Kaggle Notebook for 11th Dataset(Xception Model-90% Accuracy).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### This is the Notebook that I found on Kaggle that works on the 11th dataset and achieves good accuracy on it.   It uses Xception model.\n",
    "\n",
    "- *Training Accuracy= 97% and Test Accuracy= >90%* \n",
    "\n",
    "- Try and Run it on Kaggle. (Link)[https://www.kaggle.com/code/mostafaabla/garbage-classification-keras-transfer-learning]."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "papermill": {
     "duration": 0.025054,
     "end_time": "2021-02-11T20:31:04.915467",
     "exception": false,
     "start_time": "2021-02-11T20:31:04.890413",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Garbage Classification using keras and transfer learning"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.022508,
     "end_time": "2021-02-11T20:31:04.964163",
     "exception": false,
     "start_time": "2021-02-11T20:31:04.941655",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "This kernel shows how to classify Garbage images into 12 different classes using transfer learning.If you find this kernel useful please upvote it!\n",
    "\n",
    "Transfer learning means that instead of your model learning everything from scratch, it uses another model that was trained on a similar problem, so that you can \"transfer\" the learned \"knowledge\" of the pretrained model to your model, and then learn some new features.\n",
    "\n",
    "The ImageNet Data set is huge data set consisting of more than 14 million images from more than 22,000 different categories, here we are using a smaller version of it which has 1000 different categories.\n",
    "\n",
    "In this kernel we use an xception model which is pretrained on the ImageNet dataset and then build some layers on top of it to be able to classify the garbage images.\n",
    "\n",
    "Transfer learning makes sense here because the ImageNet data set has a much larger number of images (14 million) than the Garbage Classification data set (around 15,500 image). This increases the speed of training for our model and the accuracy of our predictions."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.022726,
     "end_time": "2021-02-11T20:31:05.010149",
     "exception": false,
     "start_time": "2021-02-11T20:31:04.987423",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Import Required Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-11T20:31:05.069000Z",
     "iopub.status.busy": "2021-02-11T20:31:05.068147Z",
     "iopub.status.idle": "2021-02-11T20:31:11.144375Z",
     "shell.execute_reply": "2021-02-11T20:31:11.145113Z"
    },
    "papermill": {
     "duration": 6.111744,
     "end_time": "2021-02-11T20:31:11.145317",
     "exception": false,
     "start_time": "2021-02-11T20:31:05.033573",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "setup successful!\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd \n",
    "import random\n",
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import keras.applications.xception as xception\n",
    "import zipfile\n",
    "import sys\n",
    "import time\n",
    "import tensorflow.keras as keras\n",
    "import tensorflow as tf\n",
    "import re\n",
    "\n",
    "from PIL import Image\n",
    "from keras.layers import Input, Conv2D, Dense, Flatten, MaxPooling2D, Input, GlobalAveragePooling2D\n",
    "from keras.layers.experimental.preprocessing import Normalization\n",
    "from keras.models import Model, Sequential\n",
    "from keras.preprocessing import image\n",
    "from keras.utils import to_categorical\n",
    "from keras.layers import Lambda\n",
    "from keras.callbacks import EarlyStopping\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import classification_report\n",
    "\n",
    "print('setup successful!')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.024309,
     "end_time": "2021-02-11T20:31:11.195109",
     "exception": false,
     "start_time": "2021-02-11T20:31:11.170800",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Define Constants"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-11T20:31:11.251089Z",
     "iopub.status.busy": "2021-02-11T20:31:11.250180Z",
