scaler_features.fit_transform(X_train)\n",
    "Y_train_scaled = scaler_target.fit_transform(Y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 246/250\n",
      "30447/30447 [==============================] - 18s 600us/step - loss: 0.0051 - acc: 0.8484\n",
      "Epoch 247/250\n",
      "30447/30447 [==============================] - 18s 595us/step - loss: 0.0053 - acc: 0.8465\n",
      "Epoch 248/250\n",
      "30447/30447 [==============================] - 18s 597us/step - loss: 0.0051 - acc: 0.8459\n",
      "Epoch 249/250\n",
      "30447/30447 [==============================] - 20s 643us/step - loss: 0.0056 - acc: 0.8415\n",
      "Epoch 250/250\n",
      "30447/30447 [==============================] - 18s 578us/step - loss: 0.0051 - acc: 0.8412\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7f2b96fd0080>"
      ]
     },
     "execution_count": 145,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# use this cell to find the best model architecture\n",
    "model.fit(X_train_scaled, Y_train_scaled, epochs=250, shuffle=True, verbose=1, batch_size=32, initial_epoch=245)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# predict \n",
    "Y_pred = model.predict(X_train_scaled, batch_size=10000)\n",
    "\n",
    "# convert targets back to the original scale\n",
    "Y = scaler_target.inverse_transform(Y_pred)\n",
    "\n",
    "k = 0\n",
    "# create an empty image of type unsigned int (8 bits) and same shape as the original\n",
    "im_out = np.empty(shape=im.shape, dtype=np.uint8)\n",
    "for i in range(im.shape[0]):\n",
    "    for j in range(im.shape[1]):\n",
    "        im_out[i,j]= Y[k]\n",
    "        k += 1\n",
    "        \n",
    "plt.imshow(im_out)\n",
    "plt.savefig(\"mymona_250_00051.png\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model with 4 hidden layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_7 (Dense)              (None, 500)               1500      \n",
      "_________________________________________________________________\n",
      "dense_8 (Dense)              (None, 500)               250500    \n",
      "_________________________________________________________________\n",
      "dense_9 (Dense)              (None, 500)               250500    \n",
      "_________________________________________________________________\n",
      "dense_10 (Dense)             (None, 500)               250500    \n",
      "_________________________________________________________________\n",
      "dense_11 (Dense)             (None, 3)                 1503      \n",
      "=================================================================\n",
      "Total params: 754,503\n",
      "Trainable params: 754,503\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "import keras\n",
    "from keras.models import Sequential\n",
    "from keras.layers.core import Dense, Activation, Dropout\n",
    "from sklearn import preprocessing\n",
    "from keras.optimizers import *\n",
    "from keras.initializers import *\n",
    "\n",
    "# your model here\n",
    "# hints: k*10^2 neurons per layer + good initialization + deep network (>2 layers)\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "model.add(Dense(units=500, input_shape=(2,), \n",
    "                kernel_initializer='normal',\n",
    "                bias_initializer='zeros',\n",
    "                activation='relu'))\n",
    "\n",
    "model.add(Dense(units=500, \n",
    "                kernel_initializer='normal',\n",
    "                bias_initializer='zeros',\n",
    "                activation='relu'))\n",
    "\n",
    "model.add(Dense(units=500,\n",
    "                kernel_initializer='normal',\n",
    "                bias_initializer='zeros',\n",
    "                activation='relu'))\n",
    "\n",
    "model.add(Dense(units=500,\n",
    "                kernel_initializer='normal',\n",
    "                bias_initializer='zeros',\n",
    "                activation='relu'))\n",
    "\n",
    "model.add(Dense(units=3, \n",
    "                kernel_initializer='normal', \n",
    "                bias_initializer='zeros'))\n",
    "\n",
    "model.compile(loss='mean_squared_error', \n",
    "              optimizer=RMSprop(lr=0.001, rho=0.9, epsilon=None, decay=10e-6))\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1001/1500\n",
      "30447/30447 [==============================] - 12s 407us/step - loss: 0.0018\n",
      "Epoch 1002/1500\n",
      "30447/30447 [==============================] - 12s 404us/step - loss: 0.0018\n",
      "Epoch 1003/1500\n",
      "30447/30447 [==============================] - 13s 416us/step - loss: 0.0019\n",
      "Epoch 1004/1500\n",
      "30447/30447 [==============================] - 12s 391us/step - loss: 0.0018\n",
      "Epoch 1005/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0018\n",
      "Epoch 1006/1500\n",
      "30447/30447 [==============================] - 11s 365us/step - loss: 0.0018\n",
      "Epoch 1007/1500\n",
      "30447/30447 [==============================] - 11s 365us/step - loss: 0.0018\n",
      "Epoch 1008/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0018\n",
      "Epoch 1009/1500\n",
      "30447/30447 [==============================] - 12s 410us/step - loss: 0.0018\n",
      "Epoch 1010/1500\n",
      "30447/30447 [==============================] - 12s 399us/step - loss: 0.0018\n",
      "Epoch 1011/1500\n",
      "30447/30447 [==============================] - 12s 402us/step - loss: 0.0018\n",
      "Epoch 1012/1500\n",
      "30447/30447 [==============================] - 13s 415us/step - loss: 0.0018\n",
      "Epoch 1013/1500\n",
      "30447/30447 [==============================] - 13s 413us/step - loss: 0.0018\n",
      "Epoch 1014/1500\n",
      "30447/30447 [==============================] - 12s 409us/step - loss: 0.0018\n",
      "Epoch 1015/1500\n",
      "30447/30447 [==============================] - 13s 416us/step - loss: 0.0018\n",
      "Epoch 1016/1500\n",
      "30447/30447 [==============================] - 12s 404us/step - loss: 0.0018\n",
      "Epoch 1017/1500\n",
      "30447/30447 [==============================] - 13s 416us/step - loss: 0.0018\n",
      "Epoch 1018/1500\n",
      "30447/30447 [==============================] - 13s 424us/step - loss: 0.0018\n",
      "Epoch 1019/1500\n",
      "30447/30447 [==============================] - 13s 423us/step - loss: 0.0018\n",
      "Epoch 1020/1500\n",
      "30447/30447 [==============================] - 12s 406us/step - loss: 0.0018\n",
      "Epoch 1021/1500\n",
