Baseline_CNN.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "id": "HeBr3U9dzBBg",
    "outputId": "e9658d5f-c4d8-478c-ddcf-d687edddd7e5"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "from keras.utils.np_utils import to_categorical # convert to one-hot-encoding\n",
    "\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "from tensorflow.keras import layers\n",
    "from tensorflow.keras import Model\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "from tensorflow.keras.callbacks import ReduceLROnPlateau\n",
    "%matplotlib inline\n",
    "\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "M1BzjwNzUB1T"
   },
   "source": [
    "## Load In the Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "F3zhcgJD4lhk"
   },
   "outputs": [],
   "source": [
    "skin_df = pd.read_csv('hmnist_64_64_RGB.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 247
    },
    "colab_type": "code",
    "id": "W-viELZ74uN2",
    "outputId": "44f1763e-1b74-4434-e99c-51e295052387"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>pixel0000</th>\n",
       "      <th>pixel0001</th>\n",
       "      <th>pixel0002</th>\n",
       "      <th>pixel0003</th>\n",
       "      <th>pixel0004</th>\n",
       "      <th>pixel0005</th>\n",
       "      <th>pixel0006</th>\n",
       "      <th>pixel0007</th>\n",
       "      <th>pixel0008</th>\n",
       "      <th>pixel0009</th>\n",
       "      <th>...</th>\n",
       "      <th>pixel12279</th>\n",
       "      <th>pixel12280</th>\n",
       "      <th>pixel12281</th>\n",
       "      <th>pixel12282</th>\n",
       "      <th>pixel12283</th>\n",
       "      <th>pixel12284</th>\n",
       "      <th>pixel12285</th>\n",
       "      <th>pixel12286</th>\n",
       "      <th>pixel12287</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>191</td>\n",
       "      <td>152</td>\n",
       "      <td>194</td>\n",
       "      <td>191</td>\n",
       "      <td>153</td>\n",
       "      <td>195</td>\n",
       "      <td>192</td>\n",
       "      <td>149</td>\n",
       "      <td>192</td>\n",
       "      <td>194</td>\n",
       "      <td>...</td>\n",
       "      <td>185</td>\n",
       "      <td>154</td>\n",
       "      <td>180</td>\n",
       "      <td>186</td>\n",
       "      <td>156</td>\n",
       "      <td>184</td>\n",
       "      <td>182</td>\n",
       "      <td>152</td>\n",
       "      <td>173</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>24</td>\n",
       "      <td>13</td>\n",
       "      <td>23</td>\n",
       "      <td>24</td>\n",
       "      <td>14</td>\n",
       "      <td>28</td>\n",
       "      <td>37</td>\n",
       "      <td>24</td>\n",
       "      <td>46</td>\n",
       "      <td>61</td>\n",
       "      <td>...</td>\n",
       "      <td>26</td>\n",
       "      <td>15</td>\n",
       "      <td>29</td>\n",
       "      <td>27</td>\n",
       "      <td>15</td>\n",
       "      <td>28</td>\n",
       "      <td>24</td>\n",
       "      <td>13</td>\n",
       "      <td>25</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>185</td>\n",
       "      <td>129</td>\n",
       "      <td>140</td>\n",
       "      <td>192</td>\n",
       "      <td>136</td>\n",
       "      <td>151</td>\n",
       "      <td>198</td>\n",
       "      <td>142</td>\n",
       "      <td>156</td>\n",
       "      <td>198</td>\n",
       "      <td>...</td>\n",
       "      <td>152</td>\n",
       "      <td>120</td>\n",
       "      <td>138</td>\n",
       "      <td>144</td>\n",
       "      <td>113</td>\n",
       "      <td>123</td>\n",
       "      <td>115</td>\n",
       "      <td>81</td>\n",
       "      <td>84</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>24</td>\n",
       "      <td>11</td>\n",
       "      <td>19</td>\n",
       "      <td>36</td>\n",
       "      <td>19</td>\n",
       "      <td>30</td>\n",
