resnet34 (2).ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "b740c315",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-06-10T10:55:37.714210Z",
     "iopub.status.busy": "2024-06-10T10:55:37.713942Z",
     "iopub.status.idle": "2024-06-10T10:55:51.811048Z",
     "shell.execute_reply": "2024-06-10T10:55:51.810273Z"
    },
    "papermill": {
     "duration": 14.105555,
     "end_time": "2024-06-10T10:55:51.813435",
     "exception": false,
     "start_time": "2024-06-10T10:55:37.707880",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2024-06-10 10:55:40.870636: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered\n",
      "2024-06-10 10:55:40.870736: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered\n",
      "2024-06-10 10:55:41.032218: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import os\n",
    "import cv2\n",
    "import random\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Conv2D, MaxPooling2D, Dense, Flatten, Dropout, Activation, Add, GlobalAveragePooling2D\n",
    "from tensorflow.keras import layers, models\n",
    "from tensorflow.keras.layers import Input, BatchNormalization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "c8552ba0",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-06-10T10:55:51.836118Z",
     "iopub.status.busy": "2024-06-10T10:55:51.835069Z",
     "iopub.status.idle": "2024-06-10T10:55:51.839615Z",
     "shell.execute_reply": "2024-06-10T10:55:51.838722Z"
    },
    "papermill": {
     "duration": 0.014117,
     "end_time": "2024-06-10T10:55:51.841432",
     "exception": false,
     "start_time": "2024-06-10T10:55:51.827315",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "dataset_path='/kaggle/input/realifake/Realifake'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "e75f5706",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-06-10T10:55:51.851810Z",
     "iopub.status.busy": "2024-06-10T10:55:51.851568Z",
     "iopub.status.idle": "2024-06-10T10:55:51.858223Z",
     "shell.execute_reply": "2024-06-10T10:55:51.857349Z"
    },
    "papermill": {
     "duration": 0.014015,
     "end_time": "2024-06-10T10:55:51.860077",
     "exception": false,
     "start_time": "2024-06-10T10:55:51.846062",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def projection_block(x, filters, strides,padding='same'):\n",
    "    shortcut = x\n",
    "    x = conv_block(x, filters=filters, kernel_size=(1, 1), strides=strides,padding='same')\n",
    "    x = conv_block(x, filters=filters, kernel_size=(3, 3), strides=(1, 1),padding='same')\n",
    "    x = Conv2D(filters=filters * 4, kernel_size=(1, 1),padding='same')(x)\n",
    "    x = BatchNormalization()(x)\n",
    "    shortcut = Conv2D(filters=filters * 4, kernel_size=(1, 1), strides=strides,padding='same')(shortcut)\n",
    "    shortcut = BatchNormalization()(shortcut)\n",
    "    x = Add()([x, shortcut])\n",
    "    x = Activation('relu')(x)\n",
    "    return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "e766498c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-06-10T10:55:51.870264Z",
     "iopub.status.busy": "2024-06-10T10:55:51.870029Z",
     "iopub.status.idle": "2024-06-10T10:55:51.876780Z",
     "shell.execute_reply": "2024-06-10T10:55:51.875984Z"
    },
    "papermill": {
     "duration": 0.01391,
     "end_time": "2024-06-10T10:55:51.878571",
     "exception": false,
     "start_time": "2024-06-10T10:55:51.864661",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def convolutional_block(x, filter):\n",
    "    # copy tensor to variable called x_skip\n",
    "    x_skip = x\n",
    "    # Layer 1\n",
    "    x = tf.keras.layers.Conv2D(filter, (3,3), padding = 'same', strides = (2,2))(x)\n",
    "    x = tf.keras.layers.BatchNormalization(axis=3)(x)\n",
    "    x = tf.keras.layers.Activation('relu')(x)\n",
    "    # Layer 2\n",
    "    x = tf.keras.layers.Conv2D(filter, (3,3), padding = 'same')(x)\n",
    "    x = tf.keras.layers.BatchNormalization(axis=3)(x)\n",
    "    # Processing Residue with conv(1,1)\n",
    "    x_skip = tf.keras.layers.Conv2D(filter, (1,1), strides = (2,2))(x_skip)\n",
    "    # Add Residue\n",
    "    x = tf.keras.layers.Add()([x, x_skip])     \n",
    "    x = tf.keras.layers.Activation('relu')(x)\n",
    "    return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "a800d2f3",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-06-10T10:55:51.888851Z",
     "iopub.status.busy": "2024-06-10T10:55:51.888614Z",
     "iopub.status.idle": "2024-06-10T10:55:51.894767Z",
     "shell.execute_reply": "2024-06-10T10:55:51.893988Z"
    },
    "papermill": {
     "duration": 0.013387,
