model_mobilenet.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "from __future__ import absolute_import, division, print_function, unicode_literals\n",
    "import os\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import tensorflow as tf\n",
    "import os\n",
    "import sys\n",
    "from PIL import Image\n",
    "from tensorflow import keras"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "base_dir = os.getcwd()+\"/data\"\n",
    "train_dir = os.path.join(base_dir, 'train')\n",
    "validation_dir = os.path.join(base_dir, 'validation')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'keras' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-2-0baa7f93dd30>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0mimage_size\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m160\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;31m# Rescale all images by 1./255 and apply image augmentation\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m train_datagen = keras.preprocessing.image.ImageDataGenerator(\n\u001b[0m\u001b[1;32m      5\u001b[0m                 rescale=1./255)\n\u001b[1;32m      6\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mNameError\u001b[0m: name 'keras' is not defined"
     ]
    }
   ],
   "source": [
    "batch_size = 32\n",
    "image_size = 160\n",
    "# Rescale all images by 1./255 and apply image augmentation\n",
    "train_datagen = keras.preprocessing.image.ImageDataGenerator(\n",
    "                rescale=1./255)\n",
    "\n",
    "validation_datagen = keras.preprocessing.image.ImageDataGenerator(rescale=1./255)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 5364 images belonging to 6 classes.\n",
      "Found 1338 images belonging to 6 classes.\n"
     ]
    }
   ],
   "source": [
    "# Flow training images in batches of 20 using train_datagen generator\n",
    "train_generator = train_datagen.flow_from_directory(\n",
    "                train_dir,  # Source directory for the training images\n",
    "                target_size=(image_size, image_size),  \n",
    "                batch_size=batch_size,\n",
    "                # Since we use binary_crossentropy loss, we need binary labels\n",
    "                class_mode='categorical')\n",
    "\n",
    "# Flow validation images in batches of 20 using test_datagen generator\n",
    "validation_generator = validation_datagen.flow_from_directory(\n",
    "                validation_dir, # Source directory for the validation images\n",
    "                target_size=(image_size, image_size),\n",
    "                batch_size=batch_size,\n",
    "                class_mode='categorical')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'image_size' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-1-20da494ae631>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mIMG_SHAPE\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mimage_size\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mimage_size\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m3\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;31m# Create the base model from the pre-trained model MobileNet V2\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m base_model = tf.keras.applications.VGG16(input_shape=IMG_SHAPE,\n\u001b[1;32m      5\u001b[0m                                                \u001b[0minclude_top\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mNameError\u001b[0m: name 'image_size' is not defined"
     ]
    }
   ],
   "source": [
    "IMG_SHAPE = (image_size, image_size, 3)\n",
    "\n",
    "# Create the base model from the pre-trained model MobileNet V2\n",
    "base_model = tf.keras.applications.MobileNet(input_shape=IMG_SHAPE,\n",
    "                                               include_top=False, \n",
    "                                               weights='imagenet')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "base_model.trainable = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         (None, 160, 160, 3)       0         \n",
      "_________________________________________________________________\n",
      "conv1 (Conv2D)               (None, 80, 80, 32)        864       \n",
      "_________________________________________________________________\n",
      "conv1_bn (BatchNormalization (None, 80, 80, 32)        128       \n",
      "_________________________________________________________________\n",