     "iopub.status.idle": "2021-02-11T20:31:11.253976Z",
     "shell.execute_reply": "2021-02-11T20:31:11.254625Z"
    },
    "papermill": {
     "duration": 0.035925,
     "end_time": "2021-02-11T20:31:11.254781",
     "exception": false,
     "start_time": "2021-02-11T20:31:11.218856",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "defining constants successful!\n"
     ]
    }
   ],
   "source": [
    "# Increasing the image size didn't result in increasing the training accuracy\n",
    "IMAGE_WIDTH = 320    \n",
    "IMAGE_HEIGHT = 320\n",
    "IMAGE_SIZE=(IMAGE_WIDTH, IMAGE_HEIGHT)\n",
    "IMAGE_CHANNELS = 3\n",
    "\n",
    "\n",
    "# Path where our data is located\n",
    "base_path = \"../input/garbage-classification/garbage_classification/\"\n",
    "\n",
    "# Dictionary to save our 12 classes\n",
    "categories = {0: 'paper', 1: 'cardboard', 2: 'plastic', 3: 'metal', 4: 'trash', 5: 'battery',\n",
    "              6: 'shoes', 7: 'clothes', 8: 'green-glass', 9: 'brown-glass', 10: 'white-glass',\n",
    "              11: 'biological'}\n",
    "\n",
    "print('defining constants successful!')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.023065,
     "end_time": "2021-02-11T20:31:11.302078",
     "exception": false,
     "start_time": "2021-02-11T20:31:11.279013",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Create DataFrame"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.023671,
     "end_time": "2021-02-11T20:31:11.349827",
     "exception": false,
     "start_time": "2021-02-11T20:31:11.326156",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "We want to create a data frame that has in one column the filenames of all our images and in the other column the corresponding category. \n",
    "We Open the directories in the dataset one by one, save the filenames in the filenames_list and add the corresponding category in the categories_list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-11T20:31:11.408728Z",
     "iopub.status.busy": "2021-02-11T20:31:11.408070Z",
     "iopub.status.idle": "2021-02-11T20:31:15.523520Z",
     "shell.execute_reply": "2021-02-11T20:31:15.524276Z"
    },
    "papermill": {
     "duration": 4.150669,
     "end_time": "2021-02-11T20:31:15.524465",
     "exception": false,
     "start_time": "2021-02-11T20:31:11.373796",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number of elements =  15515\n"
     ]
    }
   ],
   "source": [
    "# Add class name prefix to filename. So for example \"/paper104.jpg\" become \"paper/paper104.jpg\"\n",
    "def add_class_name_prefix(df, col_name):\n",
    "    df[col_name] = df[col_name].apply(lambda x: x[:re.search(\"\\d\",x).start()] + '/' + x)\n",
    "    return df\n",
    "\n",
    "# list conatining all the filenames in the dataset\n",
    "filenames_list = []\n",
    "# list to store the corresponding category, note that each folder of the dataset has one class of data\n",
    "categories_list = []\n",
    "\n",
    "for category in categories:\n",
    "    filenames = os.listdir(base_path + categories[category])\n",
    "    \n",
    "    filenames_list = filenames_list  +filenames\n",
    "    categories_list = categories_list + [category] * len(filenames)\n",
    "    \n",
    "df = pd.DataFrame({\n",
    "    'filename': filenames_list,\n",
    "    'category': categories_list\n",
    "})\n",
    "\n",
    "df = add_class_name_prefix(df, 'filename')\n",
    "\n",
    "# Shuffle the dataframe\n",
    "df = df.sample(frac=1).reset_index(drop=True)\n",
    "\n",
    "print('number of elements = ' , len(df))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-11T20:31:15.587866Z",
     "iopub.status.busy": "2021-02-11T20:31:15.587015Z",
     "iopub.status.idle": "2021-02-11T20:31:15.595997Z",
     "shell.execute_reply": "2021-02-11T20:31:15.595466Z"
    },
    "papermill": {
     "duration": 0.045722,
     "end_time": "2021-02-11T20:31:15.596102",
     "exception": false,
     "start_time": "2021-02-11T20:31:15.550380",
     "status": "completed"
    },
    "tags": []
   },
   "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>filename</th>\n",
