tutorial_screw_training_airgapped.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "18630765",
   "metadata": {},
   "source": [
    "### [Optional] Reset the data folder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "bbc7bac9",
   "metadata": {},
   "outputs": [],
   "source": [
    "!rm -rf ./data/arrange/ ./data/model/"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d295ffdd",
   "metadata": {},
   "source": [
    "## Part 1: Fill in the Variable \n",
    "Please modify the variable to get the correct information."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "50a8abc3",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Label Data file\n",
    "label_data_file = \"project-6-at-2022-09-19-04-17-b9f72b54.json\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "dbc26cbf",
   "metadata": {
    "tags": [
     "parameters"
    ]
   },
   "outputs": [],
   "source": [
    "# Model learning rate\n",
    "base_learning_rate = 0.01"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "0886a32a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# MLFlow experiment name\n",
    "mlflow_experiment_name = \"tutorial_screw_train\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "14e53a00",
   "metadata": {},
   "source": [
    "## Part 2: Install python package"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "2a4d6b92",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Skip this method. The environment is airgapped.\n",
      "You need to use the correct jupyter image.\n"
     ]
    }
   ],
   "source": [
    "print(\"Skip this method. The environment is airgapped.\")\n",
    "print(\"You need to use the correct jupyter image.\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c1aeb4e5",
   "metadata": {},
   "source": [
    "## Part 3: Data Processing"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "560037e6",
   "metadata": {},
   "source": [
    "### Step 1: Prepare the Data\n",
    "Download and extract the zip file, then follow the tutorial to use Label Studio to label the data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "bd152829",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Kaggle dataset: https://www.kaggle.com/datasets/ruruamour/screw-dataset\n",
    "\n",
    "original_data_folder_path = \"./data/original/\"\n",
    "label_data_folder_path = \"./data/label/\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e4c4061b",
   "metadata": {},
   "source": [
    "### Step 2: Data ELT pipeline\n",
    "\n",
    "We will do the data ELT pipeline in the following code."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f82e660a",
   "metadata": {},
   "source": [
    "#### Import the python package"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "156f296e",
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "import shutil\n",
    "import os\n",
    "from pathlib import Path"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "daf05764",
   "metadata": {},
   "source": [
    "#### Data Extraction: Read the label studio export json-format file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "b2f298bc",
   "metadata": {},
   "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>image</th>\n",
       "      <th>id</th>\n",
       "      <th>choice</th>\n",
       "      <th>annotator</th>\n",
       "      <th>annotation_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ok001.png</td>\n",
       "      <td>40</td>\n",
       "      <td>good</td>\n",
       "      <td>demo@infuseai.io</td>\n",
       "      <td>44</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ok002.png</td>\n",
       "      <td>39</td>\n",
       "      <td>good</td>\n",
       "      <td>demo@infuseai.io</td>\n",
       "      <td>43</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ok003.png</td>\n",
       "      <td>38</td>\n",
       "      <td>good</td>\n",
       "      <td>demo@infuseai.io</td>\n",
       "      <td>42</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ok004.png</td>\n",
       "      <td>37</td>\n",
       "      <td>good</td>\n",
       "      <td>demo@infuseai.io</td>\n",
       "      <td>41</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ok005.png</td>\n",
       "      <td>36</td>\n",
       "      <td>good</td>\n",
       "      <td>demo@infuseai.io</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       image  id choice         annotator  annotation_id\n",
       "0  ok001.png  40   good  demo@infuseai.io             44\n",
       "1  ok002.png  39   good  demo@infuseai.io             43\n",
       "2  ok003.png  38   good  demo@infuseai.io             42\n",
