Convnet Pretrained.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Using a pre-trained convnet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'2.2.4'"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import keras\n",
    "keras.__version__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Feature extraction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.applications import VGG16\n",
    "\n",
    "conv_base = VGG16(weights='imagenet',\n",
    "                  include_top=False,\n",
    "                  input_shape=(150, 150, 3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         (None, 150, 150, 3)       0         \n",
      "_________________________________________________________________\n",
      "block1_conv1 (Conv2D)        (None, 150, 150, 64)      1792      \n",
      "_________________________________________________________________\n",
      "block1_conv2 (Conv2D)        (None, 150, 150, 64)      36928     \n",
      "_________________________________________________________________\n",
      "block1_pool (MaxPooling2D)   (None, 75, 75, 64)        0         \n",
      "_________________________________________________________________\n",
      "block2_conv1 (Conv2D)        (None, 75, 75, 128)       73856     \n",
      "_________________________________________________________________\n",
      "block2_conv2 (Conv2D)        (None, 75, 75, 128)       147584    \n",
      "_________________________________________________________________\n",
      "block2_pool (MaxPooling2D)   (None, 37, 37, 128)       0         \n",
      "_________________________________________________________________\n",
      "block3_conv1 (Conv2D)        (None, 37, 37, 256)       295168    \n",
      "_________________________________________________________________\n",
      "block3_conv2 (Conv2D)        (None, 37, 37, 256)       590080    \n",
      "_________________________________________________________________\n",
      "block3_conv3 (Conv2D)        (None, 37, 37, 256)       590080    \n",
      "_________________________________________________________________\n",
      "block3_pool (MaxPooling2D)   (None, 18, 18, 256)       0         \n",
      "_________________________________________________________________\n",
      "block4_conv1 (Conv2D)        (None, 18, 18, 512)       1180160   \n",
      "_________________________________________________________________\n",
      "block4_conv2 (Conv2D)        (None, 18, 18, 512)       2359808   \n",
      "_________________________________________________________________\n",
      "block4_conv3 (Conv2D)        (None, 18, 18, 512)       2359808   \n",
      "_________________________________________________________________\n",
      "block4_pool (MaxPooling2D)   (None, 9, 9, 512)         0         \n",
      "_________________________________________________________________\n",
      "block5_conv1 (Conv2D)        (None, 9, 9, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv2 (Conv2D)        (None, 9, 9, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv3 (Conv2D)        (None, 9, 9, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_pool (MaxPooling2D)   (None, 4, 4, 512)         0         \n",
      "=================================================================\n",
      "Total params: 14,714,688\n",
      "Trainable params: 14,714,688\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "conv_base.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import numpy as np\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "\n",
    "base_dir = '/Users/GAO/Downloads/cats_and_dogs_small'\n",
    "\n",
    "train_dir = os.path.join(base_dir, 'train')\n",
    "validation_dir = os.path.join(base_dir, 'validation')\n",
    "test_dir = os.path.join(base_dir, 'test')\n",
    "\n",
    "datagen = ImageDataGenerator(rescale=1./255)\n",
    "batch_size = 20"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "def extract_features(directory, sample_count):\n",
    "    features = np.zeros(shape=(sample_count, 4, 4, 512))\n",
    "    labels = np.zeros(shape=(sample_count))\n",
    "    generator = datagen.flow_from_directory(\n",
    "        directory,\n",
    "        target_size=(150, 150),\n",
    "        batch_size=batch_size,\n",
    "        class_mode='binary')\n",
    "    i = 0\n",
    "    for inputs_batch, labels_batch in generator:\n",
    "        features_batch = conv_base.predict(inputs_batch)\n",
    "        features[i * batch_size : (i + 1) * batch_size] = features_batch\n",
    "        labels[i * batch_size : (i + 1) * batch_size] = labels_batch\n",
    "        i += 1\n",
    "        if i * batch_size >= sample_count:\n",
