Chapter 21 Object recognition in photographs.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.datasets import cifar10\n",
    "from matplotlib import pyplot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading data from https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz\n",
      "170500096/170498071 [==============================] - 256s 2us/step\n"
     ]
    }
   ],
   "source": [
    "(X_train,y_train), (X_test,y_test) = cifar10.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for i in range(0,9):\n",
    "    pyplot.subplot(330+i+1)\n",
    "    pyplot.imshow(X_train[i])\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Simple CNN for CIFAR-10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy\n",
    "from keras.datasets import cifar10\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.layers import Dropout\n",
    "from keras.layers import Flatten\n",
    "from keras.constraints import maxnorm\n",
    "from keras.optimizers import SGD\n",
    "from keras.layers.convolutional import Convolution2D\n",
    "from keras.layers.convolutional import MaxPooling2D\n",
    "from keras.utils import np_utils\n",
    "from keras import backend as K"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "seed = 7\n",
    "numpy.random.seed(seed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "(X_train,y_train),(X_test,y_test) = cifar10.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = X_train.astype('float32')\n",
    "X_test = X_test.astype('float32')\n",
    "X_train = X_train/255.0\n",
    "X_test = X_test/255.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train = np_utils.to_categorical(y_train)\n",
    "y_test = np_utils.to_categorical(y_test)\n",
    "num_classes = y_test.shape[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/arvind/.local/lib/python3.6/site-packages/ipykernel_launcher.py:2: UserWarning: Update your `Conv2D` call to the Keras 2 API: `Conv2D(32, (3, 3), input_shape=(32, 32, 3..., padding=\"same\", activation=\"relu\", kernel_constraint=<keras.con...)`\n",
      "  \n",
      "/home/arvind/.local/lib/python3.6/site-packages/ipykernel_launcher.py:4: UserWarning: Update your `Conv2D` call to the Keras 2 API: `Conv2D(32, (3, 3), activation=\"relu\", padding=\"same\", kernel_constraint=<keras.con...)`\n",
      "  after removing the cwd from sys.path.\n",
      "/home/arvind/.local/lib/python3.6/site-packages/ipykernel_launcher.py:7: UserWarning: Update your `Dense` call to the Keras 2 API: `Dense(512, activation=\"relu\", kernel_constraint=<keras.con...)`\n",
      "  import sys\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "model.add(Convolution2D(32,(3,3),input_shape = (32,32,3),padding = \"same\", activation = 'relu',W_constraint = maxnorm(3)))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(Convolution2D(32,(3,3),activation=\"relu\",padding='same',W_constraint=maxnorm(3)))\n",
    "model.add(MaxPooling2D(pool_size=(2,2)))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(512,activation='relu',W_constraint=maxnorm(3)))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(num_classes, activation=\"softmax\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs = 25\n",
    "lrate=0.01\n",
    "decay = lrate/epochs\n",
    "sgd = SGD(lr=lrate,momentum=0.9,decay=decay,nesterov=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(loss=\"categorical_crossentropy\",optimizer=sgd,metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_5\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_9 (Conv2D)            (None, 32, 32, 32)        896       \n",
      "_________________________________________________________________\n",
      "dropout_8 (Dropout)          (None, 32, 32, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_10 (Conv2D)           (None, 32, 32, 32)        9248      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_5 (MaxPooling2 (None, 16, 16, 32)        0         \n",
      "_________________________________________________________________\n",
      "flatten_5 (Flatten)          (None, 8192)              0         \n",