      "conv1_relu (Activation)      (None, 80, 80, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv_dw_1 (DepthwiseConv2D)  (None, 80, 80, 32)        288       \n",
      "_________________________________________________________________\n",
      "conv_dw_1_bn (BatchNormaliza (None, 80, 80, 32)        128       \n",
      "_________________________________________________________________\n",
      "conv_dw_1_relu (Activation)  (None, 80, 80, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv_pw_1 (Conv2D)           (None, 80, 80, 64)        2048      \n",
      "_________________________________________________________________\n",
      "conv_pw_1_bn (BatchNormaliza (None, 80, 80, 64)        256       \n",
      "_________________________________________________________________\n",
      "conv_pw_1_relu (Activation)  (None, 80, 80, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv_dw_2 (DepthwiseConv2D)  (None, 40, 40, 64)        576       \n",
      "_________________________________________________________________\n",
      "conv_dw_2_bn (BatchNormaliza (None, 40, 40, 64)        256       \n",
      "_________________________________________________________________\n",
      "conv_dw_2_relu (Activation)  (None, 40, 40, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv_pw_2 (Conv2D)           (None, 40, 40, 128)       8192      \n",
      "_________________________________________________________________\n",
      "conv_pw_2_bn (BatchNormaliza (None, 40, 40, 128)       512       \n",
      "_________________________________________________________________\n",
      "conv_pw_2_relu (Activation)  (None, 40, 40, 128)       0         \n",
      "_________________________________________________________________\n",
      "conv_dw_3 (DepthwiseConv2D)  (None, 40, 40, 128)       1152      \n",
      "_________________________________________________________________\n",
      "conv_dw_3_bn (BatchNormaliza (None, 40, 40, 128)       512       \n",
      "_________________________________________________________________\n",
      "conv_dw_3_relu (Activation)  (None, 40, 40, 128)       0         \n",
      "_________________________________________________________________\n",
      "conv_pw_3 (Conv2D)           (None, 40, 40, 128)       16384     \n",
      "_________________________________________________________________\n",
      "conv_pw_3_bn (BatchNormaliza (None, 40, 40, 128)       512       \n",
      "_________________________________________________________________\n",
      "conv_pw_3_relu (Activation)  (None, 40, 40, 128)       0         \n",
      "_________________________________________________________________\n",
      "conv_dw_4 (DepthwiseConv2D)  (None, 20, 20, 128)       1152      \n",
      "_________________________________________________________________\n",
      "conv_dw_4_bn (BatchNormaliza (None, 20, 20, 128)       512       \n",
      "_________________________________________________________________\n",
      "conv_dw_4_relu (Activation)  (None, 20, 20, 128)       0         \n",
      "_________________________________________________________________\n",
      "conv_pw_4 (Conv2D)           (None, 20, 20, 256)       32768     \n",
      "_________________________________________________________________\n",
      "conv_pw_4_bn (BatchNormaliza (None, 20, 20, 256)       1024      \n",
      "_________________________________________________________________\n",
      "conv_pw_4_relu (Activation)  (None, 20, 20, 256)       0         \n",
      "_________________________________________________________________\n",
      "conv_dw_5 (DepthwiseConv2D)  (None, 20, 20, 256)       2304      \n",
      "_________________________________________________________________\n",
      "conv_dw_5_bn (BatchNormaliza (None, 20, 20, 256)       1024      \n",
      "_________________________________________________________________\n",
      "conv_dw_5_relu (Activation)  (None, 20, 20, 256)       0         \n",
      "_________________________________________________________________\n",
      "conv_pw_5 (Conv2D)           (None, 20, 20, 256)       65536     \n",
      "_________________________________________________________________\n",
      "conv_pw_5_bn (BatchNormaliza (None, 20, 20, 256)       1024      \n",
      "_________________________________________________________________\n",
      "conv_pw_5_relu (Activation)  (None, 20, 20, 256)       0         \n",