    "            # Note that since generators yield data indefinitely in a loop,\n",
    "            # we must `break` after every image has been seen once.\n",
    "            break\n",
    "    return features, labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 2000 images belonging to 2 classes.\n",
      "Found 1000 images belonging to 2 classes.\n",
      "Found 1000 images belonging to 2 classes.\n"
     ]
    }
   ],
   "source": [
    "train_features, train_labels = extract_features(train_dir, 2000)\n",
    "validation_features, validation_labels = extract_features(validation_dir, 1000)\n",
    "test_features, test_labels = extract_features(test_dir, 1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_features = np.reshape(train_features, (2000, 4 * 4 * 512))\n",
    "validation_features = np.reshape(validation_features, (1000, 4 * 4 * 512))\n",
    "test_features = np.reshape(test_features, (1000, 4 * 4 * 512))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 2000 samples, validate on 1000 samples\n",
      "Epoch 1/30\n",
      "2000/2000 [==============================] - 2s 1ms/step - loss: 0.6156 - acc: 0.6585 - val_loss: 0.4562 - val_acc: 0.8090\n",
      "Epoch 2/30\n",
      "2000/2000 [==============================] - 1s 553us/step - loss: 0.4324 - acc: 0.8115 - val_loss: 0.3651 - val_acc: 0.8630\n",
      "Epoch 3/30\n",
      "2000/2000 [==============================] - 1s 563us/step - loss: 0.3543 - acc: 0.8495 - val_loss: 0.3255 - val_acc: 0.8760\n",
      "Epoch 4/30\n",
      "2000/2000 [==============================] - 1s 556us/step - loss: 0.3195 - acc: 0.8720 - val_loss: 0.3062 - val_acc: 0.8720\n",
      "Epoch 5/30\n",
      "2000/2000 [==============================] - 1s 561us/step - loss: 0.2803 - acc: 0.8840 - val_loss: 0.2921 - val_acc: 0.8810\n",
      "Epoch 6/30\n",
      "2000/2000 [==============================] - 1s 562us/step - loss: 0.2659 - acc: 0.8950 - val_loss: 0.2781 - val_acc: 0.8850\n",
      "Epoch 7/30\n",
      "2000/2000 [==============================] - 1s 568us/step - loss: 0.2423 - acc: 0.9070 - val_loss: 0.2662 - val_acc: 0.8910\n",
      "Epoch 8/30\n",
      "2000/2000 [==============================] - 1s 555us/step - loss: 0.2366 - acc: 0.9080 - val_loss: 0.2627 - val_acc: 0.8920\n",
      "Epoch 9/30\n",
      "2000/2000 [==============================] - 1s 559us/step - loss: 0.2178 - acc: 0.9190 - val_loss: 0.2591 - val_acc: 0.8880\n",
      "Epoch 10/30\n",
      "2000/2000 [==============================] - 1s 573us/step - loss: 0.2014 - acc: 0.9220 - val_loss: 0.2566 - val_acc: 0.8890\n",
      "Epoch 11/30\n",
      "2000/2000 [==============================] - 1s 570us/step - loss: 0.1942 - acc: 0.9275 - val_loss: 0.2509 - val_acc: 0.8910\n",
      "Epoch 12/30\n",
      "2000/2000 [==============================] - 1s 585us/step - loss: 0.1834 - acc: 0.9350 - val_loss: 0.2513 - val_acc: 0.8910\n",
      "Epoch 13/30\n",
      "2000/2000 [==============================] - 1s 574us/step - loss: 0.1796 - acc: 0.9360 - val_loss: 0.2453 - val_acc: 0.9010\n",
      "Epoch 14/30\n",
      "2000/2000 [==============================] - 1s 574us/step - loss: 0.1700 - acc: 0.9385 - val_loss: 0.2429 - val_acc: 0.9000\n",
      "Epoch 15/30\n",
      "2000/2000 [==============================] - 1s 551us/step - loss: 0.1594 - acc: 0.9450 - val_loss: 0.2443 - val_acc: 0.9010\n",
      "Epoch 16/30\n",
      "2000/2000 [==============================] - 1s 562us/step - loss: 0.1522 - acc: 0.9460 - val_loss: 0.2440 - val_acc: 0.9020\n",
      "Epoch 17/30\n",
      "2000/2000 [==============================] - 1s 552us/step - loss: 0.1495 - acc: 0.9505 - val_loss: 0.2401 - val_acc: 0.9000\n",
      "Epoch 18/30\n",
      "2000/2000 [==============================] - 1s 565us/step - loss: 0.1392 - acc: 0.9490 - val_loss: 0.2391 - val_acc: 0.8970\n",
      "Epoch 19/30\n",
      "2000/2000 [==============================] - 1s 559us/step - loss: 0.1336 - acc: 0.9575 - val_loss: 0.2406 - val_acc: 0.9000\n",
      "Epoch 20/30\n",
      "2000/2000 [==============================] - 1s 571us/step - loss: 0.1296 - acc: 0.9585 - val_loss: 0.2386 - val_acc: 0.9000\n",
      "Epoch 21/30\n",
      "2000/2000 [==============================] - 1s 593us/step - loss: 0.1244 - acc: 0.9580 - val_loss: 0.2388 - val_acc: 0.9020\n",
      "Epoch 22/30\n",
      "2000/2000 [==============================] - 1s 587us/step - loss: 0.1188 - acc: 0.9600 - val_loss: 0.2457 - val_acc: 0.9060\n",
      "Epoch 23/30\n",
      "2000/2000 [==============================] - 1s 563us/step - loss: 0.1156 - acc: 0.9590 - val_loss: 0.2413 - val_acc: 0.9000\n",
