MixedDefectWafer_ViT_v1b.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "fcd4259e",
   "metadata": {},
   "source": [
    "# Visual transformer for mixed-type wafer defect pattern recognition\n",
    "- This notebook is belong to https://github.com/PanithanS/Visual-Transformer-Mixed-type-Wafers-Defect-Recognition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "da6ab406",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "<style>\n",
       "    div#notebook-container    { width: 80%; }\n",
       "    div#menubar-container     { width: 65%; }\n",
       "    div#maintoolbar-container { width: 80%; }\n",
       "</style>\n"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#pip install -U tensorflow-addons\n",
    "\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras import layers\n",
    "import tensorflow_addons as tfa\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import train_test_split\n",
    "import pandas as pd\n",
    "import pickle\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "from IPython.display import display, clear_output, HTML\n",
    "display(HTML(data=\"\"\"\n",
    "<style>\n",
    "    div#notebook-container    { width: 80%; }\n",
    "    div#menubar-container     { width: 65%; }\n",
    "    div#maintoolbar-container { width: 80%; }\n",
    "</style>\n",
    "\"\"\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "c79e2bfc",
   "metadata": {},
   "outputs": [],
   "source": [
    "class data_loader:\n",
    "    \n",
    "    def __init__(self, dataset_path, vis=False):\n",
    "        \n",
    "        \"\"\"this funtion is to initialize the data_loader class\n",
    "        \n",
    "        Args:\n",
    "            dataset_path(string): the dataset object address\n",
    "        \n",
    "        \"\"\"\n",
    "        \n",
    "        self.dataset_path = dataset_path\n",
    "        self.is_vis = vis\n",
    "        \n",
    "        # The list that use to translate the label to defects type\n",
    "        self.label_keys = [\"Center, \", \"Donut, \", \"Edge_Loc, \", \"Edge_Ring, \", \n",
    "                           \"Loc, \", \"Near_Full, \", \"Scratch, \", \"Random, \"]\n",
    "\n",
    "    def load_data(self):\n",
    "        \n",
    "        \"\"\"this funtion is to load data to the vaiable \"self.train\" and \"self.label\"  \"\"\"\n",
    "        \n",
    "        self.data = np.load(self.dataset_path)\n",
    "        self.train = self.data[\"arr_0\"]\n",
    "        self.label = self.data[\"arr_1\"]\n",
    "        \n",
    "        #print(\"data_x\",np.shape(self.train))\n",
    "        print(\"MixedWM38:\", np.shape(self.label)[0], 'wafer loaded')\n",
    "        \n",
    "    def prep_data(self):\n",
    "        \n",
    "        \"\"\"this funtion is to do train_test_split and format the data\"\"\"\n",
    "        \n",
    "        self.x_train, self.x_test, self.y_train, self.y_test = train_test_split(self.train, self.label, test_size = 0.2)\n",
    "        \n",
    "        # Make sure images have shape (Batch, image_size, image_size, channel)\n",
    "        self.x_train = np.expand_dims(self.x_train, -1)\n",
    "        self.x_test = np.expand_dims(self.x_test, -1)\n",
    "        \n",
    "        # Convert class vectors to right format\n",
    "        self.y_train = tf.convert_to_tensor(np.array(self.y_train))\n",
    "        self.y_test = tf.convert_to_tensor(np.array(self.y_test))\n",
    "    \n",
    "    def read_label(self, label, defect_types =''):\n",
    "        \n",
    "        \"\"\"this funtion is to translate the label into defect_types\n",
    "        \n",
    "        Args:\n",
    "            label(list): the label that indicate type of  defect, for instance [0 1 0 1 0 0 0 1]\n",
    "\n",
    "        Return:\n",
    "            defect_types(string): the string that indicate defect type\n",
    "        \"\"\"\n",
    "       \n",
    "        if np.sum(label) == 0:\n",
    "            defect_types = 'Normal wafer'\n",
    "        \n",
    "        else:\n",
    "            for digit in range(np.shape(label)[0]):\n",
    "\n",
    "                if label[digit] == 1:\n",
    "\n",
    "                    defect_types = defect_types + self.label_keys[digit]  \n",
    "        \n",
    "        return defect_types\n",
    "    \n",
    "    def see_wafer(self, wafer_num):\n",