      "30447/30447 [==============================] - 12s 408us/step - loss: 0.0018\n",
      "Epoch 1022/1500\n",
      "30447/30447 [==============================] - 14s 444us/step - loss: 0.0018\n",
      "Epoch 1023/1500\n",
      "30447/30447 [==============================] - 13s 434us/step - loss: 0.0018\n",
      "Epoch 1024/1500\n",
      "30447/30447 [==============================] - 13s 415us/step - loss: 0.0018\n",
      "Epoch 1025/1500\n",
      "30447/30447 [==============================] - 13s 411us/step - loss: 0.0018\n",
      "Epoch 1026/1500\n",
      "30447/30447 [==============================] - 13s 436us/step - loss: 0.0018\n",
      "Epoch 1027/1500\n",
      "30447/30447 [==============================] - 13s 413us/step - loss: 0.0018\n",
      "Epoch 1028/1500\n",
      "30447/30447 [==============================] - 12s 403us/step - loss: 0.0018\n",
      "Epoch 1029/1500\n",
      "30447/30447 [==============================] - 12s 406us/step - loss: 0.0018\n",
      "Epoch 1030/1500\n",
      "30447/30447 [==============================] - 12s 406us/step - loss: 0.0018\n",
      "Epoch 1031/1500\n",
      "30447/30447 [==============================] - 12s 408us/step - loss: 0.0018\n",
      "Epoch 1032/1500\n",
      "30447/30447 [==============================] - 12s 409us/step - loss: 0.0018\n",
      "Epoch 1033/1500\n",
      "30447/30447 [==============================] - 12s 410us/step - loss: 0.0018\n",
      "Epoch 1034/1500\n",
      "30447/30447 [==============================] - 13s 413us/step - loss: 0.0018\n",
      "Epoch 1035/1500\n",
      "30447/30447 [==============================] - 13s 414us/step - loss: 0.0018\n",
      "Epoch 1036/1500\n",
      "30447/30447 [==============================] - 13s 436us/step - loss: 0.0018\n",
      "Epoch 1037/1500\n",
      "30447/30447 [==============================] - 14s 464us/step - loss: 0.0018\n",
      "Epoch 1038/1500\n",
      "30447/30447 [==============================] - 13s 415us/step - loss: 0.0018\n",
      "Epoch 1039/1500\n",
      "30447/30447 [==============================] - 13s 415us/step - loss: 0.0018\n",
      "Epoch 1040/1500\n",
      "30447/30447 [==============================] - 13s 415us/step - loss: 0.0018\n",
      "Epoch 1041/1500\n",
      "30447/30447 [==============================] - 13s 429us/step - loss: 0.0018\n",
      "Epoch 1042/1500\n",
      "30447/30447 [==============================] - 13s 420us/step - loss: 0.0018\n",
      "Epoch 1043/1500\n",
      "30447/30447 [==============================] - 13s 422us/step - loss: 0.0018\n",
      "Epoch 1044/1500\n",
      "30447/30447 [==============================] - 13s 420us/step - loss: 0.0018\n",
      "Epoch 1045/1500\n",
      "30447/30447 [==============================] - 13s 419us/step - loss: 0.0018\n",
      "Epoch 1046/1500\n",
      "30447/30447 [==============================] - 13s 421us/step - loss: 0.0018\n",
      "Epoch 1047/1500\n",
      "30447/30447 [==============================] - 14s 471us/step - loss: 0.0018\n",
      "Epoch 1048/1500\n",
      "30447/30447 [==============================] - 12s 405us/step - loss: 0.0018\n",
      "Epoch 1049/1500\n",
      "30447/30447 [==============================] - 12s 402us/step - loss: 0.0018\n",
      "Epoch 1050/1500\n",
      "30447/30447 [==============================] - 12s 402us/step - loss: 0.0018\n",
      "Epoch 1051/1500\n",
      "30447/30447 [==============================] - 12s 402us/step - loss: 0.0018\n",
      "Epoch 1052/1500\n",
      "30447/30447 [==============================] - 12s 402us/step - loss: 0.0018\n",
      "Epoch 1053/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0018\n",
      "Epoch 1054/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0018\n",
      "Epoch 1055/1500\n",
      "30447/30447 [==============================] - 12s 383us/step - loss: 0.0018\n",
      "Epoch 1056/1500\n",
      "30447/30447 [==============================] - 12s 383us/step - loss: 0.0018\n",
      "Epoch 1057/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0018\n",
      "Epoch 1058/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0018\n",
      "Epoch 1059/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0018\n",
      "Epoch 1060/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0018\n",
      "Epoch 1061/1500\n",
      "30447/30447 [==============================] - 12s 383us/step - loss: 0.0018\n",
      "Epoch 1062/1500\n",
      "30447/30447 [==============================] - 12s 383us/step - loss: 0.0018\n",
      "Epoch 1063/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0018\n",
      "Epoch 1064/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0018\n",
      "Epoch 1065/1500\n",
      "30447/30447 [==============================] - 12s 380us/step - loss: 0.0018\n",
      "Epoch 1066/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0018\n",
      "Epoch 1067/1500\n",
      "30447/30447 [==============================] - 12s 380us/step - loss: 0.0018\n",
      "Epoch 1068/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0018\n",
      "Epoch 1069/1500\n",
      "30447/30447 [==============================] - 12s 385us/step - loss: 0.0018\n",
      "Epoch 1070/1500\n",
      "30447/30447 [==============================] - 12s 384us/step - loss: 0.0018\n",
      "Epoch 1071/1500\n",
      "30447/30447 [==============================] - 12s 383us/step - loss: 0.0018\n",
      "Epoch 1072/1500\n",
      "30447/30447 [==============================] - 12s 385us/step - loss: 0.0018\n",
      "Epoch 1073/1500\n",
      "30447/30447 [==============================] - 12s 384us/step - loss: 0.0018\n",
      "Epoch 1074/1500\n",
      "30447/30447 [==============================] - 12s 385us/step - loss: 0.0018\n",
      "Epoch 1075/1500\n",
      "30447/30447 [==============================] - 12s 384us/step - loss: 0.0018\n",
      "Epoch 1076/1500\n",
      "30447/30447 [==============================] - 12s 386us/step - loss: 0.0018\n",
      "Epoch 1077/1500\n",
      "30447/30447 [==============================] - 12s 386us/step - loss: 0.0018\n",
      "Epoch 1078/1500\n",
      "30447/30447 [==============================] - 12s 384us/step - loss: 0.0018\n",
      "Epoch 1079/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0018\n",
      "Epoch 1080/1500\n",
      "30447/30447 [==============================] - 12s 385us/step - loss: 0.0018\n",
      "Epoch 1081/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0018\n",