      "Epoch 24/30\n",
      "2000/2000 [==============================] - 1s 595us/step - loss: 0.1097 - acc: 0.9630 - val_loss: 0.2519 - val_acc: 0.8980\n",
      "Epoch 25/30\n",
      "2000/2000 [==============================] - 1s 584us/step - loss: 0.1090 - acc: 0.9615 - val_loss: 0.2457 - val_acc: 0.9040\n",
      "Epoch 26/30\n",
      "2000/2000 [==============================] - 1s 596us/step - loss: 0.1012 - acc: 0.9690 - val_loss: 0.2387 - val_acc: 0.9010\n",
      "Epoch 27/30\n",
      "2000/2000 [==============================] - 1s 586us/step - loss: 0.0967 - acc: 0.9680 - val_loss: 0.2419 - val_acc: 0.9000\n",
      "Epoch 28/30\n",
      "2000/2000 [==============================] - 1s 592us/step - loss: 0.0922 - acc: 0.9695 - val_loss: 0.2403 - val_acc: 0.9000\n",
      "Epoch 29/30\n",
      "2000/2000 [==============================] - 1s 597us/step - loss: 0.0914 - acc: 0.9690 - val_loss: 0.2428 - val_acc: 0.8990\n",
      "Epoch 30/30\n",
      "2000/2000 [==============================] - 1s 593us/step - loss: 0.0877 - acc: 0.9720 - val_loss: 0.2582 - val_acc: 0.8960\n"
     ]
    }
   ],
   "source": [
    "from keras import models\n",
    "from keras import layers\n",
    "from keras import optimizers\n",
    "\n",
    "model = models.Sequential()\n",
    "model.add(layers.Dense(256, activation='relu', input_dim=4 * 4 * 512))\n",
    "model.add(layers.Dropout(0.5))\n",
    "model.add(layers.Dense(1, activation='sigmoid'))\n",
    "\n",
    "model.compile(optimizer=optimizers.RMSprop(lr=2e-5),\n",
    "              loss='binary_crossentropy',\n",
    "              metrics=['acc'])\n",
    "\n",
    "history = model.fit(train_features, train_labels,\n",
    "                    epochs=30,\n",
    "                    batch_size=20,\n",
    "                    validation_data=(validation_features, validation_labels))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "acc = history.history['acc']\n",
    "val_acc = history.history['val_acc']\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "epochs = range(len(acc))\n",
    "\n",
    "plt.plot(epochs, acc, 'bo', label='Training acc')\n",
    "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
    "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras import models\n",
    "from keras import layers\n",
    "\n",
    "model = models.Sequential()\n",
    "model.add(conv_base)\n",
    "model.add(layers.Flatten())\n",
    "model.add(layers.Dense(256, activation='relu'))\n",
    "model.add(layers.Dense(1, activation='sigmoid'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "vgg16 (Model)                (None, 4, 4, 512)         14714688  \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 8192)              0         \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 256)               2097408   \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 1)                 257       \n",
      "=================================================================\n",
      "Total params: 16,812,353\n",
      "Trainable params: 16,812,353\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "This is the number of trainable weights before freezing the conv base: 30\n"
     ]
    }
   ],
   "source": [
    "print('This is the number of trainable weights '\n",
    "      'before freezing the conv base:', len(model.trainable_weights))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "conv_base.trainable = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "This is the number of trainable weights after freezing the conv base: 4\n"
     ]
    }
   ],
   "source": [
    "print('This is the number of trainable weights '\n",
    "      'after freezing the conv base:', len(model.trainable_weights))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 2000 images belonging to 2 classes.\n",
      "Found 1000 images belonging to 2 classes.\n",
      "Epoch 1/30\n",
      " - 34s - loss: 0.5814 - acc: 0.6985 - val_loss: 0.4386 - val_acc: 0.8360\n",
      "Epoch 2/30\n",
      " - 34s - loss: 0.4744 - acc: 0.7840 - val_loss: 0.3606 - val_acc: 0.8600\n",
      "Epoch 3/30\n",
      " - 34s - loss: 0.4362 - acc: 0.8010 - val_loss: 0.3287 - val_acc: 0.8650\n",
      "Epoch 4/30\n",
      " - 34s - loss: 0.4012 - acc: 0.8280 - val_loss: 0.3073 - val_acc: 0.8750\n",
      "Epoch 5/30\n",
      " - 34s - loss: 0.3854 - acc: 0.8295 - val_loss: 0.3023 - val_acc: 0.8740\n",
      "Epoch 6/30\n",
      " - 34s - loss: 0.3661 - acc: 0.8415 - val_loss: 0.2787 - val_acc: 0.8870\n",
      "Epoch 7/30\n",
      " - 34s - loss: 0.3614 - acc: 0.8415 - val_loss: 0.2926 - val_acc: 0.8730\n",
      "Epoch 8/30\n",