    "        \n",
    "        \"\"\"this funtion is to see the defect pattern in idicated by wafer number\n",
    "        \n",
    "        \n",
    "        Args:\n",
    "            wafer_num(float): the index of loaded data in self.train\n",
    "\n",
    "        Return:\n",
    "            plot(oject): the plot image\n",
    "        \"\"\"\n",
    "        \n",
    "        defect_types = self.read_label(self.label[wafer_num])\n",
    "        \n",
    "        plt.title(\"wafer #\"+str(wafer_num))\n",
    "        plt.imshow(self.train[wafer_num])\n",
    "        plt.colorbar()\n",
    "        plt.show()\n",
    "        \n",
    "        print(\"Defect types=\", defect_types)\n",
    "        print(\"Labeled as:\", self.label[wafer_num])\n",
    "        \n",
    "    def get_data(self):\n",
    "        \n",
    "        \"\"\"this funtion is get the loaded data \n",
    "\n",
    "        Return:\n",
    "            x_train(array): image of defect pattern for model training\n",
    "            x_test(array): image of defect pattern for model validation\n",
    "            y_train(tensor): label of defect pattern for model training\n",
    "            y_test(tensor): label of defect pattern for model validation\n",
    "        \"\"\"\n",
    "            \n",
    "        self.load_data()\n",
    "        self.prep_data()\n",
    "        \n",
    "        return self.x_train, self.x_test, self.y_train, self.y_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "69cbdefc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MixedWM38: 38015 wafer loaded\n"
     ]
    }
   ],
   "source": [
    "dataset_path = r'C:\\Users\\admin\\Desktop\\PanithanS\\Wafer_defect\\Wafer_Map_Datasets.npz'\n",
    "\n",
    "A1 = data_loader(dataset_path, vis=True)\n",
    "x_train, x_test, y_train, y_test = A1.get_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "9bc2a665",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Defect types= Donut, Edge_Ring, Loc, \n",
      "Labeled as: [0 1 0 1 1 0 0 0]\n"
     ]
    }
   ],
   "source": [
    "# 2 type\n",
    "A1.see_wafer(20000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "f3832867",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Defect types= Edge_Ring, Scratch, \n",
      "Labeled as: [0 0 0 1 0 0 1 0]\n"
     ]
    }
   ],
   "source": [
    "# 2 type\n",
    "A1.see_wafer(36015)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "3500b5e2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Defect types= Donut, Edge_Ring, Loc, Scratch, \n",
      "Labeled as: [0 1 0 1 1 0 1 0]\n"
     ]
    }
   ],
   "source": [
    "# 4 type\n",
    "A1.see_wafer(19005)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "73cb3aa3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Defect types= Normal wafer\n",
      "Labeled as: [0 0 0 0 0 0 0 0]\n"
     ]
    }
   ],
   "source": [
    "# Normal\n",
    "A1.see_wafer(33866)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "db0338f3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Defect types= Center, \n",
      "Labeled as: [1 0 0 0 0 0 0 0]\n"
     ]
    }
   ],
   "source": [
    "# 1 type\n",
    "A1.see_wafer(12000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "2028356e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Defect types= Donut, \n",
      "Labeled as: [0 1 0 0 0 0 0 0]\n"
     ]
    }
   ],
   "source": [
    "# 1 type\n",
    "A1.see_wafer(24000)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ca2d8d54",
   "metadata": {},
   "source": [
    "#  Visual transformer model and training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "8b90c7ee",
   "metadata": {},
   "outputs": [],
   "source": [
    "class vis_transformer:\n",
    "     \n",
    "    def __init__(self, x_train, x_test, y_train, y_test):\n",
    "        \n",
    "        \"\"\"this funtion is to initialize the vis_transformer class\n",
    "        \n",
    "        Args:\n",
    "            x_train(array): image of defect pattern for model training\n",
    "            x_test(array): image of defect pattern for model validation\n",
    "            y_train(tensor): label of defect pattern for model training\n",
    "            y_test(tensor): label of defect pattern for model validation\n",
    "\n",
    "        \"\"\"\n",
    "        \n",
    "        self.x_train = x_train\n",