      "Epoch 1082/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0018\n",
      "Epoch 1083/1500\n",
      "30447/30447 [==============================] - 12s 383us/step - loss: 0.0018\n",
      "Epoch 1084/1500\n",
      "30447/30447 [==============================] - 12s 383us/step - loss: 0.0018\n",
      "Epoch 1085/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0018\n",
      "Epoch 1086/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0018\n",
      "Epoch 1087/1500\n",
      "30447/30447 [==============================] - 12s 383us/step - loss: 0.0017\n",
      "Epoch 1088/1500\n",
      "30447/30447 [==============================] - 12s 383us/step - loss: 0.0017\n",
      "Epoch 1089/1500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0018\n",
      "Epoch 1090/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0018\n",
      "Epoch 1091/1500\n",
      "30447/30447 [==============================] - 11s 375us/step - loss: 0.0017\n",
      "Epoch 1092/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1093/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0018\n",
      "Epoch 1094/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1095/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1096/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0018\n",
      "Epoch 1097/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1098/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0018\n",
      "Epoch 1099/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1100/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1101/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1102/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1103/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0017\n",
      "Epoch 1104/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0017\n",
      "Epoch 1105/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0017\n",
      "Epoch 1106/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0017\n",
      "Epoch 1107/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0018\n",
      "Epoch 1108/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0017\n",
      "Epoch 1109/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0017\n",
      "Epoch 1110/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0017\n",
      "Epoch 1111/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1112/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0017\n",
      "Epoch 1113/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0017\n",
      "Epoch 1114/1500\n",
      "30447/30447 [==============================] - 11s 375us/step - loss: 0.0017\n",
      "Epoch 1115/1500\n",
      "30447/30447 [==============================] - 11s 375us/step - loss: 0.0017\n",
      "Epoch 1116/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1117/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0017\n",
      "Epoch 1118/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1119/1500\n",
      "30447/30447 [==============================] - 11s 375us/step - loss: 0.0017\n",
      "Epoch 1120/1500\n",
      "30447/30447 [==============================] - 11s 378us/step - loss: 0.0017\n",
      "Epoch 1121/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1122/1500\n",
      "30447/30447 [==============================] - 11s 378us/step - loss: 0.0017\n",
      "Epoch 1123/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1124/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1125/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1126/1500\n",
      "30447/30447 [==============================] - 11s 375us/step - loss: 0.0017\n",
      "Epoch 1127/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0017\n",
      "Epoch 1128/1500\n",
      "30447/30447 [==============================] - 11s 374us/step - loss: 0.0017\n",
      "Epoch 1129/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0017\n",
      "Epoch 1130/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0017\n",
      "Epoch 1131/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1132/1500\n",
      "30447/30447 [==============================] - 12s 380us/step - loss: 0.0017\n",
      "Epoch 1133/1500\n",
      "30447/30447 [==============================] - 12s 380us/step - loss: 0.0017\n",
      "Epoch 1134/1500\n",
      "30447/30447 [==============================] - 12s 380us/step - loss: 0.0017\n",
      "Epoch 1135/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0017\n",
      "Epoch 1136/1500\n",
      "30447/30447 [==============================] - 12s 380us/step - loss: 0.0017\n",
      "Epoch 1137/1500\n",
      "30447/30447 [==============================] - 12s 380us/step - loss: 0.0017\n",
      "Epoch 1138/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0017\n",
      "Epoch 1139/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0017\n",
      "Epoch 1140/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1141/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0017\n",
      "Epoch 1142/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1143/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1144/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0017\n",
      "Epoch 1145/1500\n",
      "30447/30447 [==============================] - 12s 394us/step - loss: 0.0017\n",
      "Epoch 1146/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0017\n",
      "Epoch 1147/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0017\n",
      "Epoch 1148/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1149/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1150/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1151/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1152/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1153/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1154/1500\n",
      "30447/30447 [==============================] - 11s 378us/step - loss: 0.0017\n",
      "Epoch 1155/1500\n",
      "30447/30447 [==============================] - 11s 378us/step - loss: 0.0017\n",
      "Epoch 1156/1500\n",
      "30447/30447 [==============================] - 11s 378us/step - loss: 0.0017\n",
      "Epoch 1157/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1158/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1159/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1160/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0017\n",
      "Epoch 1161/1500\n",
      "30447/30447 [==============================] - 11s 378us/step - loss: 0.0017\n",
