Predicting continuous variables with Images.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import all the necessary files!\n",
    "import os\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras import layers\n",
    "from tensorflow.keras import Model\n",
    "\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "\n",
    "import pandas as pd\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Hyperparameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss_used = 'mean_squared_error'\n",
    "metrics = ['accuracy']\n",
    "batch_size=64\n",
    "epochs = 10"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Saving locations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.applications import VGG16\n",
    "\n",
    "pre_trained_model = VGG16(input_shape = (150, 150, 3), \n",
    "                                include_top = False, \n",
    "                                weights = 'imagenet')\n",
    "\n",
    "# Make all the layers in the pre-trained model non-trainable\n",
    "for layer in pre_trained_model.layers:\n",
    "    layer.trainable = False"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Transfer Learning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.applications import VGG16\n",
    "\n",
    "pre_trained_model = VGG16(input_shape = (150, 150, 3), \n",
    "                                include_top = False, \n",
    "                                weights = 'imagenet')\n",
    "\n",
    "# Make all the layers in the pre-trained model non-trainable\n",
    "for layer in pre_trained_model.layers:\n",
    "    layer.trainable = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "if not os.path.exists(folder):\n",
    "    os.makedirs(folder)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "last layer output shape:  (None, 9, 9, 512)\n"
     ]
    }
   ],
   "source": [
    "last_layer = pre_trained_model.get_layer('block5_conv3')\n",
    "print('last layer output shape: ', last_layer.output_shape)\n",
    "last_output = last_layer.output "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_2\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_2 (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",
      "flatten_2 (Flatten)          (None, 41472)             0         \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 1024)              42468352  \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 1024)              0         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 1)                 1025      \n",
      "=================================================================\n",
      "Total params: 57,184,065\n",
      "Trainable params: 42,469,377\n",
      "Non-trainable params: 14,714,688\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from tensorflow.keras.optimizers import RMSprop\n",
    "\n",
    "# Flatten the output layer to 1 dimension\n",
    "x = layers.Flatten()(last_output)\n",
    "# Add a fully connected layer with 1,024 hidden units and ReLU activation\n",
    "x = layers.Dense(1024, activation='relu')(x)\n",
    "# Add a dropout rate of 0.2\n",
    "x = layers.Dropout(0.2)(x)                  \n",
    "# Add a final sigmoid layer for classification\n",
    "x = layers.Dense  (1, activation='linear')(x)           \n",
    "\n",
    "model = Model( pre_trained_model.input, x) \n",
    "\n",
    "model.compile(optimizer = RMSprop(lr=0.0001), \n",
    "              loss = loss_used,\n",
    "              metrics = metrics) \n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Loading the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "all_data_df = pd.read_csv('train_set.csv')\n",
    "all_data_df_use = all_data_df[['img_path','ROS']]\n",
    "train_df=all_data_df_use[all_data_df.img_path.notnull()]\n",
    "\n",
    "result = pd.read_csv('test_set.csv')\n",
    "test_df = result[['img_path','ROS']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "scrolled": true
   },
   "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>img_path</th>\n",
       "      <th>ROS</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.109929</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.655773</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.245059</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.389480</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.033828</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.047089</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.306019</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.319010</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.285231</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.571916</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.469944</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.030314</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.076493</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.051188</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.069174</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.112128</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.419816</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.474496</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.283748</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.922543</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.745688</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