    "        self.x_test = x_test\n",
    "        self.y_train = y_train\n",
    "        self.y_test = y_test\n",
    "        \n",
    "        (_, self.image_size, b, c) = np.shape(self.x_train) # We'll resize input images to this size\n",
    "        \n",
    "        self.input_shape = (self.image_size, self.image_size, c)\n",
    "\n",
    "        self.label_size = np.shape(y_test)[1]\n",
    "        self.patch_size = 13  # Size of the patches to be extract from the input images\n",
    "        self.num_patches = (self.image_size // self.patch_size) ** 2\n",
    "\n",
    "        self.projection_dim = 96\n",
    "        self.num_heads = 4\n",
    "        self.transformer_units = [self.projection_dim * 2, self.projection_dim,]  # Size of the transformer layers\n",
    "        self.transformer_layers = 16\n",
    "        self.mlp_head_units = [2048, 1024]  # Size of the dense layers of the final classifier\n",
    "        \n",
    "    def augmentation(self):\n",
    "        \n",
    "        \"\"\"this funtion is as oject function to do data augmentation\"\"\"\n",
    "        \n",
    "        data_augmentation = keras.Sequential(\n",
    "            [\n",
    "                tf.keras.layers.experimental.preprocessing.Normalization(),\n",
    "                tf.keras.layers.experimental.preprocessing.RandomRotation(factor=0.02),\n",
    "                tf.keras.layers.experimental.preprocessing.RandomZoom(height_factor=0.2, width_factor=0.2),\n",
    "            ], name=\"data_augmentation\",)\n",
    "\n",
    "        # Compute the mean and the variance of the training data for normalization.\n",
    "        data_augmentation.layers[0].adapt(self.x_train)\n",
    "        \n",
    "        \n",
    "    def mlp(self, x, hidden_units, dropout_rate):\n",
    "        \n",
    "        \"\"\"this funtion is multilayer perceptron(mpl) head\n",
    "        \n",
    "        Args:\n",
    "            hidden_units(list): the hidden dimenstion of the mpl\n",
    "            dropout_rate(float): dropout rate of mpl\n",
    "\n",
    "        \"\"\"\n",
    "        \n",
    "        for units in hidden_units:\n",
    "\n",
    "            x = layers.Dense(units, activation=tf.nn.gelu)(x)\n",
    "            x = layers.Dropout(dropout_rate)(x)\n",
    "\n",
    "        return x\n",
    "\n",
    "    def get_patches(self, images):\n",
    "                \n",
    "        \"\"\"this funtion is to split image into patches of \"self.patch_size\" x \"self.patch_size\"\n",
    "        \n",
    "        Args:\n",
    "            images(array): the images to applied patches\n",
    "\n",
    "        Return:\n",
    "            patches(array): the image patches\n",
    "        \n",
    "        \"\"\"\n",
    "            \n",
    "        batch_size = tf.shape(images)[0]\n",
    "        patches = tf.image.extract_patches(\n",
    "            images=images,\n",
    "            sizes=[1, self.patch_size, self.patch_size, 1],\n",
    "            strides=[1, self.patch_size, self.patch_size, 1],\n",
    "            rates=[1, 1, 1, 1],\n",
    "            padding=\"VALID\",\n",
    "        )\n",
    "        patch_dims = patches.shape[-1]\n",
    "        \n",
    "        patches = tf.reshape(patches, [batch_size, -1, patch_dims])\n",
    "        \n",
    "        return patches\n",
    "    \n",
    "    def see_patches(self, image, num_image):\n",
    "        \n",
    "        patches = self.get_patches(image)\n",
    "        \n",
    "        image = self.x_train[num_image]\n",
    "        plt.imshow(image)\n",
    "        plt.clim(0,2)\n",
    "        plt.axis(\"off\")\n",
    "\n",
    "        resized_image = tf.image.resize(tf.convert_to_tensor([image]), size=(self.image_size, self.image_size))\n",
    "        \n",
    "        patches = self.get_patches(resized_image)\n",
    "        \n",
    "        print(f\"Image size: {self.image_size} X {self.image_size}\")\n",
    "        print(f\"Patch size: {self.patch_size} X {self.patch_size}\")\n",
    "        print(f\"Patches per image: {patches.shape[1]}\")\n",
    "        print(f\"Elements per patch: {patches.shape[-1]}\")\n",
    "        print(f\"Num patch:\", self.num_patches)\n",
    "\n",
    "        n = int(np.sqrt(patches.shape[1]))\n",
    "        plt.figure(figsize=(4, 4))\n",