       "      <th>category</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>battery/battery460.jpg</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>clothes/clothes1487.jpg</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>plastic/plastic839.jpg</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>battery/battery168.jpg</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>battery/battery876.jpg</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  filename  category\n",
       "0   battery/battery460.jpg         5\n",
       "1  clothes/clothes1487.jpg         7\n",
       "2   plastic/plastic839.jpg         2\n",
       "3   battery/battery168.jpg         5\n",
       "4   battery/battery876.jpg         5"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-11T20:31:15.653252Z",
     "iopub.status.busy": "2021-02-11T20:31:15.652557Z",
     "iopub.status.idle": "2021-02-11T20:31:15.969722Z",
     "shell.execute_reply": "2021-02-11T20:31:15.970268Z"
    },
    "papermill": {
     "duration": 0.349301,
     "end_time": "2021-02-11T20:31:15.970432",
     "exception": false,
     "start_time": "2021-02-11T20:31:15.621131",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "clothes/clothes4547.jpg\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fe3252ee5d0>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# see sample image, you can run the same cell again to get a different image\n",
    "random_row = random.randint(0, len(df)-1)\n",
    "sample = df.iloc[random_row]\n",
    "randomimage = image.load_img(base_path +sample['filename'])\n",
    "print(sample['filename'])\n",
    "plt.imshow(randomimage)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.029738,
     "end_time": "2021-02-11T20:31:16.030617",
     "exception": false,
     "start_time": "2021-02-11T20:31:16.000879",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Viusalize the Categories Distribution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-11T20:31:16.110777Z",
     "iopub.status.busy": "2021-02-11T20:31:16.109933Z",
     "iopub.status.idle": "2021-02-11T20:31:16.311827Z",
     "shell.execute_reply": "2021-02-11T20:31:16.312375Z"
    },
    "papermill": {
     "duration": 0.251652,
     "end_time": "2021-02-11T20:31:16.312553",
     "exception": false,
     "start_time": "2021-02-11T20:31:16.060901",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_visualization = df.copy()\n",
    "# Change the catgegories from numbers to names\n",
    "df_visualization['category'] = df_visualization['category'].apply(lambda x:categories[x] )\n",
    "\n",
    "df_visualization['category'].value_counts().plot.bar(x = 'count', y = 'category' )\n",
    "\n",
    "plt.xlabel(\"Garbage Classes\", labelpad=14)\n",
    "plt.ylabel(\"Images Count\", labelpad=14)\n",
    "plt.title(\"Count of images per class\", y=1.02);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.033412,
     "end_time": "2021-02-11T20:31:16.379075",
     "exception": false,
     "start_time": "2021-02-11T20:31:16.345663",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Create the model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.03117,
     "end_time": "2021-02-11T20:31:16.442381",
     "exception": false,
     "start_time": "2021-02-11T20:31:16.411211",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "The steps are:\n",
    "1. Create an xception model without the last layer and load the ImageNet pretrained weights\n",
    "2. Add a pre-processing layer\n",
    "3. Add a pooling layer followed by a softmax layer at the end"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-11T20:31:16.516166Z",
     "iopub.status.busy": "2021-02-11T20:31:16.515354Z",
     "iopub.status.idle": "2021-02-11T20:31:23.983771Z",
     "shell.execute_reply": "2021-02-11T20:31:23.983134Z"
    },
    "papermill": {
     "duration": 7.509921,
     "end_time": "2021-02-11T20:31:23.983900",
     "exception": false,
     "start_time": "2021-02-11T20:31:16.473979",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "lambda (Lambda)              (None, 320, 320, 3)       0         \n",
      "_________________________________________________________________\n",
      "xception (Functional)        (None, 10, 10, 2048)      20861480  \n",