      "_________________________________________________________________\n",
      "conv_dw_6 (DepthwiseConv2D)  (None, 10, 10, 256)       2304      \n",
      "_________________________________________________________________\n",
      "conv_dw_6_bn (BatchNormaliza (None, 10, 10, 256)       1024      \n",
      "_________________________________________________________________\n",
      "conv_dw_6_relu (Activation)  (None, 10, 10, 256)       0         \n",
      "_________________________________________________________________\n",
      "conv_pw_6 (Conv2D)           (None, 10, 10, 512)       131072    \n",
      "_________________________________________________________________\n",
      "conv_pw_6_bn (BatchNormaliza (None, 10, 10, 512)       2048      \n",
      "_________________________________________________________________\n",
      "conv_pw_6_relu (Activation)  (None, 10, 10, 512)       0         \n",
      "_________________________________________________________________\n",
      "conv_dw_7 (DepthwiseConv2D)  (None, 10, 10, 512)       4608      \n",
      "_________________________________________________________________\n",
      "conv_dw_7_bn (BatchNormaliza (None, 10, 10, 512)       2048      \n",
      "_________________________________________________________________\n",
      "conv_dw_7_relu (Activation)  (None, 10, 10, 512)       0         \n",
      "_________________________________________________________________\n",
      "conv_pw_7 (Conv2D)           (None, 10, 10, 512)       262144    \n",
      "_________________________________________________________________\n",
      "conv_pw_7_bn (BatchNormaliza (None, 10, 10, 512)       2048      \n",
      "_________________________________________________________________\n",
      "conv_pw_7_relu (Activation)  (None, 10, 10, 512)       0         \n",
      "_________________________________________________________________\n",
      "conv_dw_8 (DepthwiseConv2D)  (None, 10, 10, 512)       4608      \n",
      "_________________________________________________________________\n",
      "conv_dw_8_bn (BatchNormaliza (None, 10, 10, 512)       2048      \n",
      "_________________________________________________________________\n",
      "conv_dw_8_relu (Activation)  (None, 10, 10, 512)       0         \n",
      "_________________________________________________________________\n",
      "conv_pw_8 (Conv2D)           (None, 10, 10, 512)       262144    \n",
      "_________________________________________________________________\n",
      "conv_pw_8_bn (BatchNormaliza (None, 10, 10, 512)       2048      \n",
      "_________________________________________________________________\n",
      "conv_pw_8_relu (Activation)  (None, 10, 10, 512)       0         \n",
      "_________________________________________________________________\n",
      "conv_dw_9 (DepthwiseConv2D)  (None, 10, 10, 512)       4608      \n",
      "_________________________________________________________________\n",
      "conv_dw_9_bn (BatchNormaliza (None, 10, 10, 512)       2048      \n",
      "_________________________________________________________________\n",
      "conv_dw_9_relu (Activation)  (None, 10, 10, 512)       0         \n",
      "_________________________________________________________________\n",
      "conv_pw_9 (Conv2D)           (None, 10, 10, 512)       262144    \n",
      "_________________________________________________________________\n",
      "conv_pw_9_bn (BatchNormaliza (None, 10, 10, 512)       2048      \n",
      "_________________________________________________________________\n",
      "conv_pw_9_relu (Activation)  (None, 10, 10, 512)       0         \n",
      "_________________________________________________________________\n",
      "conv_dw_10 (DepthwiseConv2D) (None, 10, 10, 512)       4608      \n",
      "_________________________________________________________________\n",
      "conv_dw_10_bn (BatchNormaliz (None, 10, 10, 512)       2048      \n",
      "_________________________________________________________________\n",
      "conv_dw_10_relu (Activation) (None, 10, 10, 512)       0         \n",
      "_________________________________________________________________\n",
      "conv_pw_10 (Conv2D)          (None, 10, 10, 512)       262144    \n",
      "_________________________________________________________________\n",
      "conv_pw_10_bn (BatchNormaliz (None, 10, 10, 512)       2048      \n",