    "        for i, patch in enumerate(patches[0]):\n",
    "            ax = plt.subplot(n, n, i + 1)\n",
    "            patch_img = tf.reshape(patch, (self.patch_size, self.patch_size))\n",
    "            plt.imshow(patch_img.numpy())\n",
    "            plt.clim(0,2)\n",
    "            plt.axis(\"off\")\n",
    "        \n",
    "        print(np.shape(patches))\n",
    "        \n",
    "    def get_patchencoder(self, patch):\n",
    "        \n",
    "        \"\"\"this funtion is encode the patch as projection on dense layer with position embedding \n",
    "        \n",
    "        Args:\n",
    "            patch(array): the images to applied patches\n",
    "\n",
    "        Return:\n",
    "            encoded(oject): the encoded patches\n",
    "        \n",
    "        \"\"\"\n",
    "        \n",
    "        self.projection = layers.Dense(units=self.projection_dim)\n",
    "        self.position_embedding = layers.Embedding(input_dim=self.num_patches, output_dim=self.projection_dim)\n",
    "\n",
    "        positions = tf.range(start=0, limit=self.num_patches, delta=1)\n",
    "        encoded = self.projection(patch) + self.position_embedding(positions)\n",
    "\n",
    "        return encoded\n",
    "    \n",
    "    def create_model(self):\n",
    "        \n",
    "        \"\"\"this funtion is to create visual transformer\n",
    "        \n",
    "        Return:\n",
    "            model(object): visual transformer model\n",
    "        \n",
    "        \"\"\"\n",
    "        inputs = layers.Input(shape=(52,52,1))\n",
    "\n",
    "        # Augment data.\n",
    "        images = inputs #data_augmentation(inputs)\n",
    "\n",
    "        # Create patches.\n",
    "        patches = self.get_patches(images)\n",
    "\n",
    "        # Encode patches.\n",
    "        encoded_patches = self.get_patchencoder(patches)\n",
    "\n",
    "        #print(np.shape(encoded_patches))\n",
    "\n",
    "        # Create multiple layers of the Transformer block.\n",
    "\n",
    "        for _ in range(self.transformer_layers):\n",
    "\n",
    "            # Layer normalization 1.\n",
    "            x1 = layers.LayerNormalization(epsilon=1e-6)(encoded_patches)\n",
    "\n",
    "            # Create a multi-head attention layer.\n",
    "            attention_output = layers.MultiHeadAttention(num_heads=self.num_heads, key_dim=self.projection_dim, dropout=0.1)(x1, x1)\n",
    "\n",
    "            #print(np.shape(attention_output))\n",
    "\n",
    "            # Skip connection 1.\n",
    "            x2 = layers.Add()([attention_output, encoded_patches])\n",
    "\n",
    "            # Layer normalization 2.\n",
    "            x3 = layers.LayerNormalization(epsilon=1e-6)(x2)\n",
    "\n",
    "            # MLP.\n",
    "            x3 = self.mlp(x3, hidden_units=self.transformer_units, dropout_rate=0.1)\n",
    "\n",
    "            # Skip connection 2.\n",
    "            encoded_patches = layers.Add()([x3, x2])\n",
    "\n",
    "        #print('encoded_patches:', np.shape(encoded_patches))\n",
    "\n",
    "        # Create a [batch_size, projection_dim] tensor.\n",
    "        representation = layers.LayerNormalization(epsilon=1e-6)(encoded_patches)\n",
    "        representation = layers.Flatten()(representation)\n",
    "        representation = layers.Dropout(0.2)(representation)\n",
    "\n",
    "        #print('representation:', np.shape(representation))\n",
    "\n",
    "        # Add MLP.\n",
    "        features = self.mlp(representation, hidden_units=self.mlp_head_units, dropout_rate=0.5)\n",
    "\n",
    "        #print('features:', np.shape(features))\n",
    "\n",
    "        # Classify outputs.\n",
    "        #features = layers.Dense(label_size*2, kernel_initializer='he_uniform', activation='relu')(features)\n",
    "        logits = layers.Dense(self.label_size, activation='sigmoid')(features)\n",
    "\n",
    "        # Create the Keras model.\n",
    "        model = keras.Model(inputs=inputs, outputs=logits)\n",
    "\n",
    "        #print(model.summary())\n",
    "\n",
    "        return model\n",
    "    \n",
    "    \n",
    "    def train_model(self, batch_size = 512, num_epochs = 40, save_path='none'):\n",
    "        \n",
    "        \"\"\"this funtion is to train the visual transformer model\n",
    "        \n",