      "_________________________________________________________________\n",
      "dense_7 (Dense)              (None, 512)               4194816   \n",
      "_________________________________________________________________\n",
      "dropout_9 (Dropout)          (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dense_8 (Dense)              (None, 10)                5130      \n",
      "=================================================================\n",
      "Total params: 4,210,090\n",
      "Trainable params: 4,210,090\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 50000 samples, validate on 10000 samples\n",
      "Epoch 1/25\n",
      " - 150s - loss: 1.7327 - accuracy: 0.3715 - val_loss: 1.4096 - val_accuracy: 0.4941\n",
      "Epoch 2/25\n",
      " - 151s - loss: 1.3530 - accuracy: 0.5136 - val_loss: 1.1812 - val_accuracy: 0.5774\n",
      "Epoch 3/25\n",
      " - 162s - loss: 1.1798 - accuracy: 0.5798 - val_loss: 1.1036 - val_accuracy: 0.6135\n",
      "Epoch 4/25\n",
      " - 181s - loss: 1.0561 - accuracy: 0.6251 - val_loss: 1.0653 - val_accuracy: 0.6251\n",
      "Epoch 5/25\n",
      " - 158s - loss: 0.9603 - accuracy: 0.6584 - val_loss: 0.9911 - val_accuracy: 0.6491\n",
      "Epoch 6/25\n",
      " - 147s - loss: 0.8799 - accuracy: 0.6851 - val_loss: 0.9694 - val_accuracy: 0.6586\n",
      "Epoch 7/25\n",
      " - 144s - loss: 0.8066 - accuracy: 0.7143 - val_loss: 0.9350 - val_accuracy: 0.6732\n",
      "Epoch 8/25\n",
      " - 145s - loss: 0.7436 - accuracy: 0.7359 - val_loss: 0.9218 - val_accuracy: 0.6841\n",
      "Epoch 9/25\n",
      " - 144s - loss: 0.6839 - accuracy: 0.7575 - val_loss: 0.9177 - val_accuracy: 0.6881\n",
      "Epoch 10/25\n",
      " - 144s - loss: 0.6377 - accuracy: 0.7755 - val_loss: 0.9174 - val_accuracy: 0.6925\n",
      "Epoch 11/25\n",
      " - 149s - loss: 0.5905 - accuracy: 0.7913 - val_loss: 0.9232 - val_accuracy: 0.6915\n",
      "Epoch 12/25\n",
      " - 145s - loss: 0.5441 - accuracy: 0.8076 - val_loss: 0.9208 - val_accuracy: 0.6954\n",
      "Epoch 13/25\n",
      " - 145s - loss: 0.5088 - accuracy: 0.8187 - val_loss: 0.9410 - val_accuracy: 0.6964\n",
      "Epoch 14/25\n",
      " - 146s - loss: 0.4756 - accuracy: 0.8299 - val_loss: 0.9299 - val_accuracy: 0.6999\n",
      "Epoch 15/25\n",
      " - 145s - loss: 0.4451 - accuracy: 0.8431 - val_loss: 0.9457 - val_accuracy: 0.6986\n",
      "Epoch 16/25\n",
      " - 154s - loss: 0.4185 - accuracy: 0.8507 - val_loss: 0.9583 - val_accuracy: 0.6982\n",
      "Epoch 17/25\n",
      " - 162s - loss: 0.3893 - accuracy: 0.8616 - val_loss: 0.9762 - val_accuracy: 0.6955\n",
      "Epoch 18/25\n",
      " - 149s - loss: 0.3709 - accuracy: 0.8692 - val_loss: 0.9810 - val_accuracy: 0.7007\n",
      "Epoch 19/25\n",
      " - 170s - loss: 0.3475 - accuracy: 0.8756 - val_loss: 0.9887 - val_accuracy: 0.7019\n",
      "Epoch 20/25\n",
      " - 152s - loss: 0.3339 - accuracy: 0.8835 - val_loss: 0.9925 - val_accuracy: 0.7000\n",
      "Epoch 21/25\n",
      " - 147s - loss: 0.3108 - accuracy: 0.8903 - val_loss: 1.0109 - val_accuracy: 0.7044\n",
      "Epoch 22/25\n",
      " - 146s - loss: 0.3001 - accuracy: 0.8952 - val_loss: 1.0139 - val_accuracy: 0.7049\n",
      "Epoch 23/25\n",
      " - 145s - loss: 0.2838 - accuracy: 0.9023 - val_loss: 1.0365 - val_accuracy: 0.7053\n",
      "Epoch 24/25\n",
      " - 146s - loss: 0.2634 - accuracy: 0.9094 - val_loss: 1.0638 - val_accuracy: 0.7051\n",
      "Epoch 25/25\n",
      " - 155s - loss: 0.2574 - accuracy: 0.9101 - val_loss: 1.0714 - val_accuracy: 0.7056\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.callbacks.History at 0x7f3a440d7240>"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(X_train,y_train,validation_data=(X_test,y_test),epochs=epochs,batch_size=32,verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "scores = model.evaluate(X_test,y_test,verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 70.56%\n"
     ]
    }
   ],
   "source": [