      "Epoch 1162/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1163/1500\n",
      "30447/30447 [==============================] - 11s 378us/step - loss: 0.0017\n",
      "Epoch 1164/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0017\n",
      "Epoch 1165/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1166/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1167/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1168/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0017\n",
      "Epoch 1169/1500\n",
      "30447/30447 [==============================] - 12s 398us/step - loss: 0.0017\n",
      "Epoch 1170/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0017\n",
      "Epoch 1171/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1172/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0017\n",
      "Epoch 1173/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0017\n",
      "Epoch 1174/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1175/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1176/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0017\n",
      "Epoch 1177/1500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1178/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1179/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1180/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1181/1500\n",
      "30447/30447 [==============================] - 12s 380us/step - loss: 0.0017\n",
      "Epoch 1182/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1183/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1184/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1185/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1186/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0017\n",
      "Epoch 1187/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1188/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0017\n",
      "Epoch 1189/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0017\n",
      "Epoch 1190/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1191/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0017\n",
      "Epoch 1192/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1193/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0017\n",
      "Epoch 1194/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0016\n",
      "Epoch 1195/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0016\n",
      "Epoch 1196/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0016\n",
      "Epoch 1197/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0016\n",
      "Epoch 1198/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0016\n",
      "Epoch 1199/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1200/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0016\n",
      "Epoch 1201/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0016\n",
      "Epoch 1202/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0016\n",
      "Epoch 1203/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0016\n",
      "Epoch 1204/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0016\n",
      "Epoch 1205/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0017\n",
      "Epoch 1206/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0016\n",
      "Epoch 1207/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0016\n",
      "Epoch 1208/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0016\n",
      "Epoch 1209/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0016\n",
      "Epoch 1210/1500\n",
      "30447/30447 [==============================] - 11s 375us/step - loss: 0.0016\n",
      "Epoch 1211/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0016\n",
      "Epoch 1212/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0017\n",
      "Epoch 1213/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0016\n",
      "Epoch 1214/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0016\n",
      "Epoch 1215/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0016\n",
      "Epoch 1216/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0016\n",
      "Epoch 1217/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0016\n",
      "Epoch 1218/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0016\n",
      "Epoch 1219/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0016\n",
      "Epoch 1220/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0016\n",
      "Epoch 1221/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0016\n",
      "Epoch 1222/1500\n",
      "30447/30447 [==============================] - 12s 378us/step - loss: 0.0016\n",
      "Epoch 1223/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0016\n",
      "Epoch 1224/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0016\n",
      "Epoch 1225/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0016\n",
      "Epoch 1226/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0016\n",
      "Epoch 1227/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0016\n",
      "Epoch 1228/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0016\n",
      "Epoch 1229/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0016\n",
      "Epoch 1230/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0016\n",
      "Epoch 1231/1500\n",
      "30447/30447 [==============================] - 12s 403us/step - loss: 0.0016\n",
      "Epoch 1232/1500\n",
      "30447/30447 [==============================] - 13s 424us/step - loss: 0.0016\n",
      "Epoch 1233/1500\n",
      "30447/30447 [==============================] - 14s 459us/step - loss: 0.0016\n",
      "Epoch 1234/1500\n",
      "30447/30447 [==============================] - 13s 417us/step - loss: 0.0016\n",
      "Epoch 1235/1500\n",
      "30447/30447 [==============================] - 15s 492us/step - loss: 0.0016\n",
      "Epoch 1236/1500\n",
      "30447/30447 [==============================] - 12s 403us/step - loss: 0.0016\n",
      "Epoch 1237/1500\n",
      "30447/30447 [==============================] - 11s 371us/step - loss: 0.0016\n",
      "Epoch 1238/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0016\n",
      "Epoch 1239/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1240/1500\n",
      "30447/30447 [==============================] - 11s 371us/step - loss: 0.0016\n",
      "Epoch 1241/1500\n",
      "30447/30447 [==============================] - 11s 373us/step - loss: 0.0016\n",
      "Epoch 1242/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0016\n",
      "Epoch 1243/1500\n",
      "30447/30447 [==============================] - 11s 371us/step - loss: 0.0016\n",