      "_________________________________________________________________\n",
      "conv_pw_10_relu (Activation) (None, 10, 10, 512)       0         \n",
      "_________________________________________________________________\n",
      "conv_dw_11 (DepthwiseConv2D) (None, 10, 10, 512)       4608      \n",
      "_________________________________________________________________\n",
      "conv_dw_11_bn (BatchNormaliz (None, 10, 10, 512)       2048      \n",
      "_________________________________________________________________\n",
      "conv_dw_11_relu (Activation) (None, 10, 10, 512)       0         \n",
      "_________________________________________________________________\n",
      "conv_pw_11 (Conv2D)          (None, 10, 10, 512)       262144    \n",
      "_________________________________________________________________\n",
      "conv_pw_11_bn (BatchNormaliz (None, 10, 10, 512)       2048      \n",
      "_________________________________________________________________\n",
      "conv_pw_11_relu (Activation) (None, 10, 10, 512)       0         \n",
      "_________________________________________________________________\n",
      "conv_dw_12 (DepthwiseConv2D) (None, 5, 5, 512)         4608      \n",
      "_________________________________________________________________\n",
      "conv_dw_12_bn (BatchNormaliz (None, 5, 5, 512)         2048      \n",
      "_________________________________________________________________\n",
      "conv_dw_12_relu (Activation) (None, 5, 5, 512)         0         \n",
      "_________________________________________________________________\n",
      "conv_pw_12 (Conv2D)          (None, 5, 5, 1024)        524288    \n",
      "_________________________________________________________________\n",
      "conv_pw_12_bn (BatchNormaliz (None, 5, 5, 1024)        4096      \n",
      "_________________________________________________________________\n",
      "conv_pw_12_relu (Activation) (None, 5, 5, 1024)        0         \n",
      "_________________________________________________________________\n",
      "conv_dw_13 (DepthwiseConv2D) (None, 5, 5, 1024)        9216      \n",
      "_________________________________________________________________\n",
      "conv_dw_13_bn (BatchNormaliz (None, 5, 5, 1024)        4096      \n",
      "_________________________________________________________________\n",
      "conv_dw_13_relu (Activation) (None, 5, 5, 1024)        0         \n",
      "_________________________________________________________________\n",
      "conv_pw_13 (Conv2D)          (None, 5, 5, 1024)        1048576   \n",
      "_________________________________________________________________\n",
      "conv_pw_13_bn (BatchNormaliz (None, 5, 5, 1024)        4096      \n",
      "_________________________________________________________________\n",
      "conv_pw_13_relu (Activation) (None, 5, 5, 1024)        0         \n",
      "=================================================================\n",
      "Total params: 3,228,864\n",
      "Trainable params: 0\n",
      "Non-trainable params: 3,228,864\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "base_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = tf.keras.Sequential([\n",
    "  base_model,\n",
    "  keras.layers.GlobalAveragePooling2D(),\n",
    "  keras.layers.Dense(6, activation='softmax')\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(optimizer=tf.keras.optimizers.RMSprop(lr=0.0001), \n",
    "              loss='categorical_crossentropy', \n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "mobilenet_1.00_160 (Model)   (None, 5, 5, 1024)        3228864   \n",
      "_________________________________________________________________\n",
      "global_average_pooling2d_1 ( (None, 1024)              0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 6)                 6150      \n",
      "=================================================================\n",
      "Total params: 3,235,014\n",
      "Trainable params: 6,150\n",
      "Non-trainable params: 3,228,864\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(model.trainable_variables)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "167/167 [==============================]167/167 [==============================] - 394s 2s/step - loss: 1.2797 - acc: 0.5587 - val_loss: 1.4549 - val_acc: 0.5465\n",
      "\n",
      "Epoch 2/10\n",