    "print(\"Accuracy: %.2f%%\"%(scores[1]*100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Larger CNN for CIFAR-10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_6\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_11 (Conv2D)           (None, 32, 32, 32)        896       \n",
      "_________________________________________________________________\n",
      "dropout_10 (Dropout)         (None, 32, 32, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_12 (Conv2D)           (None, 32, 32, 32)        9248      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_6 (MaxPooling2 (None, 16, 16, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_13 (Conv2D)           (None, 16, 16, 64)        18496     \n",
      "_________________________________________________________________\n",
      "dropout_11 (Dropout)         (None, 16, 16, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_14 (Conv2D)           (None, 16, 16, 64)        36928     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_7 (MaxPooling2 (None, 8, 8, 64)          0         \n",
      "_________________________________________________________________\n",
      "conv2d_15 (Conv2D)           (None, 8, 8, 128)         73856     \n",
      "_________________________________________________________________\n",
      "dropout_12 (Dropout)         (None, 8, 8, 128)         0         \n",
      "_________________________________________________________________\n",
      "conv2d_16 (Conv2D)           (None, 8, 8, 128)         147584    \n",
      "_________________________________________________________________\n",
      "max_pooling2d_8 (MaxPooling2 (None, 4, 4, 128)         0         \n",
      "_________________________________________________________________\n",
      "flatten_6 (Flatten)          (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dropout_13 (Dropout)         (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dense_9 (Dense)              (None, 1024)              2098176   \n",
      "_________________________________________________________________\n",
      "dropout_14 (Dropout)         (None, 1024)              0         \n",
      "_________________________________________________________________\n",
      "dense_10 (Dense)             (None, 512)               524800    \n",
      "_________________________________________________________________\n",
      "dropout_15 (Dropout)         (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dense_11 (Dense)             (None, 10)                5130      \n",
      "=================================================================\n",
      "Total params: 2,915,114\n",
      "Trainable params: 2,915,114\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/arvind/.local/lib/python3.6/site-packages/ipykernel_launcher.py:16: UserWarning: Update your `Dense` call to the Keras 2 API: `Dense(1024, activation=\"relu\", kernel_constraint=<keras.con...)`\n",
      "  app.launch_new_instance()\n",
      "/home/arvind/.local/lib/python3.6/site-packages/ipykernel_launcher.py:18: UserWarning: Update your `Dense` call to the Keras 2 API: `Dense(512, activation=\"relu\", kernel_constraint=<keras.con...)`\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "model.add(Convolution2D(32,(3,3),input_shape= (32,32,3),activation=\"relu\",padding = \"same\"))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(Convolution2D(32,(3,3),activation=\"relu\",padding=\"same\"))\n",
    "model.add(MaxPooling2D(pool_size=(2,2)))\n",
    "model.add(Convolution2D(64,(3,3),activation=\"relu\",padding=\"same\"))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(Convolution2D(64,(3,3),activation=\"relu\",padding=\"same\"))\n",
    "model.add(MaxPooling2D(pool_size=(2,2)))\n",
    "model.add(Convolution2D(128,(3,3),activation=\"relu\",padding=\"same\"))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(Convolution2D(128,(3,3),activation=\"relu\",padding=\"same\"))\n",
    "model.add(MaxPooling2D(pool_size=(2,2)))\n",
    "model.add(Flatten())\n",
    "model.add(Dropout(0.2))\n",