       "3  ok004.png  37   good  demo@infuseai.io             41\n",
       "4  ok005.png  36   good  demo@infuseai.io             40"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "df = pd.read_json(os.path.join(label_data_folder_path, label_data_file))\n",
    "df['image'] = df['image'].str.split(pat=\"/\", expand=True)[8]\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "132740bd",
   "metadata": {},
   "source": [
    "#### Data Load-To: Create a training and validation data folder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "0e2e804e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'train': {'folder_path': './data/arrange/train',\n",
       "  'good': {'folder_path': './data/arrange/train/good'},\n",
       "  'bad': {'folder_path': './data/arrange/train/bad'}},\n",
       " 'valid': {'folder_path': './data/arrange/valid',\n",
       "  'good': {'folder_path': './data/arrange/valid/good'},\n",
       "  'bad': {'folder_path': './data/arrange/valid/bad'}}}"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "arrange_data_folder_path = './data/arrange/'\n",
    "dataset_type_list = ['train', 'valid']\n",
    "label_list = ['good', 'bad']\n",
    "dataset_dict = {}\n",
    "\n",
    "for dataset_type in dataset_type_list:\n",
    "    dataset_dict[dataset_type] = {}\n",
    "    dataset_dict[dataset_type]['folder_path'] = os.path.join(arrange_data_folder_path, dataset_type)\n",
    "    for label in label_list:\n",
    "        arrange_subfolder_path = os.path.join(arrange_data_folder_path, dataset_type, label)\n",
    "        dataset_dict[dataset_type][label] = {}\n",
    "        dataset_dict[dataset_type][label]['folder_path'] = arrange_subfolder_path\n",
    "        os.makedirs(arrange_subfolder_path, exist_ok=True)\n",
    "    \n",
    "dataset_dict"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f93c7ef9",
   "metadata": {},
   "source": [
    "#### Data Transformation: Put the original image data into arrangement folder."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "102dc586",
   "metadata": {},
   "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>image</th>\n",
       "      <th>id</th>\n",
       "      <th>choice</th>\n",
       "      <th>annotator</th>\n",
       "      <th>annotation_id</th>\n",
       "      <th>arrange_path</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ok001.png</td>\n",
       "      <td>40</td>\n",
       "      <td>good</td>\n",
       "      <td>demo@infuseai.io</td>\n",
       "      <td>44</td>\n",
       "      <td>./data/arrange/valid/good/ok001.png</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ok002.png</td>\n",
       "      <td>39</td>\n",
       "      <td>good</td>\n",
       "      <td>demo@infuseai.io</td>\n",
       "      <td>43</td>\n",
       "      <td>./data/arrange/valid/good/ok002.png</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ok003.png</td>\n",
       "      <td>38</td>\n",
       "      <td>good</td>\n",
       "      <td>demo@infuseai.io</td>\n",
       "      <td>42</td>\n",
       "      <td>./data/arrange/valid/good/ok003.png</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ok004.png</td>\n",
       "      <td>37</td>\n",
       "      <td>good</td>\n",
       "      <td>demo@infuseai.io</td>\n",
       "      <td>41</td>\n",
       "      <td>./data/arrange/valid/good/ok004.png</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ok005.png</td>\n",
       "      <td>36</td>\n",
       "      <td>good</td>\n",
       "      <td>demo@infuseai.io</td>\n",
       "      <td>40</td>\n",
       "      <td>./data/arrange/train/good/ok005.png</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       image  id choice         annotator  annotation_id  \\\n",
       "0  ok001.png  40   good  demo@infuseai.io             44   \n",
       "1  ok002.png  39   good  demo@infuseai.io             43   \n",
       "2  ok003.png  38   good  demo@infuseai.io             42   \n",
       "3  ok004.png  37   good  demo@infuseai.io             41   \n",
       "4  ok005.png  36   good  demo@infuseai.io             40   \n",
       "\n",
       "                          arrange_path  \n",
       "0  ./data/arrange/valid/good/ok001.png  \n",
       "1  ./data/arrange/valid/good/ok002.png  \n",
       "2  ./data/arrange/valid/good/ok003.png  \n",
       "3  ./data/arrange/valid/good/ok004.png  \n",
       "4  ./data/arrange/train/good/ok005.png  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "val_count = 4\n",
    "val_good_count = 0\n",
    "val_bad_count = 0\n",