      "Epoch 1244/1500\n",
      "30447/30447 [==============================] - 13s 429us/step - loss: 0.0016\n",
      "Epoch 1245/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1246/1500\n",
      "30447/30447 [==============================] - 11s 372us/step - loss: 0.0016\n",
      "Epoch 1247/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1248/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1249/1500\n",
      "30447/30447 [==============================] - 11s 370us/step - loss: 0.0016\n",
      "Epoch 1250/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1251/1500\n",
      "30447/30447 [==============================] - 11s 372us/step - loss: 0.0016\n",
      "Epoch 1252/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1253/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1254/1500\n",
      "30447/30447 [==============================] - 11s 370us/step - loss: 0.0016\n",
      "Epoch 1255/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0016\n",
      "Epoch 1256/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1257/1500\n",
      "30447/30447 [==============================] - 11s 370us/step - loss: 0.0016\n",
      "Epoch 1258/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1259/1500\n",
      "30447/30447 [==============================] - 11s 371us/step - loss: 0.0016\n",
      "Epoch 1260/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1261/1500\n",
      "30447/30447 [==============================] - 11s 370us/step - loss: 0.0016\n",
      "Epoch 1262/1500\n",
      "30447/30447 [==============================] - 11s 376us/step - loss: 0.0016\n",
      "Epoch 1263/1500\n",
      "30447/30447 [==============================] - 11s 367us/step - loss: 0.0016\n",
      "Epoch 1264/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0016\n",
      "Epoch 1265/1500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0016\n",
      "Epoch 1266/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1267/1500\n",
      "30447/30447 [==============================] - 11s 375us/step - loss: 0.0016\n",
      "Epoch 1268/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0016\n",
      "Epoch 1269/1500\n",
      "30447/30447 [==============================] - 11s 367us/step - loss: 0.0016\n",
      "Epoch 1270/1500\n",
      "30447/30447 [==============================] - 11s 370us/step - loss: 0.0016\n",
      "Epoch 1271/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0016\n",
      "Epoch 1272/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1273/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0016\n",
      "Epoch 1274/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1275/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0016\n",
      "Epoch 1276/1500\n",
      "30447/30447 [==============================] - 11s 367us/step - loss: 0.0016\n",
      "Epoch 1277/1500\n",
      "30447/30447 [==============================] - 11s 366us/step - loss: 0.0016\n",
      "Epoch 1278/1500\n",
      "30447/30447 [==============================] - 11s 371us/step - loss: 0.0016\n",
      "Epoch 1279/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1280/1500\n",
      "30447/30447 [==============================] - 11s 367us/step - loss: 0.0016\n",
      "Epoch 1281/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0016\n",
      "Epoch 1282/1500\n",
      "30447/30447 [==============================] - 11s 367us/step - loss: 0.0016\n",
      "Epoch 1283/1500\n",
      "30447/30447 [==============================] - 11s 370us/step - loss: 0.0016\n",
      "Epoch 1284/1500\n",
      "30447/30447 [==============================] - 11s 367us/step - loss: 0.0016\n",
      "Epoch 1285/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1286/1500\n",
      "30447/30447 [==============================] - 11s 370us/step - loss: 0.0016\n",
      "Epoch 1287/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1288/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1289/1500\n",
      "30447/30447 [==============================] - 11s 370us/step - loss: 0.0016\n",
      "Epoch 1290/1500\n",
      "30447/30447 [==============================] - 11s 367us/step - loss: 0.0016\n",
      "Epoch 1291/1500\n",
      "30447/30447 [==============================] - 11s 370us/step - loss: 0.0016\n",
      "Epoch 1292/1500\n",
      "30447/30447 [==============================] - 11s 367us/step - loss: 0.0016\n",
      "Epoch 1293/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0016\n",
      "Epoch 1294/1500\n",
      "30447/30447 [==============================] - 11s 370us/step - loss: 0.0016\n",
      "Epoch 1295/1500\n",
      "30447/30447 [==============================] - 11s 367us/step - loss: 0.0016\n",
      "Epoch 1296/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1297/1500\n",
      "30447/30447 [==============================] - 11s 370us/step - loss: 0.0016\n",
      "Epoch 1298/1500\n",
      "30447/30447 [==============================] - 11s 367us/step - loss: 0.0016\n",
      "Epoch 1299/1500\n",
      "30447/30447 [==============================] - 11s 371us/step - loss: 0.0016\n",
      "Epoch 1300/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0016\n",
      "Epoch 1301/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1302/1500\n",
      "30447/30447 [==============================] - 11s 370us/step - loss: 0.0016\n",
      "Epoch 1303/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0016\n",
      "Epoch 1304/1500\n",
      "30447/30447 [==============================] - 12s 385us/step - loss: 0.0016\n",
      "Epoch 1305/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0016\n",
      "Epoch 1306/1500\n",
      "30447/30447 [==============================] - 11s 367us/step - loss: 0.0016\n",
      "Epoch 1307/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1308/1500\n",
      "30447/30447 [==============================] - 11s 365us/step - loss: 0.0016\n",
      "Epoch 1309/1500\n",
      "30447/30447 [==============================] - 11s 366us/step - loss: 0.0016\n",
      "Epoch 1310/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1311/1500\n",
      "30447/30447 [==============================] - 11s 365us/step - loss: 0.0016\n",
      "Epoch 1312/1500\n",
      "30447/30447 [==============================] - 11s 366us/step - loss: 0.0016\n",
      "Epoch 1313/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1314/1500\n",
      "30447/30447 [==============================] - 11s 366us/step - loss: 0.0016\n",