    "        Args:\n",
    "            batch_size(float): training batch_size\n",
    "            num_epochs(float): training num_epochs\n",
    "            num_epochs(string): model oject address, and it will be use for history saving file too\n",
    "        \n",
    "        Return:\n",
    "            history(dict): training history\n",
    "        \n",
    "        \"\"\"\n",
    "        self.model_vit = self.create_model()\n",
    "        \n",
    "        if save_path != 'none':\n",
    "            print(\"set model-saving: \" + save_path )\n",
    "            print(\"set history-saving: \" + save_path+'_history' )\n",
    "        \n",
    "        lr_schedule = keras.optimizers.schedules.ExponentialDecay(initial_learning_rate=1e-3, decay_steps=10000, decay_rate=0.9)\n",
    "        opt = tf.keras.optimizers.Adam(lr_schedule)\n",
    "\n",
    "        self.model_vit.compile(optimizer=opt, loss='binary_crossentropy',metrics=keras.metrics.BinaryAccuracy(name=\"accuracy\"))\n",
    "\n",
    "        checkpoint_filepath = \"./tmp/checkpoint\"\n",
    "        checkpoint_callback = keras.callbacks.ModelCheckpoint(\n",
    "            checkpoint_filepath,\n",
    "            monitor=\"val_accuracy\",\n",
    "            save_best_only=True,\n",
    "            save_weights_only=True,\n",
    "        )\n",
    "\n",
    "        history = self.model_vit.fit(\n",
    "            x=self.x_train,\n",
    "            y=self.y_train,\n",
    "            batch_size=batch_size,\n",
    "            epochs=num_epochs,\n",
    "            validation_split=0.1,\n",
    "            callbacks=[checkpoint_callback],\n",
    "        )\n",
    "\n",
    "        self.model_vit.load_weights(checkpoint_filepath)\n",
    "        _, accuracy = self.model_vit.evaluate(self.x_test, self.y_test)\n",
    "        print(f\"Test accuracy: {round(accuracy * 100, 2)}%\")\n",
    "        \n",
    "        if save_path!= 'none':\n",
    "            self.model_vit.save_weights(save_path)\n",
    "            \n",
    "            with open(save_path+'_history', 'wb') as file_pi:\n",
    "                pickle.dump(history.history, file_pi)\n",
    "        \n",
    "        return history\n",
    "    \n",
    "    def load_model(self, path, plot=False):\n",
    "        \n",
    "        \"\"\"this funtion is to load the training weight of the visual transformer model\n",
    "        \n",
    "        Args:\n",
    "            path(string): model oject address, and it will be use for history loading file too\n",
    "        \n",
    "        Return:\n",
    "            self.model_vit(oject): the visual transformer model with loaded weight\n",
    "        \n",
    "        \"\"\"\n",
    "            \n",
    "        try:\n",
    "            del self.model_vit\n",
    "        except:\n",
    "            print(\"load model\")\n",
    "        \n",
    "        self.model_vit = self.create_model()\n",
    "        self.model_vit.load_weights(path)\n",
    "        lr_schedule = keras.optimizers.schedules.ExponentialDecay(initial_learning_rate=1e-3, decay_steps=10000, decay_rate=0.9)\n",
    "        opt = tf.keras.optimizers.Adam(lr_schedule)\n",
    "\n",
    "        self.model_vit.compile(optimizer=opt, loss='binary_crossentropy',metrics=keras.metrics.BinaryAccuracy(name=\"accuracy\"))\n",
    "        \n",
    "        with open(path+'_history', \"rb\") as file_pi:\n",
    "            history = pickle.load(file_pi)\n",
    "        \n",
    "        if (plot):\n",
    "            # Plot training and validation loss\n",
    "            plt.figure(dpi=100)  # You can adjust the figure size if needed\n",
    "            \n",
    "            plt.plot(history['val_accuracy'], label='Validation accuracy')\n",
    "            plt.plot(history['accuracy'], label='Training accuracy')\n",
    "            \n",
    "            plt.plot(history['val_loss'], label='Validation loss')\n",
    "            plt.plot(history['loss'], label='Training Loss')\n",
    "            \n",
    "            plt.legend()\n",
    "            plt.xlabel('Epochs')\n",
    "            plt.ylabel('Loss')\n",
    "            plt.show()\n",
    "\n",
    "        return self.model_vit\n",
    "    \n",
    "    def model_eval(self, model, x_test, y_test):\n",
    "        \n",
    "        \"\"\"this funtion is for model accuracy checking\"\"\"\n",