      "_________________________________________________________________\n",
      "global_average_pooling2d (Gl (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 12)                24588     \n",
      "=================================================================\n",
      "Total params: 20,886,068\n",
      "Trainable params: 24,588\n",
      "Non-trainable params: 20,861,480\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Conv2D, MaxPooling2D, Dropout, Flatten, Dense, Activation, BatchNormalization\n",
    "import keras.applications.xception as xception\n",
    "\n",
    "xception_layer = xception.Xception(include_top = False, input_shape = (IMAGE_WIDTH, IMAGE_HEIGHT,IMAGE_CHANNELS),\n",
    "                       weights = '../input/xception/xception_weights_tf_dim_ordering_tf_kernels_notop.h5')\n",
    "\n",
    "# We don't want to train the imported weights\n",
    "xception_layer.trainable = False\n",
    "\n",
    "\n",
    "model = Sequential()\n",
    "model.add(keras.Input(shape=(IMAGE_WIDTH, IMAGE_HEIGHT, IMAGE_CHANNELS)))\n",
    "\n",
    "#create a custom layer to apply the preprocessing\n",
    "def xception_preprocessing(img):\n",
    "  return xception.preprocess_input(img)\n",
    "\n",
    "model.add(Lambda(xception_preprocessing))\n",
    "\n",
    "model.add(xception_layer)\n",
    "model.add(tf.keras.layers.GlobalAveragePooling2D())\n",
    "model.add(Dense(len(categories), activation='softmax')) \n",
    "\n",
    "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['categorical_accuracy'])\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.033601,
     "end_time": "2021-02-11T20:31:24.051182",
     "exception": false,
     "start_time": "2021-02-11T20:31:24.017581",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "We will use the EarlyStopping call back to stop our training if the validation_accuray is not improving for a certain number of epochs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-11T20:31:24.163989Z",
     "iopub.status.busy": "2021-02-11T20:31:24.163161Z",
     "iopub.status.idle": "2021-02-11T20:31:24.168163Z",
     "shell.execute_reply": "2021-02-11T20:31:24.169006Z"
    },
    "papermill": {
     "duration": 0.045082,
     "end_time": "2021-02-11T20:31:24.169191",
     "exception": false,
     "start_time": "2021-02-11T20:31:24.124109",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "call back defined!\n"
     ]
    }
   ],
   "source": [
    "early_stop = EarlyStopping(patience = 2, verbose = 1, monitor='val_categorical_accuracy' , mode='max', min_delta=0.001, restore_best_weights = True)\n",
    "\n",
    "callbacks = [early_stop]\n",
    "\n",
    "print('call back defined!')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.033573,
     "end_time": "2021-02-11T20:31:24.237523",
     "exception": false,
     "start_time": "2021-02-11T20:31:24.203950",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Split the Data Set"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.032922,
     "end_time": "2021-02-11T20:31:24.304532",
     "exception": false,
     "start_time": "2021-02-11T20:31:24.271610",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "We split the training set into three separate sets:\n",
    "\n",
    "1. **The training set:** used to train our model.\n",
    "1. **The validation set**: used to double check that our model is not overfitting the training set, i.e. it can also generalise to other data other than the train data\n",
    "1. **The Test set:** Used to estimate the accuracy of the model on new data other than the ones the model used for training\n",
    "For a competition  or for some other cases, you can split the data only to training and validation sets in order to achieve the highest  possible accuracy, without the need to properly estimate how accurate the model really is.\n",
    "\n",
    "We split the data set as follows: 80% train set, 10% cross_validation set, and 10% test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-11T20:31:24.383122Z",
     "iopub.status.busy": "2021-02-11T20:31:24.381978Z",
     "iopub.status.idle": "2021-02-11T20:31:24.399106Z",
     "shell.execute_reply": "2021-02-11T20:31:24.399844Z"