       "      <td>64</td>\n",
       "      <td>38</td>\n",
       "      <td>50</td>\n",
       "      <td>91</td>\n",
       "      <td>...</td>\n",
       "      <td>27</td>\n",
       "      <td>13</td>\n",
       "      <td>16</td>\n",
       "      <td>25</td>\n",
       "      <td>12</td>\n",
       "      <td>16</td>\n",
       "      <td>24</td>\n",
       "      <td>9</td>\n",
       "      <td>14</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>138</td>\n",
       "      <td>94</td>\n",
       "      <td>117</td>\n",
       "      <td>158</td>\n",
       "      <td>113</td>\n",
       "      <td>138</td>\n",
       "      <td>178</td>\n",
       "      <td>133</td>\n",
       "      <td>161</td>\n",
       "      <td>191</td>\n",
       "      <td>...</td>\n",
       "      <td>140</td>\n",
       "      <td>106</td>\n",
       "      <td>118</td>\n",
       "      <td>108</td>\n",
       "      <td>77</td>\n",
       "      <td>92</td>\n",
       "      <td>67</td>\n",
       "      <td>40</td>\n",
       "      <td>55</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 12289 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   pixel0000  pixel0001  pixel0002  pixel0003  pixel0004  pixel0005  \\\n",
       "0        191        152        194        191        153        195   \n",
       "1         24         13         23         24         14         28   \n",
       "2        185        129        140        192        136        151   \n",
       "3         24         11         19         36         19         30   \n",
       "4        138         94        117        158        113        138   \n",
       "\n",
       "   pixel0006  pixel0007  pixel0008  pixel0009  ...    pixel12279  pixel12280  \\\n",
       "0        192        149        192        194  ...           185         154   \n",
       "1         37         24         46         61  ...            26          15   \n",
       "2        198        142        156        198  ...           152         120   \n",
       "3         64         38         50         91  ...            27          13   \n",
       "4        178        133        161        191  ...           140         106   \n",
       "\n",
       "   pixel12281  pixel12282  pixel12283  pixel12284  pixel12285  pixel12286  \\\n",
       "0         180         186         156         184         182         152   \n",
       "1          29          27          15          28          24          13   \n",
       "2         138         144         113         123         115          81   \n",
       "3          16          25          12          16          24           9   \n",
       "4         118         108          77          92          67          40   \n",
       "\n",
       "   pixel12287  label  \n",
       "0         173      2  \n",
       "1          25      2  \n",
       "2          84      2  \n",
       "3          14      2  \n",
       "4          55      2  \n",
       "\n",
       "[5 rows x 12289 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "skin_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "C1psqaTj45AM"
   },
   "outputs": [],
   "source": [
    "X = skin_df.drop(\"label\", axis=1).as_matrix()\n",
    "label = skin_df[\"label\"].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "id": "1e2D7SCz5LQh",
    "outputId": "cb30d359-a2b0-4e0b-c501-6a3288c3ffb5"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((10015, 12288), (10015,))"
      ]
     },
     "execution_count": 5,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape, label.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "E3iU_vSjUIpn"
   },
   "source": [
    "## Scaling and Split Data into Train, Validation and Test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "q25NL3-X5jry"
   },
   "outputs": [],
   "source": [
    "X_mean = np.mean(X)\n",
    "X_std = np.std(X)\n",