    "model.add(Dense(1024,activation=\"relu\",W_constraint=maxnorm(3)))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(Dense(512,activation=\"relu\",W_constraint=maxnorm(3)))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(Dense(num_classes,activation=\"softmax\"))\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs = 25\n",
    "lrate=0.01\n",
    "decay=lrate/epochs\n",
    "sgd = SGD(lr=lrate, momentum=0.9,decay=decay,nesterov=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(loss=\"categorical_crossentropy\",optimizer=sgd,metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_6\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_11 (Conv2D)           (None, 32, 32, 32)        896       \n",
      "_________________________________________________________________\n",
      "dropout_10 (Dropout)         (None, 32, 32, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_12 (Conv2D)           (None, 32, 32, 32)        9248      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_6 (MaxPooling2 (None, 16, 16, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_13 (Conv2D)           (None, 16, 16, 64)        18496     \n",
      "_________________________________________________________________\n",
      "dropout_11 (Dropout)         (None, 16, 16, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_14 (Conv2D)           (None, 16, 16, 64)        36928     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_7 (MaxPooling2 (None, 8, 8, 64)          0         \n",
      "_________________________________________________________________\n",
      "conv2d_15 (Conv2D)           (None, 8, 8, 128)         73856     \n",
      "_________________________________________________________________\n",
      "dropout_12 (Dropout)         (None, 8, 8, 128)         0         \n",
      "_________________________________________________________________\n",
      "conv2d_16 (Conv2D)           (None, 8, 8, 128)         147584    \n",
      "_________________________________________________________________\n",
      "max_pooling2d_8 (MaxPooling2 (None, 4, 4, 128)         0         \n",
      "_________________________________________________________________\n",
      "flatten_6 (Flatten)          (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dropout_13 (Dropout)         (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dense_9 (Dense)              (None, 1024)              2098176   \n",
      "_________________________________________________________________\n",
      "dropout_14 (Dropout)         (None, 1024)              0         \n",
      "_________________________________________________________________\n",
      "dense_10 (Dense)             (None, 512)               524800    \n",
      "_________________________________________________________________\n",
      "dropout_15 (Dropout)         (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dense_11 (Dense)             (None, 10)                5130      \n",
      "=================================================================\n",
      "Total params: 2,915,114\n",
      "Trainable params: 2,915,114\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 50000 samples, validate on 10000 samples\n",
      "Epoch 1/25\n",
      "50000/50000 [==============================] - 238s 5ms/step - loss: 1.9247 - accuracy: 0.2882 - val_loss: 1.6916 - val_accuracy: 0.3990\n",
      "Epoch 2/25\n",
      "50000/50000 [==============================] - 236s 5ms/step - loss: 1.4934 - accuracy: 0.4553 - val_loss: 1.4139 - val_accuracy: 0.4862\n",
      "Epoch 3/25\n",
      "50000/50000 [==============================] - 238s 5ms/step - loss: 1.3198 - accuracy: 0.5223 - val_loss: 1.1991 - val_accuracy: 0.5702\n",
      "Epoch 4/25\n",
      "50000/50000 [==============================] - 254s 5ms/step - loss: 1.1867 - accuracy: 0.5729 - val_loss: 1.0903 - val_accuracy: 0.6057\n",
      "Epoch 5/25\n",
      "50000/50000 [==============================] - 291s 6ms/step - loss: 1.0713 - accuracy: 0.6155 - val_loss: 1.0617 - val_accuracy: 0.6197\n",