    "        \n",
    "        loss, accuracy = model.evaluate(x_test, y_test)\n",
    "        \n",
    "        print(\"loss: \", loss, \"accuracy: \", accuracy )\n",
    "        print(f\"Test accuracy: {round(accuracy*100, 2)}%\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6b0cab05",
   "metadata": {},
   "source": [
    "###  Briefly check on the patches"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "a4f38c13",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Image size: 52 X 52\n",
      "Patch size: 13 X 13\n",
      "Patches per image: 16\n",
      "Elements per patch: 169\n",
      "Num patch: 16\n",
      "(1, 16, 169)\n"
     ]
    },
    {
     "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": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 16 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "A_test_patches = vis_transformer(x_train, x_test, y_train, y_test)\n",
    "A_test_patches.see_patches(x_test, num_image=250)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "05575bc4",
   "metadata": {},
   "source": [
    "### ViT intitialize to train a new model or create the model for weight loading"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "e48633b3",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "set model-saving: C:\\Users\\admin\\Desktop\\PanithanS\\Wafer_defect\\model_vit_v1b\n",
      "set history-saving: C:\\Users\\admin\\Desktop\\PanithanS\\Wafer_defect\\model_vit_v1b_history\n",
      "Epoch 1/45\n",
      "54/54 [==============================] - 293s 5s/step - loss: 0.6839 - accuracy: 0.6611 - val_loss: 0.5090 - val_accuracy: 0.7012\n",
      "Epoch 2/45\n",
      "54/54 [==============================] - 307s 6s/step - loss: 0.5495 - accuracy: 0.7035 - val_loss: 0.5124 - val_accuracy: 0.7105\n",
      "Epoch 3/45\n",
      "54/54 [==============================] - 322s 6s/step - loss: 0.5125 - accuracy: 0.7079 - val_loss: 0.5017 - val_accuracy: 0.7105\n",
      "Epoch 4/45\n",
      "54/54 [==============================] - 330s 6s/step - loss: 0.5036 - accuracy: 0.7099 - val_loss: 0.4933 - val_accuracy: 0.7388\n",
      "Epoch 5/45\n",
      "54/54 [==============================] - 319s 6s/step - loss: 0.4717 - accuracy: 0.7291 - val_loss: 0.3476 - val_accuracy: 0.8236\n",
      "Epoch 6/45\n",
      "54/54 [==============================] - 313s 6s/step - loss: 0.3527 - accuracy: 0.8130 - val_loss: 0.2504 - val_accuracy: 0.8777\n",
      "Epoch 7/45\n",
      "54/54 [==============================] - 294s 5s/step - loss: 0.2713 - accuracy: 0.8660 - val_loss: 0.1850 - val_accuracy: 0.9222\n",
      "Epoch 8/45\n",
      "54/54 [==============================] - 274s 5s/step - loss: 0.2009 - accuracy: 0.9128 - val_loss: 0.1366 - val_accuracy: 0.9511\n",
      "Epoch 9/45\n",
      "54/54 [==============================] - 271s 5s/step - loss: 0.1667 - accuracy: 0.9331 - val_loss: 0.1099 - val_accuracy: 0.9589\n",
      "Epoch 10/45\n",
      "54/54 [==============================] - 283s 5s/step - loss: 0.1391 - accuracy: 0.9464 - val_loss: 0.0953 - val_accuracy: 0.9663\n",
      "Epoch 11/45\n",
      "54/54 [==============================] - 271s 5s/step - loss: 0.1199 - accuracy: 0.9554 - val_loss: 0.0790 - val_accuracy: 0.9723\n",
      "Epoch 12/45\n",
      "54/54 [==============================] - 269s 5s/step - loss: 0.1055 - accuracy: 0.9613 - val_loss: 0.0701 - val_accuracy: 0.9762\n",
      "Epoch 13/45\n",
      "54/54 [==============================] - 269s 5s/step - loss: 0.0926 - accuracy: 0.9665 - val_loss: 0.0673 - val_accuracy: 0.9764\n",
      "Epoch 14/45\n",
      "54/54 [==============================] - 272s 5s/step - loss: 0.0951 - accuracy: 0.9664 - val_loss: 0.0562 - val_accuracy: 0.9811\n",
      "Epoch 15/45\n",
      "54/54 [==============================] - 271s 5s/step - loss: 0.0811 - accuracy: 0.9718 - val_loss: 0.0750 - val_accuracy: 0.9741\n",
      "Epoch 16/45\n",
      "54/54 [==============================] - 269s 5s/step - loss: 0.0798 - accuracy: 0.9714 - val_loss: 0.0669 - val_accuracy: 0.9751\n",
      "Epoch 17/45\n",
      "54/54 [==============================] - 269s 5s/step - loss: 0.0720 - accuracy: 0.9752 - val_loss: 0.0550 - val_accuracy: 0.9822\n",
      "Epoch 18/45\n",
      "54/54 [==============================] - 276s 5s/step - loss: 0.0605 - accuracy: 0.9795 - val_loss: 0.0481 - val_accuracy: 0.9839\n",