      "167/167 [==============================]167/167 [==============================] - 419s 3s/step - loss: 0.3193 - acc: 0.9480 - val_loss: 1.0449 - val_acc: 0.7213\n",
      "\n",
      "Epoch 3/10\n",
      "167/167 [==============================]167/167 [==============================] - 418s 3s/step - loss: 0.1116 - acc: 0.9899 - val_loss: 0.9237 - val_acc: 0.7688\n",
      "\n",
      "Epoch 4/10\n",
      "167/167 [==============================]167/167 [==============================] - 411s 2s/step - loss: 0.0503 - acc: 0.9976 - val_loss: 0.8482 - val_acc: 0.7825\n",
      "\n",
      "Epoch 5/10\n",
      "167/167 [==============================]167/167 [==============================] - 416s 2s/step - loss: 0.0295 - acc: 0.9981 - val_loss: 0.7627 - val_acc: 0.7940\n",
      "\n",
      "Epoch 6/10\n",
      "167/167 [==============================]167/167 [==============================] - 408s 2s/step - loss: 0.0177 - acc: 0.9993 - val_loss: 0.7885 - val_acc: 0.7879\n",
      "\n",
      "Epoch 7/10\n",
      "167/167 [==============================]167/167 [==============================] - 405s 2s/step - loss: 0.0117 - acc: 0.9998 - val_loss: 0.6861 - val_acc: 0.7979\n",
      "\n",
      "Epoch 8/10\n",
      "167/167 [==============================]167/167 [==============================] - 403s 2s/step - loss: 0.0087 - acc: 0.9998 - val_loss: 0.7456 - val_acc: 0.8002\n",
      "\n",
      "Epoch 9/10\n",
      "167/167 [==============================]167/167 [==============================] - 403s 2s/step - loss: 0.0067 - acc: 0.9994 - val_loss: 0.6875 - val_acc: 0.8070\n",
      "\n",
      "Epoch 10/10\n",
      "167/167 [==============================]167/167 [==============================] - 404s 2s/step - loss: 0.0056 - acc: 0.9998 - val_loss: 0.6825 - val_acc: 0.8101\n",
      "\n"
     ]
    }
   ],
   "source": [
    "epochs = 10\n",
    "steps_per_epoch = train_generator.n // batch_size\n",
    "validation_steps = validation_generator.n // batch_size\n",
    "\n",
    "history = model.fit_generator(train_generator, \n",
    "                              steps_per_epoch = steps_per_epoch,\n",
    "                              epochs=epochs, \n",
    "                              workers=4,\n",
    "                              validation_data=validation_generator, \n",
    "                              validation_steps=validation_steps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f115430ae10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "acc = history.history['acc']\n",
    "val_acc = history.history['val_acc']\n",
    "\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "plt.figure(figsize=(8, 8))\n",
    "plt.subplot(2, 1, 1)\n",
    "plt.plot(acc, label='Training Accuracy')\n",
    "plt.plot(val_acc, label='Validation Accuracy')\n",
    "plt.legend(loc='lower right')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.ylim([min(plt.ylim()),1])\n",
    "plt.title('Training and Validation Accuracy')\n",
    "\n",
    "plt.subplot(2, 1, 2)\n",
    "plt.plot(loss, label='Training Loss')\n",
    "plt.plot(val_loss, label='Validation Loss')\n",
    "plt.legend(loc='upper right')\n",
    "plt.ylabel('Cross Entropy')\n",
    "plt.ylim([0,max(plt.ylim())])\n",
    "plt.title('Training and Validation Loss')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_json = model.to_json()\n",
    "with open(\"model.json\", \"w\") as json_file:\n",
    "    json_file.write(model_json)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save_weights(\"model.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# base_model.trainable = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of layers in the base model:  82\n"
     ]
    }
   ],
   "source": [
    "# # Let's take a look to see how many layers are in the base model\n",
    "# print(\"Number of layers in the base model: \", len(base_model.layers))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# # Fine tune from this layer onwards\n",
    "# fine_tune_at = 70\n",
    "\n",
    "# # Freeze all the layers before the `fine_tune_at` layer\n",
    "# for layer in base_model.layers[:fine_tune_at]:\n",
    "#   layer.trainable =  False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# model.compile(loss='categorical_crossentropy',\n",
    "#               optimizer = tf.keras.optimizers.RMSprop(lr=2e-5),\n",
    "#               metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "mobilenet_1.00_160 (Model)   (None, 5, 5, 1024)        3228864   \n",