    "\n",
    "X = (X - X_mean)/X_std"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "oGH8wMmy6Mew"
   },
   "outputs": [],
   "source": [
    "X_train_orig, X_test, y_train_orig, y_test = train_test_split(X, label, test_size=0.1,random_state=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "id": "hU0BQ6c6B1tu",
    "outputId": "2a213eec-e04a-4bef-fd97-27321013c5a7"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((9013, 12288), (1002, 12288), (9013,), (1002,))"
      ]
     },
     "execution_count": 8,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_orig.shape, X_test.shape, y_train_orig.shape, y_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "_WzFxSy1ByyY"
   },
   "outputs": [],
   "source": [
    "X_train, X_val, y_train, y_val = train_test_split(X_train_orig, y_train_orig, test_size=0.2, random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "id": "HiK1KmsMCF5B",
    "outputId": "731610d1-06ec-4df9-c9b5-867bf7a33138"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((7210, 12288), (1803, 12288), (7210,), (1803,))"
      ]
     },
     "execution_count": 10,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape, X_val.shape, y_train.shape, y_val.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "4Zl7b9mFURFJ"
   },
   "source": [
    "## Reshape the Data to Input in CNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "mdbb6U1W6mx2"
   },
   "outputs": [],
   "source": [
    "X_train = X_train.reshape(X_train.shape[0], *(64, 64, 3))\n",
    "X_val = X_val.reshape(X_val.shape[0], *(64, 64, 3))\n",
    "X_test = X_test.reshape(X_test.shape[0], *(64, 64, 3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "id": "odlh2eEz629K",
    "outputId": "98c81273-d8af-43da-8aa6-eca193a59dbd"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((7210, 64, 64, 3), (1803, 64, 64, 3), (1002, 64, 64, 3))"
      ]
     },
     "execution_count": 12,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape, X_val.shape, X_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "id": "ToBfm6FCE82c",
    "outputId": "4289e6c7-c7e4-4562-dee2-ecaa11b6815c"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(7210,)"
      ]
     },
     "execution_count": 13,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "bv1nE9AkBMXJ"
   },
   "outputs": [],
   "source": [
    "y_train = to_categorical(y_train)\n",
    "y_val = to_categorical(y_val)\n",
    "y_test = to_categorical(y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "id": "UuRfmdeYCsW5",
    "outputId": "ae8001b5-53a6-4ee9-f3c8-498150246a55"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((7210, 7), (1803, 7), (1002, 7))"
      ]
     },
     "execution_count": 15,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train.shape, y_val.shape, y_test.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "5vVaAuWtUVJU"
   },
   "source": [
    "## CNN Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "9i2prFW67mwJ"
   },
   "outputs": [],
   "source": [
    "# Our input feature map is 64x64x3: 64x64 for the image pixels, and 3 for\n",
    "# the three color channels: R, G, and B\n",
    "img_input = layers.Input(shape=(64, 64, 3))\n",
    "\n",
    "# First convolution extracts 16 filters that are 3x3\n",
    "# Convolution is followed by max-pooling layer with a 2x2 window\n",
    "x = layers.Conv2D(16, 3, activation='relu', padding='same')(img_input)\n",
    "x = layers.MaxPooling2D(2)(x)\n",
    "\n",
    "# Second convolution extracts 32 filters that are 3x3\n",
    "# Convolution is followed by max-pooling layer with a 2x2 window\n",
    "x = layers.Conv2D(32, 3, activation='relu', padding='same')(x)\n",
    "x = layers.MaxPooling2D(2)(x)\n",
    "\n",
    "# Third convolution extracts 64 filters that are 3x3\n",
    "# Convolution is followed by max-pooling layer with a 2x2 window\n",
    "x = layers.Convolution2D(64, 3, activation='relu', padding='same')(x)\n",