      "Epoch 19/45\n",
      "54/54 [==============================] - 298s 6s/step - loss: 0.0609 - accuracy: 0.9796 - val_loss: 0.0559 - val_accuracy: 0.9818\n",
      "Epoch 20/45\n",
      "54/54 [==============================] - 316s 6s/step - loss: 0.0570 - accuracy: 0.9807 - val_loss: 0.0511 - val_accuracy: 0.9836\n",
      "Epoch 21/45\n",
      "54/54 [==============================] - 296s 5s/step - loss: 0.0549 - accuracy: 0.9813 - val_loss: 0.0435 - val_accuracy: 0.9851\n",
      "Epoch 22/45\n",
      "54/54 [==============================] - 307s 6s/step - loss: 0.0483 - accuracy: 0.9835 - val_loss: 0.0481 - val_accuracy: 0.9842\n",
      "Epoch 23/45\n",
      "54/54 [==============================] - 303s 6s/step - loss: 0.0520 - accuracy: 0.9827 - val_loss: 0.0441 - val_accuracy: 0.9853\n",
      "Epoch 24/45\n",
      "54/54 [==============================] - 306s 6s/step - loss: 0.0460 - accuracy: 0.9849 - val_loss: 0.0422 - val_accuracy: 0.9871\n",
      "Epoch 25/45\n",
      "54/54 [==============================] - 310s 6s/step - loss: 0.0414 - accuracy: 0.9864 - val_loss: 0.0604 - val_accuracy: 0.9793\n",
      "Epoch 26/45\n",
      "54/54 [==============================] - 313s 6s/step - loss: 0.0457 - accuracy: 0.9849 - val_loss: 0.0415 - val_accuracy: 0.9873\n",
      "Epoch 27/45\n",
      "54/54 [==============================] - 325s 6s/step - loss: 0.0422 - accuracy: 0.9860 - val_loss: 0.0447 - val_accuracy: 0.9866\n",
      "Epoch 28/45\n",
      "54/54 [==============================] - 318s 6s/step - loss: 0.0411 - accuracy: 0.9864 - val_loss: 0.0557 - val_accuracy: 0.9841\n",
      "Epoch 29/45\n",
      "54/54 [==============================] - 332s 6s/step - loss: 0.0484 - accuracy: 0.9839 - val_loss: 0.0552 - val_accuracy: 0.9827\n",
      "Epoch 30/45\n",
      "54/54 [==============================] - 312s 6s/step - loss: 0.0397 - accuracy: 0.9867 - val_loss: 0.0421 - val_accuracy: 0.9876\n",
      "Epoch 31/45\n",
      "54/54 [==============================] - 309s 6s/step - loss: 0.0346 - accuracy: 0.9884 - val_loss: 0.0336 - val_accuracy: 0.9906\n",
      "Epoch 32/45\n",
      "54/54 [==============================] - 311s 6s/step - loss: 0.0301 - accuracy: 0.9902 - val_loss: 0.0430 - val_accuracy: 0.9870\n",
      "Epoch 33/45\n",
      "54/54 [==============================] - 304s 6s/step - loss: 0.0334 - accuracy: 0.9892 - val_loss: 0.0398 - val_accuracy: 0.9882\n",
      "Epoch 34/45\n",
      "54/54 [==============================] - 314s 6s/step - loss: 0.0331 - accuracy: 0.9886 - val_loss: 0.0353 - val_accuracy: 0.9888\n",
      "Epoch 35/45\n",
      "54/54 [==============================] - 309s 6s/step - loss: 0.0292 - accuracy: 0.9904 - val_loss: 0.0407 - val_accuracy: 0.9876\n",
      "Epoch 36/45\n",
      "54/54 [==============================] - 309s 6s/step - loss: 0.0309 - accuracy: 0.9896 - val_loss: 0.0406 - val_accuracy: 0.9879\n",
      "Epoch 37/45\n",
      "54/54 [==============================] - 332s 6s/step - loss: 0.0283 - accuracy: 0.9907 - val_loss: 0.0414 - val_accuracy: 0.9873\n",
      "Epoch 38/45\n",
      "54/54 [==============================] - 397s 7s/step - loss: 0.0279 - accuracy: 0.9909 - val_loss: 0.0359 - val_accuracy: 0.9901\n",
      "Epoch 39/45\n",
      "54/54 [==============================] - 313s 6s/step - loss: 0.0245 - accuracy: 0.9919 - val_loss: 0.0409 - val_accuracy: 0.9875\n",
      "Epoch 40/45\n",
      "54/54 [==============================] - 339s 6s/step - loss: 0.0272 - accuracy: 0.9906 - val_loss: 0.0340 - val_accuracy: 0.9906\n",
      "Epoch 41/45\n",
      "54/54 [==============================] - 322s 6s/step - loss: 0.0229 - accuracy: 0.9926 - val_loss: 0.0444 - val_accuracy: 0.9884\n",
      "Epoch 42/45\n",
      "54/54 [==============================] - 279s 5s/step - loss: 0.0276 - accuracy: 0.9910 - val_loss: 0.0424 - val_accuracy: 0.9864\n",
      "Epoch 43/45\n",
      "54/54 [==============================] - 302s 6s/step - loss: 0.0269 - accuracy: 0.9910 - val_loss: 0.0400 - val_accuracy: 0.9894\n",
      "Epoch 44/45\n",
      "54/54 [==============================] - 312s 6s/step - loss: 0.0251 - accuracy: 0.9918 - val_loss: 0.0393 - val_accuracy: 0.9898\n",
      "Epoch 45/45\n",