    },
    "papermill": {
     "duration": 0.061194,
     "end_time": "2021-02-11T20:31:24.400027",
     "exception": false,
     "start_time": "2021-02-11T20:31:24.338833",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train size =  1551 validate size =  1551 test size =  1552\n"
     ]
    }
   ],
   "source": [
    "#Change the categories from numbers to names\n",
    "df[\"category\"] = df[\"category\"].replace(categories) \n",
    "\n",
    "# We first split the data into two sets and then split the validate_df to two sets\n",
    "train_df, validate_df = train_test_split(df, test_size=0.2, random_state=42)\n",
    "validate_df, test_df = train_test_split(validate_df, test_size=0.5, random_state=42)\n",
    "\n",
    "train_df = train_df.reset_index(drop=True)\n",
    "validate_df = validate_df.reset_index(drop=True)\n",
    "test_df = test_df.reset_index(drop=True)\n",
    "\n",
    "total_train = train_df.shape[0]\n",
    "total_validate = validate_df.shape[0]\n",
    "\n",
    "print('train size = ', total_validate , 'validate size = ', total_validate, 'test size = ', test_df.shape[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.035464,
     "end_time": "2021-02-11T20:31:24.472100",
     "exception": false,
     "start_time": "2021-02-11T20:31:24.436636",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Train the model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.033432,
     "end_time": "2021-02-11T20:31:24.539653",
     "exception": false,
     "start_time": "2021-02-11T20:31:24.506221",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "We will first create the training data generator, that will get the images from the input data directory to train on them. We will also create a generator for the validation set.\n",
    "\n",
    "Applying Data Augmentation on the training set was taking too long to be executed and the initial results didn't show much improvement than the results without augmentation, so I commented the augmentation to make the training faster. However fell free to uncomment the Data Augmentation lines in the following cell and play a bit with it."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-11T20:31:24.643893Z",
     "iopub.status.busy": "2021-02-11T20:31:24.638718Z",
     "iopub.status.idle": "2021-02-11T20:31:39.247409Z",
     "shell.execute_reply": "2021-02-11T20:31:39.248258Z"
    },
    "papermill": {
     "duration": 14.675024,
     "end_time": "2021-02-11T20:31:39.248403",
     "exception": false,
     "start_time": "2021-02-11T20:31:24.573379",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 12412 validated image filenames belonging to 12 classes.\n"
     ]
    }
   ],
   "source": [
    "batch_size=64\n",
    "\n",
    "train_datagen = image.ImageDataGenerator(\n",
    "    \n",
    "    ###  Augmentation Start  ###\n",
    "    \n",
    "    #rotation_range=30,\n",
    "    #shear_range=0.1,\n",
    "    #zoom_range=0.3,\n",
    "    #horizontal_flip=True,\n",
    "    #vertical_flip = True,\n",
    "    #width_shift_range=0.2,\n",
    "    #height_shift_range=0.2\n",
    "    \n",
    "    ##  Augmentation End  ###\n",
    ")\n",
    "\n",
    "train_generator = train_datagen.flow_from_dataframe(\n",
    "    train_df, \n",
    "    base_path, \n",
    "    x_col='filename',\n",
    "    y_col='category',\n",
    "    target_size=IMAGE_SIZE,\n",
    "    class_mode='categorical',\n",
    "    batch_size=batch_size\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-11T20:31:39.326111Z",
     "iopub.status.busy": "2021-02-11T20:31:39.325320Z",
     "iopub.status.idle": "2021-02-11T20:31:40.978711Z",
     "shell.execute_reply": "2021-02-11T20:31:40.979236Z"
    },
    "papermill": {
     "duration": 1.69855,
     "end_time": "2021-02-11T20:31:40.979394",
     "exception": false,
     "start_time": "2021-02-11T20:31:39.280844",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 1551 validated image filenames belonging to 12 classes.\n"
     ]
    }
   ],
   "source": [
    "validation_datagen = image.ImageDataGenerator()\n",
    "\n",
    "validation_generator = validation_datagen.flow_from_dataframe(\n",
    "    validate_df, \n",