    "x = layers.MaxPooling2D(2)(x)\n",
    "\n",
    "# Flatten feature map to a 1-dim tensor\n",
    "x = layers.Flatten()(x)\n",
    "\n",
    "# Create a fully connected layer with ReLU activation and 512 hidden units\n",
    "x = layers.Dense(512, activation='relu')(x)\n",
    "\n",
    "# Add a dropout rate of 0.5\n",
    "x = layers.Dropout(0.5)(x)\n",
    "\n",
    "# Create output layer with a single node and sigmoid activation\n",
    "output = layers.Dense(7, activation='softmax')(x)\n",
    "\n",
    "# Configure and compile the model\n",
    "model = Model(img_input, output)\n",
    "\n",
    "optimizer = Adam(lr=0.001, beta_1=0.9, beta_2=0.999, epsilon=None, decay=0.0, amsgrad=False)\n",
    "\n",
    "model.compile(loss='categorical_crossentropy', optimizer=optimizer, metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 518
    },
    "colab_type": "code",
    "id": "JXh46jmDA9c6",
    "outputId": "290fa9f9-c12d-4179-e6b5-9126af32f189"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         (None, 64, 64, 3)         0         \n",
      "_________________________________________________________________\n",
      "conv2d (Conv2D)              (None, 64, 64, 16)        448       \n",
      "_________________________________________________________________\n",
      "max_pooling2d (MaxPooling2D) (None, 32, 32, 16)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_1 (Conv2D)            (None, 32, 32, 32)        4640      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 16, 16, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_2 (Conv2D)            (None, 16, 16, 64)        18496     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2 (None, 8, 8, 64)          0         \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 4096)              0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 512)               2097664   \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 7)                 3591      \n",
      "=================================================================\n",
      "Total params: 2,124,839\n",
      "Trainable params: 2,124,839\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "fSoYM0JEUbGT"
   },
   "source": [
    "## Define Data Generator for Data Augmentation and Learning Rate Adaptive Reduction to Prevent Overfitting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "_ITW1SFVFuHk"
   },
   "outputs": [],
   "source": [
    "train_datagen = ImageDataGenerator(rotation_range=40, width_shift_range=0.2, height_shift_range=0.2,\n",
    "                             shear_range=0.2, zoom_range=0.2, horizontal_flip=True, fill_mode='nearest')\n",
    "\n",
    "train_datagen.fit(X_train)\n",
    "\n",
    "val_datagen = ImageDataGenerator()\n",
    "val_datagen.fit(X_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "-FvlOkvw8GJ0"
   },
   "outputs": [],
   "source": [
    "learning_rate_reduction = ReduceLROnPlateau(monitor='val_acc', patience=3, verbose=1, factor=0.5, min_lr=0.00001)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1212
    },
    "colab_type": "code",
    "id": "epgliUU-_EWz",
    "outputId": "52c0a22e-7b64-40d5-ef4f-6ee75685e2af"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/30\n",
      "112/112 [==============================] - 12s 109ms/step - loss: 0.9876 - acc: 0.6599 - val_loss: 0.9136 - val_acc: 0.6855\n",
      "Epoch 2/30\n",
      "112/112 [==============================] - 11s 97ms/step - loss: 0.8612 - acc: 0.6783 - val_loss: 0.8339 - val_acc: 0.6961\n",
      "Epoch 3/30\n",
      "112/112 [==============================] - 11s 95ms/step - loss: 0.8202 - acc: 0.7005 - val_loss: 0.8119 - val_acc: 0.7011\n",
      "Epoch 4/30\n",
      "112/112 [==============================] - 11s 95ms/step - loss: 0.7888 - acc: 0.7108 - val_loss: 0.7633 - val_acc: 0.7194\n",
      "Epoch 5/30\n",