      " - 34s - loss: 0.3499 - acc: 0.8350 - val_loss: 0.2760 - val_acc: 0.8890\n",
      "Epoch 9/30\n",
      " - 34s - loss: 0.3522 - acc: 0.8420 - val_loss: 0.2633 - val_acc: 0.8920\n",
      "Epoch 10/30\n",
      " - 34s - loss: 0.3342 - acc: 0.8595 - val_loss: 0.2549 - val_acc: 0.9010\n",
      "Epoch 11/30\n",
      " - 34s - loss: 0.3291 - acc: 0.8600 - val_loss: 0.2551 - val_acc: 0.8970\n",
      "Epoch 12/30\n",
      " - 34s - loss: 0.3169 - acc: 0.8610 - val_loss: 0.2535 - val_acc: 0.9010\n",
      "Epoch 13/30\n",
      " - 34s - loss: 0.3060 - acc: 0.8705 - val_loss: 0.2705 - val_acc: 0.8870\n",
      "Epoch 14/30\n",
      " - 34s - loss: 0.3149 - acc: 0.8570 - val_loss: 0.2497 - val_acc: 0.8970\n",
      "Epoch 15/30\n",
      " - 34s - loss: 0.3098 - acc: 0.8595 - val_loss: 0.2575 - val_acc: 0.8980\n",
      "Epoch 16/30\n",
      " - 34s - loss: 0.2969 - acc: 0.8690 - val_loss: 0.2652 - val_acc: 0.8950\n",
      "Epoch 17/30\n",
      " - 34s - loss: 0.3076 - acc: 0.8610 - val_loss: 0.2723 - val_acc: 0.8820\n",
      "Epoch 18/30\n",
      " - 34s - loss: 0.3039 - acc: 0.8680 - val_loss: 0.2433 - val_acc: 0.9010\n",
      "Epoch 19/30\n",
      " - 34s - loss: 0.2981 - acc: 0.8710 - val_loss: 0.2411 - val_acc: 0.8950\n",
      "Epoch 20/30\n",
      " - 34s - loss: 0.2991 - acc: 0.8640 - val_loss: 0.2409 - val_acc: 0.8950\n",
      "Epoch 21/30\n",
      " - 34s - loss: 0.3070 - acc: 0.8655 - val_loss: 0.2445 - val_acc: 0.8930\n",
      "Epoch 22/30\n",
      " - 34s - loss: 0.2949 - acc: 0.8730 - val_loss: 0.2416 - val_acc: 0.8950\n",
      "Epoch 23/30\n",
      " - 34s - loss: 0.3003 - acc: 0.8680 - val_loss: 0.2426 - val_acc: 0.8980\n",
      "Epoch 24/30\n",
      " - 34s - loss: 0.2980 - acc: 0.8675 - val_loss: 0.2398 - val_acc: 0.9050\n",
      "Epoch 25/30\n",
      " - 34s - loss: 0.2872 - acc: 0.8800 - val_loss: 0.2427 - val_acc: 0.9000\n",
      "Epoch 26/30\n",
      " - 34s - loss: 0.3052 - acc: 0.8635 - val_loss: 0.2374 - val_acc: 0.9070\n",
      "Epoch 27/30\n",
      " - 34s - loss: 0.2862 - acc: 0.8740 - val_loss: 0.2391 - val_acc: 0.9110\n",
      "Epoch 28/30\n",
      " - 34s - loss: 0.2897 - acc: 0.8755 - val_loss: 0.2393 - val_acc: 0.9100\n",
      "Epoch 29/30\n",
      " - 34s - loss: 0.2846 - acc: 0.8745 - val_loss: 0.2468 - val_acc: 0.9000\n",
      "Epoch 30/30\n",
      " - 34s - loss: 0.2871 - acc: 0.8820 - val_loss: 0.2373 - val_acc: 0.9090\n"
     ]
    }
   ],
   "source": [
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "\n",
    "train_datagen = ImageDataGenerator(\n",
    "      rescale=1./255,\n",
    "      rotation_range=40,\n",
    "      width_shift_range=0.2,\n",
    "      height_shift_range=0.2,\n",
    "      shear_range=0.2,\n",
    "      zoom_range=0.2,\n",
    "      horizontal_flip=True,\n",
    "      fill_mode='nearest')\n",
    "\n",
    "# Note that the validation data should not be augmented!\n",
    "test_datagen = ImageDataGenerator(rescale=1./255)\n",
    "\n",
    "train_generator = train_datagen.flow_from_directory(\n",
    "        # This is the target directory\n",
    "        train_dir,\n",
    "        # All images will be resized to 150x150\n",
    "        target_size=(150, 150),\n",
    "        batch_size=20,\n",
    "        # Since we use binary_crossentropy loss, we need binary labels\n",
    "        class_mode='binary')\n",
    "\n",
    "validation_generator = test_datagen.flow_from_directory(\n",
    "        validation_dir,\n",
    "        target_size=(150, 150),\n",
    "        batch_size=20,\n",
    "        class_mode='binary')\n",
    "\n",
    "model.compile(loss='binary_crossentropy',\n",
    "              optimizer=optimizers.RMSprop(lr=2e-5),\n",
    "              metrics=['acc'])\n",
    "\n",
    "history = model.fit_generator(\n",
    "      train_generator,\n",
    "      steps_per_epoch=100,\n",
    "      epochs=30,\n",
    "      validation_data=validation_generator,\n",
    "      validation_steps=50,\n",
    "      verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save('data/cats_and_dogs_small_3.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "acc = history.history['acc']\n",
    "val_acc = history.history['val_acc']\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "epochs = range(len(acc))\n",
    "\n",
    "plt.plot(epochs, acc, 'bo', label='Training acc')\n",
    "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
    "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Fine-tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         (None, 150, 150, 3)       0         \n",
      "_________________________________________________________________\n",