.855777</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.589562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.196485</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.149122</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.986993</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.516696</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.275618</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.270570</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13597</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.573457</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13598</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.380295</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13599</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>1.040454</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13600</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.660623</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13601</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>1.108716</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13602</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>1.245987</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13603</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.695459</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13604</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.671265</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13605</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.642353</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13606</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.647892</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13607</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>7.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13608</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>1.327331</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13609</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.935644</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13610</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.189658</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13611</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.112748</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13612</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.696133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13613</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.432353</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13614</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.705959</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13615</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>1.108216</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13616</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.603504</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13617</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.844706</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13618</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.530621</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13619</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.804653</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13620</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.335888</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13621</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.235644</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13622</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>7.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13623</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>1.241880</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13624</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.478075</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13625</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>0.625591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13626</th>\n",
       "      <td>/home/vigneshwar/sri/square_cropped_images/ima...</td>\n",
       "      <td>2.236073</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>13627 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                img_path       ROS\n",
       "0      /home/vigneshwar/sri/square_cropped_images/ima...  0.109929\n",
       "1      /home/vigneshwar/sri/square_cropped_images/ima...  0.655773\n",
       "2      /home/vigneshwar/sri/square_cropped_images/ima...  0.245059\n",
       "3      /home/vigneshwar/sri/square_cropped_images/ima...  0.000000\n",
       "4      /home/vigneshwar/sri/square_cropped_images/ima...  0.389480\n",
       "5      /home/vigneshwar/sri/square_cropped_images/ima...  0.033828\n",
       "6      /home/vigneshwar/sri/square_cropped_images/ima...  0.047089\n",
       "7      /home/vigneshwar/sri/square_cropped_images/ima...  0.306019\n",
       "8      /home/vigneshwar/sri/square_cropped_images/ima...  0.319010\n",
       "9      /home/vigneshwar/sri/square_cropped_images/ima...  0.285231\n",
       "10     /home/vigneshwar/sri/square_cropped_images/ima...  0.571916\n",
       "11     /home/vigneshwar/sri/square_cropped_images/ima...  0.469944\n",
       "12     /home/vigneshwar/sri/square_cropped_images/ima...  0.030314\n",
       "13     /home/vigneshwar/sri/square_cropped_images/ima...  0.076493\n",