    "\n",
    "arrange_folder_path_list = []\n",
    "for original_file_name, label_result in zip(df['image'], df['choice']):\n",
    "    original_file_path = os.path.join(original_data_folder_path, 'train', original_file_name)\n",
    "    if label_result == 'good':\n",
    "        if val_good_count < val_count:\n",
    "            arrange_file_path = os.path.join(dataset_dict['valid']['good']['folder_path'], original_file_name)\n",
    "            val_good_count = val_good_count + 1\n",
    "        else:\n",
    "            arrange_file_path = os.path.join(dataset_dict['train']['good']['folder_path'], original_file_name)\n",
    "    else:\n",
    "        if val_bad_count < val_count:\n",
    "            arrange_file_path = os.path.join(dataset_dict['valid']['bad']['folder_path'], original_file_name)\n",
    "            val_bad_count = val_bad_count + 1\n",
    "        else:\n",
    "            arrange_file_path = os.path.join(dataset_dict['train']['bad']['folder_path'], original_file_name)\n",
    "    arrange_folder_path_list.append(arrange_file_path)\n",
    "    shutil.copyfile(original_file_path, arrange_file_path)\n",
    "\n",
    "df['arrange_path'] = arrange_folder_path_list\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "84705ce1",
   "metadata": {},
   "source": [
    "## Part 3: Model Training Pipeline\n",
    "\n",
    "After labeled all the images, you can run the following cells to start training and log results to MLflow."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cd52d925",
   "metadata": {},
   "source": [
    "### Step 1: Import python package"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "798fd911",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-01-03 00:41:59.031147: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcudart.so.11.0'; dlerror: libcudart.so.11.0: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: \n",
      "2023-01-03 00:41:59.031203: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine.\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "import tensorflow_hub as hub\n",
    "import matplotlib.pyplot as plt\n",
    "from tensorflow.keras.preprocessing import image_dataset_from_directory"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "afaa498b",
   "metadata": {},
   "source": [
    "### Step 2: mlflow initial"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "c47650df",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023/01/03 00:42:02 WARNING mlflow.utils.autologging_utils: You are using an unsupported version of tensorflow. If you encounter errors during autologging, try upgrading / downgrading tensorflow to a supported version, or try upgrading MLflow.\n"
     ]
    }
   ],
   "source": [
    "import mlflow\n",
    "mlflow.set_experiment(mlflow_experiment_name)\n",
    "mlflow.tensorflow.autolog()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "49e7999c",
   "metadata": {},
   "source": [
    "### Step 3: Setting dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "e8c6b7d9",
   "metadata": {},
   "outputs": [],
   "source": [
    "BATCH_SIZE = 8\n",
    "IMG_SIZE = (1024, 1024)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "8e81a437",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-01-03 00:42:02.471633: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcuda.so.1'; dlerror: libcuda.so.1: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: \n",
      "2023-01-03 00:42:02.471678: W tensorflow/stream_executor/cuda/cuda_driver.cc:269] failed call to cuInit: UNKNOWN ERROR (303)\n",
      "2023-01-03 00:42:02.471703: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:156] kernel driver does not appear to be running on this host (jupyter-phadmin): /proc/driver/nvidia/version does not exist\n",
      "2023-01-03 00:42:02.474453: I tensorflow/core/platform/cpu_feature_guard.cc:151] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 32 files belonging to 2 classes.\n"
     ]
    }
   ],
   "source": [
    "train_dataset = image_dataset_from_directory(dataset_dict['train']['folder_path'],\n",
    "                                             shuffle=True,\n",
    "                                             batch_size=BATCH_SIZE,\n",
    "                                             image_size=IMG_SIZE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "995be1c5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 8 files belonging to 2 classes.\n"
     ]
    }
   ],
   "source": [
    "validation_dataset = image_dataset_from_directory(dataset_dict['valid']['folder_path'],\n",