      "block1_conv1 (Conv2D)        (None, 150, 150, 64)      1792      \n",
      "_________________________________________________________________\n",
      "block1_conv2 (Conv2D)        (None, 150, 150, 64)      36928     \n",
      "_________________________________________________________________\n",
      "block1_pool (MaxPooling2D)   (None, 75, 75, 64)        0         \n",
      "_________________________________________________________________\n",
      "block2_conv1 (Conv2D)        (None, 75, 75, 128)       73856     \n",
      "_________________________________________________________________\n",
      "block2_conv2 (Conv2D)        (None, 75, 75, 128)       147584    \n",
      "_________________________________________________________________\n",
      "block2_pool (MaxPooling2D)   (None, 37, 37, 128)       0         \n",
      "_________________________________________________________________\n",
      "block3_conv1 (Conv2D)        (None, 37, 37, 256)       295168    \n",
      "_________________________________________________________________\n",
      "block3_conv2 (Conv2D)        (None, 37, 37, 256)       590080    \n",
      "_________________________________________________________________\n",
      "block3_conv3 (Conv2D)        (None, 37, 37, 256)       590080    \n",
      "_________________________________________________________________\n",
      "block3_pool (MaxPooling2D)   (None, 18, 18, 256)       0         \n",
      "_________________________________________________________________\n",
      "block4_conv1 (Conv2D)        (None, 18, 18, 512)       1180160   \n",
      "_________________________________________________________________\n",
      "block4_conv2 (Conv2D)        (None, 18, 18, 512)       2359808   \n",
      "_________________________________________________________________\n",
      "block4_conv3 (Conv2D)        (None, 18, 18, 512)       2359808   \n",
      "_________________________________________________________________\n",
      "block4_pool (MaxPooling2D)   (None, 9, 9, 512)         0         \n",
      "_________________________________________________________________\n",
      "block5_conv1 (Conv2D)        (None, 9, 9, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv2 (Conv2D)        (None, 9, 9, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv3 (Conv2D)        (None, 9, 9, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_pool (MaxPooling2D)   (None, 4, 4, 512)         0         \n",
      "=================================================================\n",
      "Total params: 14,714,688\n",
      "Trainable params: 0\n",
      "Non-trainable params: 14,714,688\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "conv_base.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "conv_base.trainable = True\n",
    "\n",
    "set_trainable = False\n",
    "for layer in conv_base.layers:\n",
    "    if layer.name == 'block5_conv1':\n",
    "        set_trainable = True\n",
    "    if set_trainable:\n",
    "        layer.trainable = True\n",
    "    else:\n",
    "        layer.trainable = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "100/100 [==============================] - 40s 405ms/step - loss: 0.2878 - acc: 0.8730 - val_loss: 0.2221 - val_acc: 0.9110\n",
      "Epoch 2/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.2807 - acc: 0.8750 - val_loss: 0.2082 - val_acc: 0.9170\n",
      "Epoch 3/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.2383 - acc: 0.8995 - val_loss: 0.2029 - val_acc: 0.9210\n",
      "Epoch 4/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.2051 - acc: 0.9140 - val_loss: 0.2139 - val_acc: 0.9180\n",
      "Epoch 5/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.1966 - acc: 0.9175 - val_loss: 0.2279 - val_acc: 0.9070\n",
      "Epoch 6/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.1916 - acc: 0.9150 - val_loss: 0.1984 - val_acc: 0.9250\n",
      "Epoch 7/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.1793 - acc: 0.9310 - val_loss: 0.2193 - val_acc: 0.9190\n",
      "Epoch 8/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.1635 - acc: 0.9350 - val_loss: 0.1882 - val_acc: 0.9250\n",
      "Epoch 9/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.1626 - acc: 0.9350 - val_loss: 0.1940 - val_acc: 0.9270\n",
      "Epoch 10/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.1636 - acc: 0.9295 - val_loss: 0.2276 - val_acc: 0.9100\n",
      "Epoch 11/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.1653 - acc: 0.9355 - val_loss: 0.2051 - val_acc: 0.9180\n",
      "Epoch 12/100\n",
      "100/100 [==============================] - 39s 394ms/step - loss: 0.1304 - acc: 0.9495 - val_loss: 0.1950 - val_acc: 0.9290\n",