      "Epoch 1315/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0016\n",
      "Epoch 1316/1500\n",
      "30447/30447 [==============================] - 11s 365us/step - loss: 0.0016\n",
      "Epoch 1317/1500\n",
      "30447/30447 [==============================] - 11s 364us/step - loss: 0.0016\n",
      "Epoch 1318/1500\n",
      "30447/30447 [==============================] - 13s 419us/step - loss: 0.0016\n",
      "Epoch 1319/1500\n",
      "30447/30447 [==============================] - 13s 435us/step - loss: 0.0016\n",
      "Epoch 1320/1500\n",
      "30447/30447 [==============================] - 13s 424us/step - loss: 0.0016\n",
      "Epoch 1321/1500\n",
      "30447/30447 [==============================] - 13s 427us/step - loss: 0.0016\n",
      "Epoch 1322/1500\n",
      "30447/30447 [==============================] - 14s 449us/step - loss: 0.0016\n",
      "Epoch 1323/1500\n",
      "30447/30447 [==============================] - 17s 542us/step - loss: 0.0016\n",
      "Epoch 1324/1500\n",
      "30447/30447 [==============================] - 13s 443us/step - loss: 0.0016\n",
      "Epoch 1325/1500\n",
      "30447/30447 [==============================] - 13s 438us/step - loss: 0.0016\n",
      "Epoch 1326/1500\n",
      "30447/30447 [==============================] - 12s 390us/step - loss: 0.0016\n",
      "Epoch 1327/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0016\n",
      "Epoch 1328/1500\n",
      "30447/30447 [==============================] - 11s 375us/step - loss: 0.0016\n",
      "Epoch 1329/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1330/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1331/1500\n",
      "30447/30447 [==============================] - 12s 398us/step - loss: 0.0016\n",
      "Epoch 1332/1500\n",
      "30447/30447 [==============================] - 12s 391us/step - loss: 0.0015\n",
      "Epoch 1333/1500\n",
      "30447/30447 [==============================] - 12s 398us/step - loss: 0.0015\n",
      "Epoch 1334/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0015\n",
      "Epoch 1335/1500\n",
      "30447/30447 [==============================] - 11s 373us/step - loss: 0.0015\n",
      "Epoch 1336/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0016\n",
      "Epoch 1337/1500\n",
      "30447/30447 [==============================] - 11s 372us/step - loss: 0.0015\n",
      "Epoch 1338/1500\n",
      "30447/30447 [==============================] - 11s 370us/step - loss: 0.0015\n",
      "Epoch 1339/1500\n",
      "30447/30447 [==============================] - 11s 372us/step - loss: 0.0015\n",
      "Epoch 1340/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0015\n",
      "Epoch 1341/1500\n",
      "30447/30447 [==============================] - 13s 421us/step - loss: 0.0015\n",
      "Epoch 1342/1500\n",
      "30447/30447 [==============================] - 11s 367us/step - loss: 0.0015\n",
      "Epoch 1343/1500\n",
      "30447/30447 [==============================] - 11s 370us/step - loss: 0.0016\n",
      "Epoch 1344/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0016\n",
      "Epoch 1345/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0015\n",
      "Epoch 1346/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0015\n",
      "Epoch 1347/1500\n",
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0015\n",
      "Epoch 1348/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0015\n",
      "Epoch 1349/1500\n",
      "30447/30447 [==============================] - 11s 374us/step - loss: 0.0015\n",
      "Epoch 1350/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0015\n",
      "Epoch 1351/1500\n",
      "30447/30447 [==============================] - 11s 369us/step - loss: 0.0015\n",
      "Epoch 1352/1500\n",
      "30447/30447 [==============================] - 11s 367us/step - loss: 0.0015\n",
      "Epoch 1353/1500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30447/30447 [==============================] - 11s 368us/step - loss: 0.0015\n",
      "Epoch 1354/1500\n",
      "30447/30447 [==============================] - 11s 374us/step - loss: 0.0015\n",
      "Epoch 1355/1500\n",
      "30447/30447 [==============================] - 13s 423us/step - loss: 0.0015\n",
      "Epoch 1356/1500\n",
      "30447/30447 [==============================] - 13s 426us/step - loss: 0.0015\n",
      "Epoch 1357/1500\n",
      "30447/30447 [==============================] - 12s 404us/step - loss: 0.0015\n",
      "Epoch 1358/1500\n",
      "30447/30447 [==============================] - 12s 401us/step - loss: 0.0015\n",
      "Epoch 1359/1500\n",
      "30447/30447 [==============================] - 12s 399us/step - loss: 0.0015\n",
      "Epoch 1360/1500\n",
      "30447/30447 [==============================] - 12s 395us/step - loss: 0.0015\n",
      "Epoch 1361/1500\n",
      "30447/30447 [==============================] - 12s 396us/step - loss: 0.0015\n",
      "Epoch 1362/1500\n",
      "30447/30447 [==============================] - 12s 396us/step - loss: 0.0015\n",
      "Epoch 1363/1500\n",
      "30447/30447 [==============================] - 12s 396us/step - loss: 0.0015\n",
      "Epoch 1364/1500\n",
      "30447/30447 [==============================] - 13s 438us/step - loss: 0.0015\n",
      "Epoch 1365/1500\n",
      "30447/30447 [==============================] - 12s 399us/step - loss: 0.0015\n",
      "Epoch 1366/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1367/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1368/1500\n",
      "30447/30447 [==============================] - 12s 383us/step - loss: 0.0015\n",
      "Epoch 1369/1500\n",
      "30447/30447 [==============================] - 12s 383us/step - loss: 0.0015\n",
      "Epoch 1370/1500\n",
      "30447/30447 [==============================] - 12s 380us/step - loss: 0.0015\n",
      "Epoch 1371/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1372/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1373/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0015\n",
      "Epoch 1374/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1375/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1376/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0015\n",
      "Epoch 1377/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0015\n",
      "Epoch 1378/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0015\n",
      "Epoch 1379/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0015\n",
      "Epoch 1380/1500\n",