      "54/54 [==============================] - 301s 6s/step - loss: 0.0233 - accuracy: 0.9923 - val_loss: 0.0384 - val_accuracy: 0.9900\n",
      "238/238 [==============================] - 22s 93ms/step - loss: 0.0314 - accuracy: 0.98990s - loss: 0.0315 - accuracy: \n",
      "Test accuracy: 98.99%\n"
     ]
    }
   ],
   "source": [
    "A2 = vis_transformer(x_train, x_test, y_train, y_test)\n",
    "\n",
    "# model weight file to be loaded\n",
    "path = r\"C:\\Users\\admin\\Desktop\\PanithanS\\Wafer_defect\\model_vit_v1b\"\n",
    "\n",
    "# If train new model\n",
    "history = A2.train_model(batch_size = 512, num_epochs = 45, save_path=path)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "48dd9ffd",
   "metadata": {},
   "source": [
    "### Load the trained model and evaluate accuracy of the trained model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "d93e93b4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 600x400 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "238/238 [==============================] - 24s 88ms/step - loss: 0.0318 - accuracy: 0.9895\n",
      "loss:  0.031736329197883606 accuracy:  0.9898066520690918\n",
      "Test accuracy: 98.98%\n"
     ]
    }
   ],
   "source": [
    "model = A2.load_model(path, plot=True)\n",
    "A2.model_eval(model, x_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "47be9a1f",
   "metadata": {},
   "source": [
    "# Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "d5ccbe97",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "238/238 [==============================] - 22s 83ms/step\n"
     ]
    }
   ],
   "source": [
    "prediction = model.predict(x_test, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "3b544749",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ground-truth: [0 1 0 0 0 0 1 0]\n",
      "Predictions : [0 1 0 0 0 0 1 0]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predictions : Donut, Scratch, \n"
     ]
    }
   ],
   "source": [
    "pred_img_num = 900\n",
    "image = np.squeeze(x_test[pred_img_num,:,:,0])\n",
    "#print(\"Model output:\", (prediction[pred_img_num,:]))\n",
    "print(\"Ground-truth:\", np.array(y_test[pred_img_num,:]))\n",
    "print(\"Predictions : [\" + \" \".join([\"{:.0f}\".format(val) for val in prediction[pred_img_num]]) + \"]\")\n",
    "\n",
    "plt.imshow(image)\n",
    "plt.show()\n",
    "\n",
    "label = np.array([\"{:.0f}\".format(val) for val in prediction[pred_img_num]]).astype(float)\n",
    "print(\"Predictions :\", A1.read_label(label))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "a6d0afe5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ground-truth: [0 1 1 0 1 0 1 0]\n",
      "Predictions : [0 1 1 0 1 0 1 0]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predictions : Donut, Edge_Loc, Loc, Scratch, \n"
     ]
    }
   ],
   "source": [
    "pred_img_num = 500\n",
    "image = np.squeeze(x_test[pred_img_num,:,:,0])\n",
    "#print(\"Model output:\", (prediction[pred_img_num,:]))\n",
    "print(\"Ground-truth:\", np.array(y_test[pred_img_num,:]))\n",
    "print(\"Predictions : [\" + \" \".join([\"{:.0f}\".format(val) for val in prediction[pred_img_num]]) + \"]\")\n",
    "\n",
    "plt.imshow(image)\n",
    "plt.show()\n",
    "\n",
    "label = np.array([\"{:.0f}\".format(val) for val in prediction[pred_img_num]]).astype(float)\n",
    "print(\"Predictions :\", A1.read_label(label))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "206a3ef7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ground-truth: [1 0 0 0 1 0 1 0]\n",
      "Predictions : [1 0 0 0 1 0 1 0]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predictions : Center, Loc, Scratch, \n"
     ]
    }
   ],
   "source": [
    "pred_img_num = 3000\n",
    "image = np.squeeze(x_test[pred_img_num,:,:,0])\n",
    "#print(\"Model output:\", (prediction[pred_img_num,:]))\n",
    "print(\"Ground-truth:\", np.array(y_test[pred_img_num,:]))\n",
    "print(\"Predictions : [\" + \" \".join([\"{:.0f}\".format(val) for val in prediction[pred_img_num]]) + \"]\")\n",
    "\n",
    "plt.imshow(image)\n",
    "plt.show()\n",
    "\n",
    "label = np.array([\"{:.0f}\".format(val) for val in prediction[pred_img_num]]).astype(float)\n",
    "print(\"Predictions :\", A1.read_label(label))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d88171f2",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}