      "Epoch 13/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.1309 - acc: 0.9530 - val_loss: 0.1779 - val_acc: 0.9320\n",
      "Epoch 14/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.1115 - acc: 0.9590 - val_loss: 0.1852 - val_acc: 0.9250\n",
      "Epoch 15/100\n",
      "100/100 [==============================] - 39s 391ms/step - loss: 0.1200 - acc: 0.9515 - val_loss: 0.2176 - val_acc: 0.9240\n",
      "Epoch 16/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.1116 - acc: 0.9545 - val_loss: 0.1870 - val_acc: 0.9270\n",
      "Epoch 17/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.1139 - acc: 0.9550 - val_loss: 0.2053 - val_acc: 0.9180\n",
      "Epoch 18/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.1034 - acc: 0.9600 - val_loss: 0.1912 - val_acc: 0.9340\n",
      "Epoch 19/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.1069 - acc: 0.9520 - val_loss: 0.2140 - val_acc: 0.9220\n",
      "Epoch 20/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0918 - acc: 0.9665 - val_loss: 0.1776 - val_acc: 0.9330\n",
      "Epoch 21/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0915 - acc: 0.9660 - val_loss: 0.1899 - val_acc: 0.9310\n",
      "Epoch 22/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0881 - acc: 0.9660 - val_loss: 0.2221 - val_acc: 0.9220\n",
      "Epoch 23/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0926 - acc: 0.9625 - val_loss: 0.1927 - val_acc: 0.9340\n",
      "Epoch 24/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0858 - acc: 0.9680 - val_loss: 0.1904 - val_acc: 0.9300\n",
      "Epoch 25/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0730 - acc: 0.9715 - val_loss: 0.1833 - val_acc: 0.9370\n",
      "Epoch 26/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0867 - acc: 0.9655 - val_loss: 0.1984 - val_acc: 0.9320\n",
      "Epoch 27/100\n",
      "100/100 [==============================] - 39s 394ms/step - loss: 0.0800 - acc: 0.9670 - val_loss: 0.1833 - val_acc: 0.9390\n",
      "Epoch 28/100\n",
      "100/100 [==============================] - 39s 394ms/step - loss: 0.0768 - acc: 0.9725 - val_loss: 0.2106 - val_acc: 0.9330\n",
      "Epoch 29/100\n",
      "100/100 [==============================] - 39s 394ms/step - loss: 0.0728 - acc: 0.9690 - val_loss: 0.2326 - val_acc: 0.9200\n",
      "Epoch 30/100\n",
      "100/100 [==============================] - 39s 394ms/step - loss: 0.0780 - acc: 0.9680 - val_loss: 0.2317 - val_acc: 0.9340\n",
      "Epoch 31/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0634 - acc: 0.9765 - val_loss: 0.2722 - val_acc: 0.9120\n",
      "Epoch 32/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0712 - acc: 0.9720 - val_loss: 0.1833 - val_acc: 0.9360\n",
      "Epoch 33/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0667 - acc: 0.9740 - val_loss: 0.1988 - val_acc: 0.9260\n",
      "Epoch 34/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0643 - acc: 0.9745 - val_loss: 0.1854 - val_acc: 0.9370\n",
      "Epoch 35/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0712 - acc: 0.9740 - val_loss: 0.1900 - val_acc: 0.9380\n",
      "Epoch 36/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0617 - acc: 0.9755 - val_loss: 0.1928 - val_acc: 0.9380\n",
      "Epoch 37/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0437 - acc: 0.9850 - val_loss: 0.1763 - val_acc: 0.9380\n",
      "Epoch 38/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0498 - acc: 0.9770 - val_loss: 0.1966 - val_acc: 0.9380\n",
      "Epoch 39/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0560 - acc: 0.9775 - val_loss: 0.2372 - val_acc: 0.9320\n",
      "Epoch 40/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0449 - acc: 0.9835 - val_loss: 0.3027 - val_acc: 0.9160\n",
      "Epoch 41/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0416 - acc: 0.9855 - val_loss: 0.2145 - val_acc: 0.9270\n",
      "Epoch 42/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0522 - acc: 0.9820 - val_loss: 0.2038 - val_acc: 0.9370\n",
      "Epoch 43/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0439 - acc: 0.9835 - val_loss: 0.2216 - val_acc: 0.9310\n",
      "Epoch 44/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0335 - acc: 0.9870 - val_loss: 0.3642 - val_acc: 0.9110\n",
      "Epoch 45/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0534 - acc: 0.9780 - val_loss: 0.2233 - val_acc: 0.9310\n",
      "Epoch 46/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0363 - acc: 0.9870 - val_loss: 0.1970 - val_acc: 0.9370\n",
      "Epoch 47/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0461 - acc: 0.9830 - val_loss: 0.1882 - val_acc: 0.9360\n",