       "14     /home/vigneshwar/sri/square_cropped_images/ima...  0.051188\n",
       "15     /home/vigneshwar/sri/square_cropped_images/ima...  0.069174\n",
       "16     /home/vigneshwar/sri/square_cropped_images/ima...  0.112128\n",
       "17     /home/vigneshwar/sri/square_cropped_images/ima...  0.419816\n",
       "18     /home/vigneshwar/sri/square_cropped_images/ima...  0.474496\n",
       "19     /home/vigneshwar/sri/square_cropped_images/ima...  0.283748\n",
       "20     /home/vigneshwar/sri/square_cropped_images/ima...  0.922543\n",
       "21     /home/vigneshwar/sri/square_cropped_images/ima...  0.745688\n",
       "22     /home/vigneshwar/sri/square_cropped_images/ima...  0.855777\n",
       "23     /home/vigneshwar/sri/square_cropped_images/ima...  0.589562\n",
       "24     /home/vigneshwar/sri/square_cropped_images/ima...  0.196485\n",
       "25     /home/vigneshwar/sri/square_cropped_images/ima...  0.149122\n",
       "26     /home/vigneshwar/sri/square_cropped_images/ima...  0.986993\n",
       "27     /home/vigneshwar/sri/square_cropped_images/ima...  0.516696\n",
       "28     /home/vigneshwar/sri/square_cropped_images/ima...  0.275618\n",
       "29     /home/vigneshwar/sri/square_cropped_images/ima...  0.270570\n",
       "...                                                  ...       ...\n",
       "13597  /home/vigneshwar/sri/square_cropped_images/ima...  0.573457\n",
       "13598  /home/vigneshwar/sri/square_cropped_images/ima...  0.380295\n",
       "13599  /home/vigneshwar/sri/square_cropped_images/ima...  1.040454\n",
       "13600  /home/vigneshwar/sri/square_cropped_images/ima...  0.660623\n",
       "13601  /home/vigneshwar/sri/square_cropped_images/ima...  1.108716\n",
       "13602  /home/vigneshwar/sri/square_cropped_images/ima...  1.245987\n",
       "13603  /home/vigneshwar/sri/square_cropped_images/ima...  0.695459\n",
       "13604  /home/vigneshwar/sri/square_cropped_images/ima...  0.671265\n",
       "13605  /home/vigneshwar/sri/square_cropped_images/ima...  0.642353\n",
       "13606  /home/vigneshwar/sri/square_cropped_images/ima...  0.647892\n",
       "13607  /home/vigneshwar/sri/square_cropped_images/ima...  7.000000\n",
       "13608  /home/vigneshwar/sri/square_cropped_images/ima...  1.327331\n",
       "13609  /home/vigneshwar/sri/square_cropped_images/ima...  0.935644\n",
       "13610  /home/vigneshwar/sri/square_cropped_images/ima...  0.189658\n",
       "13611  /home/vigneshwar/sri/square_cropped_images/ima...  0.112748\n",
       "13612  /home/vigneshwar/sri/square_cropped_images/ima...  0.696133\n",
       "13613  /home/vigneshwar/sri/square_cropped_images/ima...  0.432353\n",
       "13614  /home/vigneshwar/sri/square_cropped_images/ima...  0.705959\n",
       "13615  /home/vigneshwar/sri/square_cropped_images/ima...  1.108216\n",
       "13616  /home/vigneshwar/sri/square_cropped_images/ima...  0.603504\n",
       "13617  /home/vigneshwar/sri/square_cropped_images/ima...  0.844706\n",
       "13618  /home/vigneshwar/sri/square_cropped_images/ima...  0.530621\n",
       "13619  /home/vigneshwar/sri/square_cropped_images/ima...  0.804653\n",
       "13620  /home/vigneshwar/sri/square_cropped_images/ima...  0.335888\n",
       "13621  /home/vigneshwar/sri/square_cropped_images/ima...  0.235644\n",
       "13622  /home/vigneshwar/sri/square_cropped_images/ima...  7.000000\n",
       "13623  /home/vigneshwar/sri/square_cropped_images/ima...  1.241880\n",
       "13624  /home/vigneshwar/sri/square_cropped_images/ima...  0.478075\n",
       "13625  /home/vigneshwar/sri/square_cropped_images/ima...  0.625591\n",
       "13626  /home/vigneshwar/sri/square_cropped_images/ima...  2.236073\n",
       "\n",
       "[13627 rows x 2 columns]"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Generators"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 10902 images.\n",
      "Found 2725 images.\n",
      "Found 2617 images.\n"
     ]
    }
   ],
   "source": [
    "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",
    "            validation_split=0.20,\n",
    "            fill_mode='nearest')\n",
    "\n",
    "train_generator=datagen.flow_from_dataframe(\n",
    "            dataframe=train_df,\n",
    "            directory=None,\n",
    "            x_col=\"img_path\",\n",
    "            y_col=\"ROS\",\n",
    "            subset=\"training\",\n",
    "            batch_size=32,\n",
    "            seed=42,\n",
    "            shuffle=True,\n",
    "            class_mode=\"other\",\n",
    "            target_size=(150,150))\n",
    "\n",
    "valid_generator=datagen.flow_from_dataframe(\n",
    "            dataframe=train_df,\n",
    "            directory=None,\n",
    "            x_col=\"img_path\",\n",
    "            y_col=\"ROS\",\n",
    "            subset=\"validation\",\n",
    "            batch_size=32,\n",
    "            seed=42,\n",
    "            class_mode=\"other\",\n",
    "            target_size=(150,150))\n",
    "\n",
    "test_datagen=ImageDataGenerator(rescale=1./255.)\n",
    "\n",
    "test_generator=test_datagen.flow_from_dataframe(\n",
    "            dataframe=test_df,\n",
    "            directory=None,\n",
    "            x_col=\"img_path\",\n",
    "            y_col=None,\n",
    "            batch_size=batch_size,\n",
    "            seed=42,\n",