      "Epoch 6/25\n",
      "50000/50000 [==============================] - 236s 5ms/step - loss: 0.9789 - accuracy: 0.6520 - val_loss: 0.9587 - val_accuracy: 0.6627\n",
      "Epoch 7/25\n",
      "50000/50000 [==============================] - 262s 5ms/step - loss: 0.9082 - accuracy: 0.6774 - val_loss: 0.8649 - val_accuracy: 0.6937\n",
      "Epoch 8/25\n",
      "50000/50000 [==============================] - 266s 5ms/step - loss: 0.8513 - accuracy: 0.6970 - val_loss: 0.8497 - val_accuracy: 0.6980\n",
      "Epoch 9/25\n",
      "50000/50000 [==============================] - 269s 5ms/step - loss: 0.7917 - accuracy: 0.7209 - val_loss: 0.7817 - val_accuracy: 0.7266\n",
      "Epoch 10/25\n",
      "50000/50000 [==============================] - 251s 5ms/step - loss: 0.7526 - accuracy: 0.7334 - val_loss: 0.7726 - val_accuracy: 0.7296\n",
      "Epoch 11/25\n",
      "50000/50000 [==============================] - 249s 5ms/step - loss: 0.7088 - accuracy: 0.7497 - val_loss: 0.7411 - val_accuracy: 0.7464\n",
      "Epoch 12/25\n",
      "50000/50000 [==============================] - 239s 5ms/step - loss: 0.6771 - accuracy: 0.7604 - val_loss: 0.7117 - val_accuracy: 0.7525\n",
      "Epoch 13/25\n",
      "50000/50000 [==============================] - 235s 5ms/step - loss: 0.6443 - accuracy: 0.7730 - val_loss: 0.6980 - val_accuracy: 0.7586\n",
      "Epoch 14/25\n",
      "50000/50000 [==============================] - 235s 5ms/step - loss: 0.6185 - accuracy: 0.7807 - val_loss: 0.6871 - val_accuracy: 0.7633\n",
      "Epoch 15/25\n",
      "50000/50000 [==============================] - 235s 5ms/step - loss: 0.5901 - accuracy: 0.7904 - val_loss: 0.6835 - val_accuracy: 0.7659\n",
      "Epoch 16/25\n",
      "50000/50000 [==============================] - 259s 5ms/step - loss: 0.5703 - accuracy: 0.7974 - val_loss: 0.6780 - val_accuracy: 0.7624\n",
      "Epoch 17/25\n",
      "50000/50000 [==============================] - 252s 5ms/step - loss: 0.5505 - accuracy: 0.8043 - val_loss: 0.6571 - val_accuracy: 0.7747\n",
      "Epoch 18/25\n",
      "50000/50000 [==============================] - 234s 5ms/step - loss: 0.5293 - accuracy: 0.8122 - val_loss: 0.6554 - val_accuracy: 0.7757\n",
      "Epoch 19/25\n",
      "50000/50000 [==============================] - 263s 5ms/step - loss: 0.5132 - accuracy: 0.8182 - val_loss: 0.6480 - val_accuracy: 0.7776\n",
      "Epoch 20/25\n",
      "50000/50000 [==============================] - 245s 5ms/step - loss: 0.4900 - accuracy: 0.8256 - val_loss: 0.6385 - val_accuracy: 0.7809\n",
      "Epoch 21/25\n",
      "50000/50000 [==============================] - 244s 5ms/step - loss: 0.4762 - accuracy: 0.8301 - val_loss: 0.6451 - val_accuracy: 0.7795\n",
      "Epoch 22/25\n",
      "50000/50000 [==============================] - 235s 5ms/step - loss: 0.4587 - accuracy: 0.8371 - val_loss: 0.6520 - val_accuracy: 0.7812\n",
      "Epoch 23/25\n",
      "50000/50000 [==============================] - 238s 5ms/step - loss: 0.4432 - accuracy: 0.8413 - val_loss: 0.6333 - val_accuracy: 0.7825\n",
      "Epoch 24/25\n",
      "50000/50000 [==============================] - 237s 5ms/step - loss: 0.4339 - accuracy: 0.8448 - val_loss: 0.6402 - val_accuracy: 0.7861\n",
      "Epoch 25/25\n",
      "50000/50000 [==============================] - 241s 5ms/step - loss: 0.4183 - accuracy: 0.8506 - val_loss: 0.6382 - val_accuracy: 0.7831\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.callbacks.History at 0x7f3a1f176f98>"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(X_train,y_train,validation_data = (X_test,y_test),epochs = epochs, batch_size=64)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10000/10000 [==============================] - 10s 1ms/step\n"
     ]
    }
   ],
   "source": [
    "scores = model.evaluate(X_test,y_test,verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 78.31%\n"
     ]
    }
   ],
   "source": [
    "print(\"Accuracy: %.2f%%\"%(scores[1]*100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}