      "30447/30447 [==============================] - 12s 385us/step - loss: 0.0015\n",
      "Epoch 1381/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1382/1500\n",
      "30447/30447 [==============================] - 12s 384us/step - loss: 0.0015\n",
      "Epoch 1383/1500\n",
      "30447/30447 [==============================] - 12s 384us/step - loss: 0.0015\n",
      "Epoch 1384/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0015\n",
      "Epoch 1385/1500\n",
      "30447/30447 [==============================] - 12s 383us/step - loss: 0.0015\n",
      "Epoch 1386/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0015\n",
      "Epoch 1387/1500\n",
      "30447/30447 [==============================] - 12s 384us/step - loss: 0.0015\n",
      "Epoch 1388/1500\n",
      "30447/30447 [==============================] - 12s 384us/step - loss: 0.0015\n",
      "Epoch 1389/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0015\n",
      "Epoch 1390/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1391/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1392/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1393/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0015\n",
      "Epoch 1394/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0015\n",
      "Epoch 1395/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1396/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0015\n",
      "Epoch 1397/1500\n",
      "30447/30447 [==============================] - 12s 380us/step - loss: 0.0015\n",
      "Epoch 1398/1500\n",
      "30447/30447 [==============================] - 12s 380us/step - loss: 0.0015\n",
      "Epoch 1399/1500\n",
      "30447/30447 [==============================] - 12s 383us/step - loss: 0.0015\n",
      "Epoch 1400/1500\n",
      "30447/30447 [==============================] - 12s 386us/step - loss: 0.0015\n",
      "Epoch 1401/1500\n",
      "30447/30447 [==============================] - 12s 385us/step - loss: 0.0015\n",
      "Epoch 1402/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1403/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1404/1500\n",
      "30447/30447 [==============================] - 12s 379us/step - loss: 0.0015\n",
      "Epoch 1405/1500\n",
      "30447/30447 [==============================] - 12s 380us/step - loss: 0.0015\n",
      "Epoch 1406/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0015\n",
      "Epoch 1407/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0015\n",
      "Epoch 1408/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1409/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0015\n",
      "Epoch 1410/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1411/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0015\n",
      "Epoch 1412/1500\n",
      "30447/30447 [==============================] - 12s 383us/step - loss: 0.0015\n",
      "Epoch 1413/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0015\n",
      "Epoch 1414/1500\n",
      "30447/30447 [==============================] - 12s 383us/step - loss: 0.0015\n",
      "Epoch 1415/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1416/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1417/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1418/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1419/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1420/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1421/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1422/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1423/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0015\n",
      "Epoch 1424/1500\n",
      "30447/30447 [==============================] - 12s 384us/step - loss: 0.0015\n",
      "Epoch 1425/1500\n",
      "30447/30447 [==============================] - 12s 385us/step - loss: 0.0015\n",
      "Epoch 1426/1500\n",
      "30447/30447 [==============================] - 12s 385us/step - loss: 0.0015\n",
      "Epoch 1427/1500\n",
      "30447/30447 [==============================] - 12s 384us/step - loss: 0.0015\n",
      "Epoch 1428/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0015\n",
      "Epoch 1429/1500\n",
      "30447/30447 [==============================] - 12s 384us/step - loss: 0.0015\n",
      "Epoch 1430/1500\n",
      "30447/30447 [==============================] - 12s 387us/step - loss: 0.0015\n",
      "Epoch 1431/1500\n",
      "30447/30447 [==============================] - 12s 384us/step - loss: 0.0015\n",
      "Epoch 1432/1500\n",
      "30447/30447 [==============================] - 12s 385us/step - loss: 0.0015\n",
      "Epoch 1433/1500\n",
      "30447/30447 [==============================] - 12s 384us/step - loss: 0.0015\n",
      "Epoch 1434/1500\n",
      "30447/30447 [==============================] - 12s 383us/step - loss: 0.0015\n",
      "Epoch 1435/1500\n",
      "30447/30447 [==============================] - 12s 383us/step - loss: 0.0015\n",
      "Epoch 1436/1500\n",
      "30447/30447 [==============================] - 12s 385us/step - loss: 0.0015\n",
      "Epoch 1437/1500\n",
      "30447/30447 [==============================] - 12s 382us/step - loss: 0.0015\n",
      "Epoch 1438/1500\n",
      "30447/30447 [==============================] - 12s 384us/step - loss: 0.0015\n",
      "Epoch 1439/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1440/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1441/1500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30447/30447 [==============================] - 12s 380us/step - loss: 0.0015\n",
      "Epoch 1442/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0015\n",
      "Epoch 1443/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0015\n",
      "Epoch 1444/1500\n",
      "30447/30447 [==============================] - 11s 378us/step - loss: 0.0015\n",
      "Epoch 1445/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0015\n",
      "Epoch 1446/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0015\n",
      "Epoch 1447/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1448/1500\n",
      "30447/30447 [==============================] - 13s 430us/step - loss: 0.0015\n",
      "Epoch 1449/1500\n",
      "30447/30447 [==============================] - 12s 384us/step - loss: 0.0015\n",
      "Epoch 1450/1500\n",
      "30447/30447 [==============================] - 12s 385us/step - loss: 0.0015\n",
      "Epoch 1451/1500\n",
      "30447/30447 [==============================] - 13s 439us/step - loss: 0.0015\n",
      "Epoch 1452/1500\n",