    "    base_path, \n",
    "    x_col='filename',\n",
    "    y_col='category',\n",
    "    target_size=IMAGE_SIZE,\n",
    "    class_mode='categorical',\n",
    "    batch_size=batch_size\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-11T20:31:41.056281Z",
     "iopub.status.busy": "2021-02-11T20:31:41.055573Z",
     "iopub.status.idle": "2021-02-11T20:43:31.253030Z",
     "shell.execute_reply": "2021-02-11T20:43:31.252167Z"
    },
    "papermill": {
     "duration": 710.238601,
     "end_time": "2021-02-11T20:43:31.253188",
     "exception": false,
     "start_time": "2021-02-11T20:31:41.014587",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20\n",
      "193/193 [==============================] - 175s 906ms/step - loss: 0.5862 - categorical_accuracy: 0.8528 - val_loss: 0.3026 - val_categorical_accuracy: 0.9277\n",
      "Epoch 2/20\n",
      "193/193 [==============================] - 88s 458ms/step - loss: 0.2357 - categorical_accuracy: 0.9393 - val_loss: 0.2321 - val_categorical_accuracy: 0.9368\n",
      "Epoch 3/20\n",
      "193/193 [==============================] - 88s 457ms/step - loss: 0.1802 - categorical_accuracy: 0.9519 - val_loss: 0.1990 - val_categorical_accuracy: 0.9460\n",
      "Epoch 4/20\n",
      "193/193 [==============================] - 88s 455ms/step - loss: 0.1511 - categorical_accuracy: 0.9589 - val_loss: 0.1933 - val_categorical_accuracy: 0.9427\n",
      "Epoch 5/20\n",
      "193/193 [==============================] - 87s 449ms/step - loss: 0.1318 - categorical_accuracy: 0.9649 - val_loss: 0.1805 - val_categorical_accuracy: 0.9499\n",
      "Epoch 6/20\n",
      "193/193 [==============================] - 86s 446ms/step - loss: 0.1165 - categorical_accuracy: 0.9696 - val_loss: 0.1760 - val_categorical_accuracy: 0.9473\n",
      "Epoch 7/20\n",
      "193/193 [==============================] - ETA: 0s - loss: 0.1046 - categorical_accuracy: 0.9728Restoring model weights from the end of the best epoch.\n",
      "193/193 [==============================] - 86s 446ms/step - loss: 0.1046 - categorical_accuracy: 0.9728 - val_loss: 0.1704 - val_categorical_accuracy: 0.9492\n",
      "Epoch 00007: early stopping\n"
     ]
    }
   ],
   "source": [
    "EPOCHS = 20\n",
    "history = model.fit_generator(\n",
    "    train_generator, \n",
    "    epochs=EPOCHS,\n",
    "    validation_data=validation_generator,\n",
    "    validation_steps=total_validate//batch_size,\n",
    "    steps_per_epoch=total_train//batch_size,\n",
    "    callbacks=callbacks\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-11T20:43:32.250241Z",
     "iopub.status.busy": "2021-02-11T20:43:32.249292Z",
     "iopub.status.idle": "2021-02-11T20:43:32.510318Z",
     "shell.execute_reply": "2021-02-11T20:43:32.509777Z"
    },
    "papermill": {
     "duration": 0.761276,
     "end_time": "2021-02-11T20:43:32.510438",
     "exception": false,
     "start_time": "2021-02-11T20:43:31.749162",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "model.save_weights(\"model.h5\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.499679,
     "end_time": "2021-02-11T20:43:33.507688",
     "exception": false,
     "start_time": "2021-02-11T20:43:33.008009",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Visualize the training process\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-11T20:43:34.544981Z",
     "iopub.status.busy": "2021-02-11T20:43:34.544030Z",
     "iopub.status.idle": "2021-02-11T20:43:34.874208Z",
     "shell.execute_reply": "2021-02-11T20:43:34.874747Z"
    },
    "papermill": {
     "duration": 0.835036,
     "end_time": "2021-02-11T20:43:34.874921",
     "exception": false,
     "start_time": "2021-02-11T20:43:34.039885",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(2, 1)\n",
    "ax1.plot(history.history['loss'], color='b', label=\"Training loss\")\n",
    "ax1.plot(history.history['val_loss'], color='r', label=\"validation loss\")\n",
    "ax1.set_yticks(np.arange(0, 0.7, 0.1))\n",
    "ax1.legend()\n",
    "\n",
    "ax2.plot(history.history['categorical_accuracy'], color='b', label=\"Training accuracy\")\n",
    "ax2.plot(history.history['val_categorical_accuracy'], color='r',label=\"Validation accuracy\")\n",