      "_________________________________________________________________\n",
      "global_average_pooling2d_2 ( (None, 1024)              0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 6)                 6150      \n",
      "=================================================================\n",
      "Total params: 3,235,014\n",
      "Trainable params: 1,600,006\n",
      "Non-trainable params: 1,635,008\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "14"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# len(model.trainable_variables)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "94/94 [==============================]94/94 [==============================] - 368s 4s/step - loss: 0.0027 - acc: 1.0000 - val_loss: 0.0538 - val_acc: 0.9875\n",
      "\n",
      "Epoch 2/10\n",
      "94/94 [==============================]94/94 [==============================] - 368s 4s/step - loss: 6.3096e-04 - acc: 1.0000 - val_loss: 0.0013 - val_acc: 1.0000\n",
      "\n",
      "Epoch 3/10\n",
      "94/94 [==============================]94/94 [==============================] - 404s 4s/step - loss: 4.1073e-04 - acc: 1.0000 - val_loss: 1.2539e-04 - val_acc: 1.0000\n",
      "\n",
      "Epoch 4/10\n",
      "94/94 [==============================]94/94 [==============================] - 396s 4s/step - loss: 1.4760e-04 - acc: 1.0000 - val_loss: 6.0510e-05 - val_acc: 1.0000\n",
      "\n",
      "Epoch 5/10\n",
      "94/94 [==============================]94/94 [==============================] - 399s 4s/step - loss: 6.4822e-05 - acc: 1.0000 - val_loss: 2.5963e-05 - val_acc: 1.0000\n",
      "\n",
      "Epoch 6/10\n",
      "94/94 [==============================]94/94 [==============================] - 425s 5s/step - loss: 3.0874e-05 - acc: 1.0000 - val_loss: 4.8057e-06 - val_acc: 1.0000\n",
      "\n",
      "Epoch 7/10\n",
      "94/94 [==============================]94/94 [==============================] - 454s 5s/step - loss: 1.8614e-05 - acc: 1.0000 - val_loss: 9.7569e-06 - val_acc: 1.0000\n",
      "\n",
      "Epoch 8/10\n",
      "94/94 [==============================]94/94 [==============================] - 447s 5s/step - loss: 9.9300e-06 - acc: 1.0000 - val_loss: 1.6326e-06 - val_acc: 1.0000\n",
      "\n",
      "Epoch 9/10\n",
      "94/94 [==============================]94/94 [==============================] - 451s 5s/step - loss: 5.6390e-06 - acc: 1.0000 - val_loss: 3.1508e-06 - val_acc: 1.0000\n",
      "\n",
      "Epoch 10/10\n",
      "94/94 [==============================]94/94 [==============================] - 460s 5s/step - loss: 3.9303e-06 - acc: 1.0000 - val_loss: 5.9427e-07 - val_acc: 1.0000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# history_fine = model.fit_generator(train_generator, \n",
    "#                                    steps_per_epoch = steps_per_epoch,\n",
    "#                                    epochs=epochs, \n",
    "#                                    workers=4,\n",
    "#                                    validation_data=validation_generator, \n",
    "#                                    validation_steps=validation_steps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "# acc += history_fine.history['acc']\n",
    "# val_acc += history_fine.history['val_acc']\n",
    "\n",
    "# loss += history_fine.history['loss']\n",
    "# val_loss += history_fine.history['val_loss']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8029a940f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plt.figure(figsize=(8, 8))\n",
    "# plt.subplot(2, 1, 1)\n",
    "# plt.plot(acc, label='Training Accuracy')\n",
    "# plt.plot(val_acc, label='Validation Accuracy')\n",
    "# plt.ylim([0.9, 1])\n",
    "# plt.plot([epochs-1,epochs-1], plt.ylim(), label='Start Fine Tuning')\n",
    "# plt.legend(loc='lower right')\n",
    "# plt.title('Training and Validation Accuracy')\n",
    "\n",
    "# plt.subplot(2, 1, 2)\n",
    "# plt.plot(loss, label='Training Loss')\n",
    "# plt.plot(val_loss, label='Validation Loss')\n",
    "# plt.ylim([0, 0.2])\n",
    "# plt.plot([epochs-1,epochs-1], plt.ylim(), label='Start Fine Tuning')\n",
    "# plt.legend(loc='upper right')\n",
    "# plt.title('Training and Validation Loss')\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "python_dl",
   "language": "python",
   "name": "python_dl"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}