      "30447/30447 [==============================] - 14s 469us/step - loss: 0.0015\n",
      "Epoch 1453/1500\n",
      "30447/30447 [==============================] - 13s 416us/step - loss: 0.0015\n",
      "Epoch 1454/1500\n",
      "30447/30447 [==============================] - 12s 404us/step - loss: 0.0015\n",
      "Epoch 1455/1500\n",
      "30447/30447 [==============================] - 12s 392us/step - loss: 0.0015\n",
      "Epoch 1456/1500\n",
      "30447/30447 [==============================] - 13s 422us/step - loss: 0.0015\n",
      "Epoch 1457/1500\n",
      "30447/30447 [==============================] - 12s 395us/step - loss: 0.0015\n",
      "Epoch 1458/1500\n",
      "30447/30447 [==============================] - 11s 370us/step - loss: 0.0015\n",
      "Epoch 1459/1500\n",
      "30447/30447 [==============================] - 12s 384us/step - loss: 0.0015\n",
      "Epoch 1460/1500\n",
      "30447/30447 [==============================] - 12s 401us/step - loss: 0.0015\n",
      "Epoch 1461/1500\n",
      "30447/30447 [==============================] - 12s 406us/step - loss: 0.0015\n",
      "Epoch 1462/1500\n",
      "30447/30447 [==============================] - 12s 395us/step - loss: 0.0015\n",
      "Epoch 1463/1500\n",
      "30447/30447 [==============================] - 11s 373us/step - loss: 0.0015\n",
      "Epoch 1464/1500\n",
      "30447/30447 [==============================] - 11s 375us/step - loss: 0.0015\n",
      "Epoch 1465/1500\n",
      "30447/30447 [==============================] - 12s 381us/step - loss: 0.0015\n",
      "Epoch 1466/1500\n",
      "30447/30447 [==============================] - 13s 422us/step - loss: 0.0015\n",
      "Epoch 1467/1500\n",
      "30447/30447 [==============================] - 13s 413us/step - loss: 0.0015\n",
      "Epoch 1468/1500\n",
      "30447/30447 [==============================] - 12s 402us/step - loss: 0.0015\n",
      "Epoch 1469/1500\n",
      "30447/30447 [==============================] - 12s 402us/step - loss: 0.0015\n",
      "Epoch 1470/1500\n",
      "30447/30447 [==============================] - 12s 402us/step - loss: 0.0015\n",
      "Epoch 1471/1500\n",
      "30447/30447 [==============================] - 12s 403us/step - loss: 0.0015\n",
      "Epoch 1472/1500\n",
      "30447/30447 [==============================] - 14s 457us/step - loss: 0.0015\n",
      "Epoch 1473/1500\n",
      "30447/30447 [==============================] - 12s 410us/step - loss: 0.0015\n",
      "Epoch 1474/1500\n",
      "30447/30447 [==============================] - 13s 417us/step - loss: 0.0015\n",
      "Epoch 1475/1500\n",
      "30447/30447 [==============================] - 12s 408us/step - loss: 0.0015\n",
      "Epoch 1476/1500\n",
      "30447/30447 [==============================] - 13s 411us/step - loss: 0.0015\n",
      "Epoch 1477/1500\n",
      "30447/30447 [==============================] - 12s 404us/step - loss: 0.0015\n",
      "Epoch 1478/1500\n",
      "30447/30447 [==============================] - 13s 440us/step - loss: 0.0015\n",
      "Epoch 1479/1500\n",
      "30447/30447 [==============================] - 12s 398us/step - loss: 0.0015\n",
      "Epoch 1480/1500\n",
      "30447/30447 [==============================] - 13s 416us/step - loss: 0.0014\n",
      "Epoch 1481/1500\n",
      "30447/30447 [==============================] - 13s 415us/step - loss: 0.0015\n",
      "Epoch 1482/1500\n",
      "30447/30447 [==============================] - 12s 395us/step - loss: 0.0014\n",
      "Epoch 1483/1500\n",
      "30447/30447 [==============================] - 13s 437us/step - loss: 0.0015\n",
      "Epoch 1484/1500\n",
      "30447/30447 [==============================] - 14s 448us/step - loss: 0.0015\n",
      "Epoch 1485/1500\n",
      "30447/30447 [==============================] - 12s 402us/step - loss: 0.0015\n",
      "Epoch 1486/1500\n",
      "30447/30447 [==============================] - 13s 415us/step - loss: 0.0015\n",
      "Epoch 1487/1500\n",
      "30447/30447 [==============================] - 13s 419us/step - loss: 0.0015\n",
      "Epoch 1488/1500\n",
      "30447/30447 [==============================] - 13s 436us/step - loss: 0.0015\n",
      "Epoch 1489/1500\n",
      "30447/30447 [==============================] - 12s 401us/step - loss: 0.0015\n",
      "Epoch 1490/1500\n",
      "30447/30447 [==============================] - 11s 374us/step - loss: 0.0015\n",
      "Epoch 1491/1500\n",
      "30447/30447 [==============================] - 12s 403us/step - loss: 0.0014\n",
      "Epoch 1492/1500\n",
      "30447/30447 [==============================] - 14s 455us/step - loss: 0.0015\n",
      "Epoch 1493/1500\n",
      "30447/30447 [==============================] - 15s 496us/step - loss: 0.0015\n",
      "Epoch 1494/1500\n",
      "30447/30447 [==============================] - 14s 444us/step - loss: 0.0015\n",
      "Epoch 1495/1500\n",
      "30447/30447 [==============================] - 15s 488us/step - loss: 0.0014\n",
      "Epoch 1496/1500\n",
      "30447/30447 [==============================] - 16s 524us/step - loss: 0.0015\n",
      "Epoch 1497/1500\n",
      "30447/30447 [==============================] - 19s 626us/step - loss: 0.0015\n",
      "Epoch 1498/1500\n",
      "30447/30447 [==============================] - 13s 440us/step - loss: 0.0015\n",
      "Epoch 1499/1500\n",
      "30447/30447 [==============================] - 13s 417us/step - loss: 0.0015\n",
      "Epoch 1500/1500\n",
      "30447/30447 [==============================] - 11s 377us/step - loss: 0.0015\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7fc69666f160>"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(X_train_scaled, Y_train_scaled, epochs=1500, shuffle=True, verbose=1, batch_size=32, initial_epoch=1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fc684d8e3c8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# predict \n",
    "Y_pred = model.predict(X_train_scaled, batch_size=10000)\n",
    "\n",
    "# convert targets back to the original scale\n",
    "Y = scaler_target.inverse_transform(Y_pred)\n",
    "\n",
    "k = 0\n",
    "# create an empty image of type unsigned int (8 bits) and same shape as the original\n",
    "im_out = np.empty(shape=im.shape, dtype=np.uint8)\n",
    "for i in range(im.shape[0]):\n",
    "    for j in range(im.shape[1]):\n",
    "        im_out[i,j]= Y[k]\n",
    "        k += 1\n",
    "        \n",
    "plt.imshow(im_out)\n",
    "plt.savefig(\"reconstruction_00015.png\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}