    "            class_mode=None,\n",
    "            target_size=(150,150))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define a Callback class that stops training once accuracy reaches 99.9%\n",
    "class myCallback(tf.keras.callbacks.Callback):\n",
    "    def on_epoch_end(self, epoch, logs={}):\n",
    "        if(logs.get('acc')>0.999):\n",
    "            print(\"\\nReached 99.9% accuracy so cancelling training!\")\n",
    "            self.model.stop_training = True\n",
    "            "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "341/341 [==============================] - 1586s 5s/step - loss: 1.2512 - acc: 0.0085 - val_loss: 0.2069 - val_acc: 0.0011\n",
      "Epoch 2/10\n",
      "341/341 [==============================] - 1578s 5s/step - loss: 0.1871 - acc: 0.0072 - val_loss: 0.2088 - val_acc: 0.0022\n",
      "Epoch 3/10\n",
      "341/341 [==============================] - 1577s 5s/step - loss: 0.1840 - acc: 0.0083 - val_loss: 0.2014 - val_acc: 0.0000e+00\n",
      "Epoch 4/10\n",
      "341/341 [==============================] - 1580s 5s/step - loss: 0.1847 - acc: 0.0067 - val_loss: 0.1991 - val_acc: 0.0000e+00\n",
      "Epoch 5/10\n",
      "341/341 [==============================] - 1579s 5s/step - loss: 0.1821 - acc: 0.0078 - val_loss: 0.1965 - val_acc: 0.0000e+00\n",
      "Epoch 6/10\n",
      "341/341 [==============================] - 1575s 5s/step - loss: 0.1828 - acc: 0.0068 - val_loss: 0.2101 - val_acc: 3.6697e-04\n",
      "Epoch 7/10\n",
      "341/341 [==============================] - 1571s 5s/step - loss: 0.1816 - acc: 0.0069 - val_loss: 0.2030 - val_acc: 3.6697e-04\n",
      "Epoch 8/10\n",
      "341/341 [==============================] - 1576s 5s/step - loss: 0.1813 - acc: 0.0073 - val_loss: 0.1968 - val_acc: 3.6697e-04\n",
      "Epoch 9/10\n",
      "341/341 [==============================] - 1576s 5s/step - loss: 0.1804 - acc: 0.0072 - val_loss: 0.2085 - val_acc: 0.0015\n",
      "Epoch 10/10\n",
      "341/341 [==============================] - 1578s 5s/step - loss: 0.1806 - acc: 0.0069 - val_loss: 0.1977 - val_acc: 0.0000e+00\n"
     ]
    }
   ],
   "source": [
    "callbacks = myCallback()\n",
    "history = model.fit_generator(train_generator,\n",
    "                              validation_data=valid_generator,\n",
    "                              epochs=epochs,\n",
    "                              verbose=1,\n",
    "                              callbacks=[callbacks])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Saving and Ploting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save(model_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "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": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\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",
    "fig1 = plt.gcf()\n",
    "plt.plot(epochs, acc, 'r', label='Training accuracy')\n",
    "plt.plot(epochs, val_acc, 'b', label='Validation accuracy')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend(loc=0)\n",
    "fig1.savefig(accuracy_plot)\n",
    "\n",
    "plt.figure()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig1 = plt.gcf()\n",
    "plt.plot(epochs, loss, 'r', label='Training loss')\n",
    "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend(loc=0)\n",
    "plt.savefig(loss_plot)\n",
    "\n",
    "plt.figure()\n",
    "#plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "41/41 [==============================] - 284s 7s/step\n"
     ]
    }
   ],
   "source": [
    "scores = model.predict_generator(test_generator,verbose=1) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.6285243 ],\n",
       "       [0.5889201 ],\n",
       "       [0.5647993 ],\n",
       "       ...,\n",
       "       [0.60544753],\n",
       "       [0.4324256 ],\n",
       "       [0.6397518 ]], dtype=float32)"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "result['Predicted'] = scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "result['Mape'] = abs( (result['ROS'] - result['Predicted'] ) / result['ROS']) * 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "result.to_csv(output_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "MAPE = result[(result['Mape'] <= 100) & (result['Mape'] >= 0)]['Mape'].mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "34.47155357680085"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "MAPE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "MAPE_RAW = result[~result.isin([np.nan, np.inf, -np.inf]).any(1)]['Mape'].mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "54.05876906315976"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "MAPE_RAW"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import mean_squared_error\n",
    "MSE = mean_squared_error(y_true = result['ROS'], y_pred = result['Predicted'])\n",
    "RMSE = MSE**(0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.18827737825725016\n",
      "0.4339094125013309\n"
     ]
    }
   ],
   "source": [
    "print(MSE)\n",
    "print(MSE**(0.5))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "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"
    }
   ],
   "source": [
    "fig1 = plt.gcf()\n",
    "pd.value_counts(result['Predicted'].round(1)).sort_index().plot(kind=\"bar\")\n",
    "fig1.savefig(ROS_pred)\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}