      "Epoch 48/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0456 - acc: 0.9855 - val_loss: 0.1829 - val_acc: 0.9380\n",
      "Epoch 49/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0520 - acc: 0.9830 - val_loss: 0.2111 - val_acc: 0.9280\n",
      "Epoch 50/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0307 - acc: 0.9875 - val_loss: 0.2284 - val_acc: 0.9400\n",
      "Epoch 51/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0401 - acc: 0.9845 - val_loss: 0.2624 - val_acc: 0.9260\n",
      "Epoch 52/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0356 - acc: 0.9860 - val_loss: 0.2288 - val_acc: 0.9350\n",
      "Epoch 53/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0327 - acc: 0.9890 - val_loss: 0.2047 - val_acc: 0.9380\n",
      "Epoch 54/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0346 - acc: 0.9885 - val_loss: 0.3103 - val_acc: 0.9250\n",
      "Epoch 55/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0306 - acc: 0.9895 - val_loss: 0.2210 - val_acc: 0.9310\n",
      "Epoch 56/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0334 - acc: 0.9875 - val_loss: 0.2840 - val_acc: 0.9290\n",
      "Epoch 57/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0395 - acc: 0.9860 - val_loss: 0.2291 - val_acc: 0.9360\n",
      "Epoch 58/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0349 - acc: 0.9870 - val_loss: 0.2157 - val_acc: 0.9350\n",
      "Epoch 59/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0317 - acc: 0.9875 - val_loss: 0.2233 - val_acc: 0.9390\n",
      "Epoch 60/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0259 - acc: 0.9890 - val_loss: 0.2324 - val_acc: 0.9320\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 61/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0402 - acc: 0.9875 - val_loss: 0.2563 - val_acc: 0.9320\n",
      "Epoch 62/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0371 - acc: 0.9885 - val_loss: 0.2055 - val_acc: 0.9350\n",
      "Epoch 63/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0236 - acc: 0.9910 - val_loss: 0.4129 - val_acc: 0.9110\n",
      "Epoch 64/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0300 - acc: 0.9880 - val_loss: 0.3673 - val_acc: 0.9170\n",
      "Epoch 65/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0308 - acc: 0.9870 - val_loss: 0.2594 - val_acc: 0.9360\n",
      "Epoch 66/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0209 - acc: 0.9920 - val_loss: 0.2493 - val_acc: 0.9360\n",
      "Epoch 67/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0282 - acc: 0.9895 - val_loss: 0.3022 - val_acc: 0.9260\n",
      "Epoch 68/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0285 - acc: 0.9890 - val_loss: 0.2350 - val_acc: 0.9380\n",
      "Epoch 69/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0289 - acc: 0.9900 - val_loss: 0.2293 - val_acc: 0.9330\n",
      "Epoch 70/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0227 - acc: 0.9930 - val_loss: 0.3599 - val_acc: 0.9150\n",
      "Epoch 71/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0216 - acc: 0.9920 - val_loss: 0.2733 - val_acc: 0.9290\n",
      "Epoch 72/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0383 - acc: 0.9875 - val_loss: 0.3335 - val_acc: 0.9220\n",
      "Epoch 73/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0241 - acc: 0.9940 - val_loss: 0.2504 - val_acc: 0.9370\n",
      "Epoch 74/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0244 - acc: 0.9905 - val_loss: 0.2727 - val_acc: 0.9370\n",
      "Epoch 75/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0214 - acc: 0.9935 - val_loss: 0.2695 - val_acc: 0.9320\n",
      "Epoch 76/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0182 - acc: 0.9955 - val_loss: 0.2637 - val_acc: 0.9350\n",
      "Epoch 77/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0215 - acc: 0.9930 - val_loss: 0.2450 - val_acc: 0.9350\n",
      "Epoch 78/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0165 - acc: 0.9945 - val_loss: 0.2850 - val_acc: 0.9300\n",
      "Epoch 79/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0307 - acc: 0.9885 - val_loss: 0.2683 - val_acc: 0.9350\n",
      "Epoch 80/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0231 - acc: 0.9940 - val_loss: 0.2124 - val_acc: 0.9420\n",
      "Epoch 81/100\n",
      "100/100 [==============================] - 39s 391ms/step - loss: 0.0229 - acc: 0.9915 - val_loss: 0.2421 - val_acc: 0.9420\n",
      "Epoch 82/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0210 - acc: 0.9935 - val_loss: 0.2437 - val_acc: 0.9350\n",
      "Epoch 83/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0255 - acc: 0.9925 - val_loss: 0.2698 - val_acc: 0.9340\n",