    "                                                  shuffle=True,\n",
    "                                                  batch_size=BATCH_SIZE,\n",
    "                                                  image_size=IMG_SIZE)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eb9184c5",
   "metadata": {},
   "source": [
    "#### Show example\n",
    "Show some cases of training examples."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "dda94521",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x720 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "class_names = train_dataset.class_names\n",
    "\n",
    "show_img = 1\n",
    "\n",
    "if show_img:\n",
    "    plt.figure(figsize=(10, 10))\n",
    "    for images, labels in train_dataset.take(1):\n",
    "        for i in range(6):\n",
    "            ax = plt.subplot(3, 3, i + 1)\n",
    "            plt.imshow(images[i].numpy().astype(\"uint8\"))\n",
    "            plt.title(class_names[labels[i]])\n",
    "            plt.axis(\"off\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "864fd073",
   "metadata": {},
   "source": [
    "### Step 4: Build the Model\n",
    "We use mobilenet as our model."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b0879181",
   "metadata": {},
   "source": [
    "#### Build the data augmentation layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "4d854b28",
   "metadata": {},
   "outputs": [],
   "source": [
    "data_augmentation = tf.keras.Sequential([\n",
    "    tf.keras.layers.experimental.preprocessing.RandomFlip(),\n",
    "])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "94e1e93e",
   "metadata": {},
   "source": [
    "#### Build the Input preprocessing application layer\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "8424cffb",
   "metadata": {},
   "outputs": [],
   "source": [
    "preprocess_input = tf.keras.applications.mobilenet_v2.preprocess_input"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c21c52c0",
   "metadata": {},
   "source": [
    "#### Build the mobilenet v3 base model\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "498599b9",
   "metadata": {},
   "outputs": [],
   "source": [
    "!mkdir ./data/model/\n",
    "!tar -zxf ./data/imagenet_mobilenet_v3_large_100_224_feature_vector_5.tar.gz \\\n",
    " --directory ./data/model/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "4aef214b",
   "metadata": {},
   "outputs": [],
   "source": [
    "model_url = \"./data/model/\"\n",
    "base_model = tf.keras.Sequential([\n",
    "    tf.keras.layers.experimental.preprocessing.Resizing(224, 224),\n",
    "    hub.KerasLayer(model_url, trainable=False)\n",
    "])\n",
    "base_model.build([None, 224, 224, 3])\n",
    "\n",
    "# Change base model trainable is False(We only fine-tune the model in this example.)\n",
    "base_model.trainable = False"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "adc6f39a",
   "metadata": {},
   "source": [
    "#### Build the prediction layer\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "e2382697",
   "metadata": {},
   "outputs": [],
   "source": [
    "prediction_layer = tf.keras.layers.Dense(1, kernel_regularizer=tf.keras.regularizers.l2(l2=0.1))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6ca1d2d6",
   "metadata": {},
   "source": [
    "#### Combine all the model layer into one model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "415e70cd",
   "metadata": {},
   "outputs": [],
   "source": [
    "inputs = tf.keras.Input(shape=(1024, 1024, 3))\n",
    "x = data_augmentation(inputs)\n",
    "x = preprocess_input(x)\n",
    "x = base_model(x, training=False)\n",
    "outputs = prediction_layer(x)\n",
    "model = tf.keras.Model(inputs, outputs)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9ff0a3eb",
   "metadata": {
    "tags": [
     "parameters"
    ]
   },
   "source": [
    "#### Compile the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "a0571f81",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.7/site-packages/keras/optimizer_v2/gradient_descent.py:102: UserWarning: The `lr` argument is deprecated, use `learning_rate` instead.\n",
      "  super(SGD, self).__init__(name, **kwargs)\n"
     ]
    }
   ],
   "source": [
    "model.compile(optimizer=tf.keras.optimizers.SGD(lr=base_learning_rate, momentum=0.9),\n",
    "              loss=tf.keras.losses.BinaryCrossentropy(from_logits=True),\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d30de8c1",
   "metadata": {},
   "source": [