    "ax2.legend()\n",
    "\n",
    "legend = plt.legend(loc='best')\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.496799,
     "end_time": "2021-02-11T20:43:35.872581",
     "exception": false,
     "start_time": "2021-02-11T20:43:35.375782",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Evaluate the test"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.493416,
     "end_time": "2021-02-11T20:43:36.863452",
     "exception": false,
     "start_time": "2021-02-11T20:43:36.370036",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "To evaluate the performance of our model we will create a test generator to load the images from the input data directory and evaluate the results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-11T20:43:37.947298Z",
     "iopub.status.busy": "2021-02-11T20:43:37.942897Z",
     "iopub.status.idle": "2021-02-11T20:43:40.190982Z",
     "shell.execute_reply": "2021-02-11T20:43:40.191643Z"
    },
    "papermill": {
     "duration": 2.830765,
     "end_time": "2021-02-11T20:43:40.191839",
     "exception": false,
     "start_time": "2021-02-11T20:43:37.361074",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 1552 validated image filenames belonging to 12 classes.\n"
     ]
    }
   ],
   "source": [
    "test_datagen = image.ImageDataGenerator()\n",
    "\n",
    "test_generator = test_datagen.flow_from_dataframe(\n",
    "    dataframe= test_df,\n",
    "    directory=base_path,\n",
    "    x_col='filename',\n",
    "    y_col='category',\n",
    "    target_size=IMAGE_SIZE,\n",
    "    color_mode=\"rgb\",\n",
    "    class_mode=\"categorical\",\n",
    "    batch_size=1,\n",
    "    shuffle=False \n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-11T20:43:41.446337Z",
     "iopub.status.busy": "2021-02-11T20:43:41.445068Z",
     "iopub.status.idle": "2021-02-11T20:44:02.018285Z",
     "shell.execute_reply": "2021-02-11T20:44:02.018914Z"
    },
    "papermill": {
     "duration": 21.107171,
     "end_time": "2021-02-11T20:44:02.019100",
     "exception": false,
     "start_time": "2021-02-11T20:43:40.911929",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy on test set =  94.91 % \n"
     ]
    }
   ],
   "source": [
    "filenames = test_generator.filenames\n",
    "nb_samples = len(filenames)\n",
    "\n",
    "_, accuracy = model.evaluate_generator(test_generator, nb_samples)\n",
    "\n",
    "print('accuracy on test set = ',  round((accuracy * 100),2 ), '% ') "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.551399,
     "end_time": "2021-02-11T20:44:03.067758",
     "exception": false,
     "start_time": "2021-02-11T20:44:02.516359",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "Great, the accuracy is well over 90% !! (it varies from one run to another due to the random shuffling but is between 94% and 96%) \n",
    "\n",
    "But let's have a look on the F1 score for each of the categories. For that we will be using the classification_report from the sklearn package."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-11T20:44:04.230003Z",
     "iopub.status.busy": "2021-02-11T20:44:04.229259Z",
     "iopub.status.idle": "2021-02-11T20:44:04.234331Z",
     "shell.execute_reply": "2021-02-11T20:44:04.233841Z"
    },
    "papermill": {
     "duration": 0.58757,
     "end_time": "2021-02-11T20:44:04.234461",
     "exception": false,
     "start_time": "2021-02-11T20:44:03.646891",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{0: 'battery', 1: 'biological', 2: 'brown-glass', 3: 'cardboard', 4: 'clothes', 5: 'green-glass', 6: 'metal', 7: 'paper', 8: 'plastic', 9: 'shoes', 10: 'trash', 11: 'white-glass'}\n"
     ]
    }
   ],
   "source": [
    "# We defined at the beginning of this notebook a dictionary that maps the categories number to names, but the train generator\n",
    "# generated it's own dictionary and it has assigned different numbers to our categories and the predictions made by the model \n",
    "# will be made using the genrator's dictionary.\n",
    "\n",
    "gen_label_map = test_generator.class_indices\n",
    "gen_label_map = dict((v,k) for k,v in gen_label_map.items())\n",