      "Epoch 84/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0175 - acc: 0.9940 - val_loss: 0.2254 - val_acc: 0.9490\n",
      "Epoch 85/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0295 - acc: 0.9895 - val_loss: 0.2539 - val_acc: 0.9360\n",
      "Epoch 86/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0236 - acc: 0.9920 - val_loss: 0.2193 - val_acc: 0.9420\n",
      "Epoch 87/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0241 - acc: 0.9910 - val_loss: 0.2450 - val_acc: 0.9390\n",
      "Epoch 88/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0207 - acc: 0.9915 - val_loss: 0.3051 - val_acc: 0.9240\n",
      "Epoch 89/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0146 - acc: 0.9945 - val_loss: 0.3384 - val_acc: 0.9220\n",
      "Epoch 90/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0232 - acc: 0.9905 - val_loss: 0.3136 - val_acc: 0.9300\n",
      "Epoch 91/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0202 - acc: 0.9915 - val_loss: 0.2664 - val_acc: 0.9380\n",
      "Epoch 92/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0174 - acc: 0.9945 - val_loss: 0.3016 - val_acc: 0.9370\n",
      "Epoch 93/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0157 - acc: 0.9935 - val_loss: 0.3773 - val_acc: 0.9260\n",
      "Epoch 94/100\n",
      "100/100 [==============================] - 39s 394ms/step - loss: 0.0138 - acc: 0.9935 - val_loss: 0.2421 - val_acc: 0.9440\n",
      "Epoch 95/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0269 - acc: 0.9900 - val_loss: 0.2374 - val_acc: 0.9380\n",
      "Epoch 96/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0214 - acc: 0.9910 - val_loss: 0.2201 - val_acc: 0.9430\n",
      "Epoch 97/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0177 - acc: 0.9950 - val_loss: 0.2657 - val_acc: 0.9370\n",
      "Epoch 98/100\n",
      "100/100 [==============================] - 39s 393ms/step - loss: 0.0254 - acc: 0.9890 - val_loss: 0.2313 - val_acc: 0.9420\n",
      "Epoch 99/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0160 - acc: 0.9930 - val_loss: 0.2044 - val_acc: 0.9440\n",
      "Epoch 100/100\n",
      "100/100 [==============================] - 39s 392ms/step - loss: 0.0204 - acc: 0.9925 - val_loss: 0.2301 - val_acc: 0.9450\n"
     ]
    }
   ],
   "source": [
    "model.compile(loss='binary_crossentropy',\n",
    "              optimizer=optimizers.RMSprop(lr=1e-5),\n",
    "              metrics=['acc'])\n",
    "\n",
    "history = model.fit_generator(\n",
    "      train_generator,\n",
    "      steps_per_epoch=100,\n",
    "      epochs=100,\n",
    "      validation_data=validation_generator,\n",
    "      validation_steps=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save('data/cats_and_dogs_small_4.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "acc = history.history['acc']\n",
    "val_acc = history.history['val_acc']\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "epochs = range(len(acc))\n",
    "\n",
    "plt.plot(epochs, acc, 'bo', label='Training acc')\n",
    "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
    "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "def smooth_curve(points, factor=0.8):\n",
    "  smoothed_points = []\n",
    "  for point in points:\n",
    "    if smoothed_points:\n",
    "      previous = smoothed_points[-1]\n",
    "      smoothed_points.append(previous * factor + point * (1 - factor))\n",
    "    else:\n",
    "      smoothed_points.append(point)\n",
    "  return smoothed_points"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(epochs,\n",
    "         smooth_curve(acc), 'bo', label='Smoothed training acc')\n",
    "plt.plot(epochs,\n",
    "         smooth_curve(val_acc), 'b', label='Smoothed validation acc')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs,\n",
    "         smooth_curve(loss), 'bo', label='Smoothed training loss')\n",
    "plt.plot(epochs,\n",
    "         smooth_curve(val_loss), 'b', label='Smoothed validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 1000 images belonging to 2 classes.\n",
      "test acc: 0.9389999902248383\n"
     ]
    }
   ],
   "source": [
    "test_generator = test_datagen.flow_from_directory(\n",
    "        test_dir,\n",
    "        target_size=(150, 150),\n",
    "        batch_size=20,\n",
    "        class_mode='binary')\n",
    "\n",
    "test_loss, test_acc = model.evaluate_generator(test_generator, steps=50)\n",
    "print('test acc:', test_acc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "TensorFlow-GPU",
   "language": "python",
   "name": "tf-gpu"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}