    "#### Get the model summary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "8fa2a99a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " input_1 (InputLayer)        [(None, 1024, 1024, 3)]   0         \n",
      "                                                                 \n",
      " sequential (Sequential)     (None, 1024, 1024, 3)     0         \n",
      "                                                                 \n",
      " tf.math.truediv (TFOpLambda  (None, 1024, 1024, 3)    0         \n",
      " )                                                               \n",
      "                                                                 \n",
      " tf.math.subtract (TFOpLambd  (None, 1024, 1024, 3)    0         \n",
      " a)                                                              \n",
      "                                                                 \n",
      " sequential_1 (Sequential)   (None, 1280)              4226432   \n",
      "                                                                 \n",
      " dense (Dense)               (None, 1)                 1281      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 4,227,713\n",
      "Trainable params: 1,281\n",
      "Non-trainable params: 4,226,432\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "88bcc8fd",
   "metadata": {},
   "source": [
    "### Step 5: Model Training"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "798bb3ac",
   "metadata": {},
   "source": [
    "#### Check the initial dataset metrics value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "8aa34c0e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 2s 2s/step - loss: 0.8903 - accuracy: 0.5000\n"
     ]
    }
   ],
   "source": [
    "loss0, accuracy0 = model.evaluate(validation_dataset)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "e845da8c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "initial loss: 0.89\n",
      "initial accuracy: 0.50\n"
     ]
    }
   ],
   "source": [
    "print(\"initial loss: {:.2f}\".format(loss0))\n",
    "print(\"initial accuracy: {:.2f}\".format(accuracy0))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f5a317cc",
   "metadata": {},
   "source": [
    "#### Setting the callback"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "ac216b04",
   "metadata": {},
   "outputs": [],
   "source": [
    "def scheduler(epoch, lr):\n",
    "    if epoch >= 10 and epoch % 10 == 0:\n",
    "        return lr / 5\n",
    "    return lr\n",
    "\n",
    "learning_rate_scheduler = tf.keras.callbacks.LearningRateScheduler(scheduler)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "4f8e0243",
   "metadata": {},
   "outputs": [],
   "source": [
    "callback_list = [learning_rate_scheduler]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "99ae7c52",
   "metadata": {},
   "source": [
    "#### Train the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "f73363be",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023/01/03 00:42:12 INFO mlflow.utils.autologging_utils: Created MLflow autologging run with ID 'be60cbf132044a1e87359dc6eaef687e', which will track hyperparameters, performance metrics, model artifacts, and lineage information for the current tensorflow workflow\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/25\n",
      "4/4 [==============================] - 12s 1s/step - loss: 0.6919 - accuracy: 0.6875 - val_loss: 2.0975 - val_accuracy: 0.5000 - lr: 0.0100\n",
      "Epoch 2/25\n",
      "4/4 [==============================] - 5s 981ms/step - loss: 0.7779 - accuracy: 0.8125 - val_loss: 1.0295 - val_accuracy: 0.3750 - lr: 0.0100\n",
      "Epoch 3/25\n",
      "4/4 [==============================] - 5s 1s/step - loss: 0.5321 - accuracy: 0.8750 - val_loss: 1.0592 - val_accuracy: 0.7500 - lr: 0.0100\n",
      "Epoch 4/25\n",
      "4/4 [==============================] - 4s 1s/step - loss: 0.4299 - accuracy: 0.9062 - val_loss: 1.1955 - val_accuracy: 0.5000 - lr: 0.0100\n",
      "Epoch 5/25\n",
      "4/4 [==============================] - 5s 1s/step - loss: 0.4310 - accuracy: 0.8750 - val_loss: 1.0194 - val_accuracy: 0.5000 - lr: 0.0100\n",
      "Epoch 6/25\n",
      "4/4 [==============================] - 5s 1s/step - loss: 0.3644 - accuracy: 0.9375 - val_loss: 0.9425 - val_accuracy: 0.7500 - lr: 0.0100\n",
      "Epoch 7/25\n",
      "4/4 [==============================] - 4s 980ms/step - loss: 0.3058 - accuracy: 0.9688 - val_loss: 0.8732 - val_accuracy: 0.8750 - lr: 0.0100\n",
      "Epoch 8/25\n",
      "4/4 [==============================] - 5s 1s/step - loss: 0.2961 - accuracy: 1.0000 - val_loss: 0.8269 - val_accuracy: 0.8750 - lr: 0.0100\n",