    "print(gen_label_map)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-11T20:44:05.255019Z",
     "iopub.status.busy": "2021-02-11T20:44:05.254234Z",
     "iopub.status.idle": "2021-02-11T20:44:20.158564Z",
     "shell.execute_reply": "2021-02-11T20:44:20.159227Z"
    },
    "papermill": {
     "duration": 15.403567,
     "end_time": "2021-02-11T20:44:20.159410",
     "exception": false,
     "start_time": "2021-02-11T20:44:04.755843",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "     battery       0.93      0.97      0.95       102\n",
      "  biological       1.00      1.00      1.00       102\n",
      " brown-glass       0.88      0.86      0.87        49\n",
      "   cardboard       0.96      0.94      0.95        93\n",
      "     clothes       0.98      0.99      0.99       533\n",
      " green-glass       0.84      0.82      0.83        62\n",
      "       metal       0.88      0.87      0.87        83\n",
      "       paper       0.95      0.94      0.95       107\n",
      "     plastic       0.84      0.84      0.84        87\n",
      "       shoes       0.99      0.98      0.99       212\n",
      "       trash       0.96      0.96      0.96        57\n",
      " white-glass       0.87      0.82      0.84        65\n",
      "\n",
      "    accuracy                           0.95      1552\n",
      "   macro avg       0.92      0.92      0.92      1552\n",
      "weighted avg       0.95      0.95      0.95      1552\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# get the model's predictions for the test set\n",
    "preds = model.predict_generator(test_generator, nb_samples)\n",
    "\n",
    "# Get the category with the highest predicted probability, the prediction is only the category's number and not name\n",
    "preds = preds.argmax(1)\n",
    "\n",
    "# Convert the predicted category's number to name \n",
    "preds = [gen_label_map[item] for item in preds]\n",
    "\n",
    "# Convert the pandas dataframe to a numpy matrix\n",
    "labels = test_df['category'].to_numpy()\n",
    "\n",
    "print(classification_report(labels, preds))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.493724,
     "end_time": "2021-02-11T20:44:21.143187",
     "exception": false,
     "start_time": "2021-02-11T20:44:20.649463",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "The table above shows among other info the F1 score of each category. In the bottom of the F1 score column notice two numbers accuracy and macro avg. Accuracy is the same accuracy that we evaluated above for the test set, it is a weighted average. \n",
    "\n",
    "However the macro avg (unweighted average) is a bit less than accuracy. This is because the clothes category, the category which has by far the largest number of images, has a very high F1 score, so accuracy (the weighted average) is higher than the unweighted average (macro avg).\n",
    "\n",
    "For this problem I would consider the macro avg a better measure of accuracy as it takes an average of all the F1 scores regardless of how much data we have in the training data for each category."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.732447,
     "end_time": "2021-02-11T20:44:22.365045",
     "exception": false,
     "start_time": "2021-02-11T20:44:21.632598",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## End Note"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.50476,
     "end_time": "2021-02-11T20:44:23.387252",
     "exception": false,
     "start_time": "2021-02-11T20:44:22.882492",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "If you liked this kernel please upvote! I also created a small web app to classify the garbage images, https://mostafa-portfolio.azurewebsites.net/ feel free to play around with it!"
   ]
  }
 ],
 "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": "3.8.8"
  },
  "papermill": {
   "duration": 805.18131,
   "end_time": "2021-02-11T20:44:25.762612",
   "environment_variables": {},
   "exception": null,
   "input_path": "__notebook__.ipynb",
   "output_path": "__notebook__.ipynb",
   "parameters": {},
   "start_time": "2021-02-11T20:31:00.581302",
   "version": "2.1.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}