      "112/112 [==============================] - 11s 94ms/step - loss: 0.7597 - acc: 0.7208 - val_loss: 0.7293 - val_acc: 0.7399\n",
      "Epoch 6/30\n",
      "112/112 [==============================] - 11s 100ms/step - loss: 0.7340 - acc: 0.7290 - val_loss: 0.7348 - val_acc: 0.7293\n",
      "Epoch 7/30\n",
      "112/112 [==============================] - 11s 95ms/step - loss: 0.7110 - acc: 0.7350 - val_loss: 0.7122 - val_acc: 0.7349\n",
      "Epoch 8/30\n",
      "112/112 [==============================] - 11s 94ms/step - loss: 0.7041 - acc: 0.7379 - val_loss: 0.6998 - val_acc: 0.7443\n",
      "Epoch 9/30\n",
      "112/112 [==============================] - 11s 96ms/step - loss: 0.6864 - acc: 0.7462 - val_loss: 0.7064 - val_acc: 0.7321\n",
      "Epoch 10/30\n",
      "112/112 [==============================] - 11s 95ms/step - loss: 0.6867 - acc: 0.7490 - val_loss: 0.7073 - val_acc: 0.7510\n",
      "Epoch 11/30\n",
      "112/112 [==============================] - 12s 110ms/step - loss: 0.6788 - acc: 0.7436 - val_loss: 0.6646 - val_acc: 0.7499\n",
      "Epoch 12/30\n",
      "112/112 [==============================] - 12s 103ms/step - loss: 0.6614 - acc: 0.7557 - val_loss: 0.6614 - val_acc: 0.7482\n",
      "Epoch 13/30\n",
      "111/112 [============================>.] - ETA: 0s - loss: 0.6492 - acc: 0.7566\n",
      "Epoch 00013: ReduceLROnPlateau reducing learning rate to 0.0005000000237487257.\n",
      "112/112 [==============================] - 11s 96ms/step - loss: 0.6499 - acc: 0.7564 - val_loss: 0.6770 - val_acc: 0.7510\n",
      "Epoch 14/30\n",
      "112/112 [==============================] - 14s 127ms/step - loss: 0.6237 - acc: 0.7669 - val_loss: 0.6453 - val_acc: 0.7560\n",
      "Epoch 15/30\n",
      "112/112 [==============================] - 17s 150ms/step - loss: 0.6160 - acc: 0.7717 - val_loss: 0.6691 - val_acc: 0.7560\n",
      "Epoch 16/30\n",
      "112/112 [==============================] - 17s 151ms/step - loss: 0.6173 - acc: 0.7679 - val_loss: 0.6607 - val_acc: 0.7543\n",
      "Epoch 17/30\n",
      "112/112 [==============================] - 17s 150ms/step - loss: 0.6022 - acc: 0.7772 - val_loss: 0.6590 - val_acc: 0.7565\n",
      "Epoch 18/30\n",
      "112/112 [==============================] - 12s 111ms/step - loss: 0.6010 - acc: 0.7746 - val_loss: 0.6218 - val_acc: 0.7682\n",
      "Epoch 19/30\n",
      "112/112 [==============================] - 11s 95ms/step - loss: 0.6021 - acc: 0.7732 - val_loss: 0.6291 - val_acc: 0.7593\n",
      "Epoch 20/30\n",
      "112/112 [==============================] - 11s 95ms/step - loss: 0.5773 - acc: 0.7809 - val_loss: 0.6372 - val_acc: 0.7648\n",
      "Epoch 21/30\n",
      "111/112 [============================>.] - ETA: 0s - loss: 0.5886 - acc: 0.7787\n",
      "Epoch 00021: ReduceLROnPlateau reducing learning rate to 0.0002500000118743628.\n",
      "112/112 [==============================] - 11s 95ms/step - loss: 0.5885 - acc: 0.7791 - val_loss: 0.6596 - val_acc: 0.7526\n",
      "Epoch 22/30\n",
      "112/112 [==============================] - 11s 95ms/step - loss: 0.5776 - acc: 0.7814 - val_loss: 0.6223 - val_acc: 0.7720\n",
      "Epoch 23/30\n",
      "112/112 [==============================] - 11s 95ms/step - loss: 0.5593 - acc: 0.7922 - val_loss: 0.6173 - val_acc: 0.7743\n",
      "Epoch 24/30\n",
      "112/112 [==============================] - 11s 95ms/step - loss: 0.5522 - acc: 0.7934 - val_loss: 0.6154 - val_acc: 0.7726\n",
      "Epoch 25/30\n",
      "112/112 [==============================] - 11s 95ms/step - loss: 0.5600 - acc: 0.7949 - val_loss: 0.6367 - val_acc: 0.7582\n",
      "Epoch 26/30\n",
      "111/112 [============================>.] - ETA: 0s - loss: 0.5487 - acc: 0.7960\n",
      "Epoch 00026: ReduceLROnPlateau reducing learning rate to 0.0001250000059371814.\n",
      "112/112 [==============================] - 11s 95ms/step - loss: 0.5505 - acc: 0.7945 - val_loss: 0.6147 - val_acc: 0.7598\n",