      "Epoch 9/25\n",
      "4/4 [==============================] - 5s 1s/step - loss: 0.2920 - accuracy: 0.9688 - val_loss: 0.7894 - val_accuracy: 0.8750 - lr: 0.0100\n",
      "Epoch 10/25\n",
      "4/4 [==============================] - 5s 948ms/step - loss: 0.2791 - accuracy: 1.0000 - val_loss: 0.8139 - val_accuracy: 0.7500 - lr: 0.0100\n",
      "Epoch 11/25\n",
      "4/4 [==============================] - 5s 1s/step - loss: 0.2819 - accuracy: 0.9688 - val_loss: 0.8115 - val_accuracy: 0.6250 - lr: 0.0020\n",
      "Epoch 12/25\n",
      "4/4 [==============================] - 4s 952ms/step - loss: 0.2989 - accuracy: 0.9375 - val_loss: 0.7448 - val_accuracy: 0.8750 - lr: 0.0020\n",
      "Epoch 13/25\n",
      "4/4 [==============================] - 5s 1s/step - loss: 0.2405 - accuracy: 0.9688 - val_loss: 0.7166 - val_accuracy: 0.8750 - lr: 0.0020\n",
      "Epoch 14/25\n",
      "4/4 [==============================] - 5s 1s/step - loss: 0.2635 - accuracy: 0.9688 - val_loss: 0.6967 - val_accuracy: 0.8750 - lr: 0.0020\n",
      "Epoch 15/25\n",
      "4/4 [==============================] - 4s 971ms/step - loss: 0.2318 - accuracy: 1.0000 - val_loss: 0.6825 - val_accuracy: 0.8750 - lr: 0.0020\n",
      "Epoch 16/25\n",
      "4/4 [==============================] - 5s 1s/step - loss: 0.2370 - accuracy: 0.9688 - val_loss: 0.6795 - val_accuracy: 0.8750 - lr: 0.0020\n",
      "Epoch 17/25\n",
      "4/4 [==============================] - 4s 1s/step - loss: 0.2307 - accuracy: 1.0000 - val_loss: 0.7010 - val_accuracy: 0.7500 - lr: 0.0020\n",
      "Epoch 18/25\n",
      "4/4 [==============================] - 5s 1s/step - loss: 0.2682 - accuracy: 0.9688 - val_loss: 0.7004 - val_accuracy: 0.7500 - lr: 0.0020\n",
      "Epoch 19/25\n",
      "4/4 [==============================] - 5s 1s/step - loss: 0.2421 - accuracy: 0.9688 - val_loss: 0.6875 - val_accuracy: 0.7500 - lr: 0.0020\n",
      "Epoch 20/25\n",
      "4/4 [==============================] - 5s 1s/step - loss: 0.2394 - accuracy: 0.9688 - val_loss: 0.6591 - val_accuracy: 0.7500 - lr: 0.0020\n",
      "Epoch 21/25\n",
      "4/4 [==============================] - 5s 1s/step - loss: 0.2246 - accuracy: 1.0000 - val_loss: 0.6533 - val_accuracy: 0.7500 - lr: 4.0000e-04\n",
      "Epoch 22/25\n",
      "4/4 [==============================] - 5s 1s/step - loss: 0.2275 - accuracy: 1.0000 - val_loss: 0.6521 - val_accuracy: 0.7500 - lr: 4.0000e-04\n",
      "Epoch 23/25\n",
      "4/4 [==============================] - 4s 1s/step - loss: 0.2309 - accuracy: 1.0000 - val_loss: 0.6563 - val_accuracy: 0.7500 - lr: 4.0000e-04\n",
      "Epoch 24/25\n",
      "4/4 [==============================] - 5s 1s/step - loss: 0.2345 - accuracy: 0.9688 - val_loss: 0.6620 - val_accuracy: 0.7500 - lr: 4.0000e-04\n",
      "Epoch 25/25\n",
      "4/4 [==============================] - 5s 1s/step - loss: 0.2209 - accuracy: 1.0000 - val_loss: 0.6646 - val_accuracy: 0.7500 - lr: 4.0000e-04\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-01-03 00:44:34.544357: W tensorflow/python/util/util.cc:368] Sets are not currently considered sequences, but this may change in the future, so consider avoiding using them.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: /tmp/tmpjbxl262e/model/data/model/assets\n"
     ]
    }
   ],
   "source": [
    "initial_epochs = 25\n",
    "\n",
    "history = model.fit(train_dataset,\n",
    "                    validation_data=validation_dataset,\n",
    "                    epochs=initial_epochs,\n",
    "                    callbacks=callback_list)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ab455483",
   "metadata": {},
   "source": [
    "### Step 6: Model Evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "6c57541b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 1s 653ms/step - loss: 0.6646 - accuracy: 0.7500\n"
     ]
    }
   ],
   "source": [
    "lossfinal, accuracyfinal = model.evaluate(validation_dataset)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "8b3b35f0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "final loss: 0.66\n",
      "final accuracy: 0.75\n"
     ]
    }
   ],
   "source": [
    "print(\"final loss: {:.2f}\".format(lossfinal))\n",
    "print(\"final accuracy: {:.2f}\".format(accuracyfinal))"
   ]
  }
 ],
 "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": "3.9.13 (main, May 24 2022, 21:13:54) \n[Clang 13.0.0 (clang-1300.0.29.30)]"
  },
  "vscode": {
   "interpreter": {
    "hash": "b0fa6594d8f4cbf19f97940f81e996739fb7646882a419484c72d19e05852a7e"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}