      "Epoch 27/30\n",
      "112/112 [==============================] - 11s 95ms/step - loss: 0.5483 - acc: 0.7887 - val_loss: 0.6117 - val_acc: 0.7787\n",
      "Epoch 28/30\n",
      "112/112 [==============================] - 11s 96ms/step - loss: 0.5421 - acc: 0.7949 - val_loss: 0.6157 - val_acc: 0.7743\n",
      "Epoch 29/30\n",
      "112/112 [==============================] - 11s 95ms/step - loss: 0.5543 - acc: 0.7921 - val_loss: 0.6147 - val_acc: 0.7709\n",
      "Epoch 30/30\n",
      "111/112 [============================>.] - ETA: 0s - loss: 0.5271 - acc: 0.7984\n",
      "Epoch 00030: ReduceLROnPlateau reducing learning rate to 6.25000029685907e-05.\n",
      "112/112 [==============================] - 11s 97ms/step - loss: 0.5263 - acc: 0.7985 - val_loss: 0.6034 - val_acc: 0.7748\n"
     ]
    }
   ],
   "source": [
    "batch_size = 64\n",
    "epochs = 30\n",
    "history = model.fit_generator(train_datagen.flow(X_train,y_train, batch_size=batch_size),\n",
    "                              epochs = epochs, validation_data = val_datagen.flow(X_val, y_val),\n",
    "                              verbose = 1, steps_per_epoch=(X_train.shape[0] // batch_size),\n",
    "                              callbacks=[learning_rate_reduction])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 86
    },
    "colab_type": "code",
    "id": "pzmhcKchK3fQ",
    "outputId": "255f9bf8-d123-435d-809f-219c47922e5f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1002/1002 [==============================] - 0s 216us/step\n",
      "1803/1803 [==============================] - 0s 187us/step\n",
      "Validation: accuracy = 0.774820  ;  loss_v = 0.603359\n",
      "Test: accuracy = 0.765469  ;  loss = 0.646671\n"
     ]
    }
   ],
   "source": [
    "loss_test, acc_test = model.evaluate(X_test, y_test, verbose=1)\n",
    "loss_val, acc_val = model.evaluate(X_val, y_val, verbose=1)\n",
    "print(\"Validation: accuracy = %f  ;  loss_v = %f\" % (acc_val, loss_val))\n",
    "print(\"Test: accuracy = %f  ;  loss = %f\" % (acc_test, loss_test))\n",
    "model.save(\"model.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 560
    },
    "colab_type": "code",
    "id": "z-XsupKP_orF",
    "outputId": "69099e17-bae9-4886-c0e9-50a0c20be04a"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Training and validation loss')"
      ]
     },
     "execution_count": 22,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXgAAAEHCAYAAACk6V2yAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb4aa1173c8>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXMAAAEHCAYAAABcCaZFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb4ac144898>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Retrieve a list of accuracy results on training and test data\n",
    "# sets for each training epoch\n",
    "acc = history.history['acc']\n",
    "val_acc = history.history['val_acc']\n",
    "\n",
    "# Retrieve a list of list results on training and test data\n",
    "# sets for each training epoch\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "# Get number of epochs\n",
    "epochs = range(len(acc))\n",
    "\n",
    "# Plot training and validation accuracy per epoch\n",
    "plt.plot(epochs, acc)\n",
    "plt.plot(epochs, val_acc)\n",
    "plt.title('Training and validation accuracy')\n",
    "\n",
    "plt.figure()\n",
    "\n",
    "# Plot training and validation loss per epoch\n",
    "plt.plot(epochs, loss)\n",
    "plt.plot(epochs, val_loss)\n",
    "plt.title('Training and validation loss')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "F3iwk30TLRbu"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "name": "Baseline_CNN.ipynb",
   "provenance": [],
   "version": "0.3.2"
  },
  "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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}