     "end_time": "2024-06-10T10:55:51.896549",
     "exception": false,
     "start_time": "2024-06-10T10:55:51.883162",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def identity_block(x, filter):\n",
    "    # copy tensor to variable called x_skip\n",
    "    x_skip = x\n",
    "    # Layer 1\n",
    "    x = tf.keras.layers.Conv2D(filter, (3,3), padding = 'same')(x)\n",
    "    x = tf.keras.layers.BatchNormalization(axis=3)(x)\n",
    "    x = tf.keras.layers.Activation('relu')(x)\n",
    "    # Layer 2\n",
    "    x = tf.keras.layers.Conv2D(filter, (3,3), padding = 'same')(x)\n",
    "    x = tf.keras.layers.BatchNormalization(axis=3)(x)\n",
    "    # Add Residue\n",
    "    x = tf.keras.layers.Add()([x, x_skip])     \n",
    "    x = tf.keras.layers.Activation('relu')(x)\n",
    "    return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "b9f07514",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-06-10T10:55:51.906957Z",
     "iopub.status.busy": "2024-06-10T10:55:51.906678Z",
     "iopub.status.idle": "2024-06-10T10:55:51.917172Z",
     "shell.execute_reply": "2024-06-10T10:55:51.916242Z"
    },
    "papermill": {
     "duration": 0.017836,
     "end_time": "2024-06-10T10:55:51.919166",
     "exception": false,
     "start_time": "2024-06-10T10:55:51.901330",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def ResNet34(shape = (300, 300, 3), classes = 2):\n",
    "    # Step 1 (Setup Input Layer)\n",
    "    x_input = tf.keras.layers.Input(shape)\n",
    "    x = tf.keras.layers.ZeroPadding2D((3, 3))(x_input)\n",
    "    # Step 2 (Initial Conv layer along with maxPool)\n",
    "    x = tf.keras.layers.Conv2D(64, kernel_size=7, strides=2, padding='same')(x)\n",
    "    x = tf.keras.layers.BatchNormalization()(x)\n",
    "    x = tf.keras.layers.Activation('relu')(x)\n",
    "    x = tf.keras.layers.MaxPool2D(pool_size=3, strides=2, padding='same')(x)\n",
    "    # Define size of sub-blocks and initial filter size\n",
    "    block_layers = [3, 4, 6, 3]\n",
    "    filter_size = 64\n",
    "    # Step 3 Add the Resnet Blocks\n",
    "    for i in range(4):\n",
    "        if i == 0:\n",
    "            # For sub-block 1 Residual/Convolutional block not needed\n",
    "            for j in range(block_layers[i]):\n",
    "                x = identity_block(x, filter_size)\n",
    "        else:\n",
    "            # One Residual/Convolutional Block followed by Identity blocks\n",
    "            # The filter size will go on increasing by a factor of 2\n",
    "            filter_size = filter_size*2\n",
    "            x = convolutional_block(x, filter_size)\n",
    "            for j in range(block_layers[i] - 1):\n",
    "                x = identity_block(x, filter_size)\n",
    "     # Step 4 End Dense Network\n",
    "         # Step 4 End Dense Network\n",
    "    x = tf.keras.layers.AveragePooling2D((2,2), padding = 'same')(x)\n",
    "    x = tf.keras.layers.Flatten()(x)\n",
    "    x = tf.keras.layers.Dense(512, activation = 'relu')(x)\n",
    "    x = tf.keras.layers.Dense(1, activation = 'sigmoid')(x)\n",
    "    model = tf.keras.models.Model(inputs = x_input, outputs = x, name = \"ResNet34\")\n",
    "    return model\n",
    "      "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "be6a7604",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-06-10T10:55:51.929485Z",
     "iopub.status.busy": "2024-06-10T10:55:51.929187Z",
     "iopub.status.idle": "2024-06-10T10:55:51.932855Z",
     "shell.execute_reply": "2024-06-10T10:55:51.932040Z"
    },
    "papermill": {
     "duration": 0.011008,
     "end_time": "2024-06-10T10:55:51.934781",
     "exception": false,
     "start_time": "2024-06-10T10:55:51.923773",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "batch_size =256\n",
    "img_height = 300\n",
    "img_width = 300"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "91bce775",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-06-10T10:55:51.945558Z",
     "iopub.status.busy": "2024-06-10T10:55:51.944875Z",
     "iopub.status.idle": "2024-06-10T10:59:56.383359Z",
     "shell.execute_reply": "2024-06-10T10:59:56.382554Z"
    },
    "papermill": {
     "duration": 244.446204,
     "end_time": "2024-06-10T10:59:56.385688",
     "exception": false,
     "start_time": "2024-06-10T10:55:51.939484",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 174425 files belonging to 2 classes.\n",
      "Using 104655 files for training.\n"
     ]
    }
   ],
   "source": [
    "train_ds = tf.keras.preprocessing.image_dataset_from_directory(\n",
    "    dataset_path,\n",
    "    validation_split=0.4,\n",
    "    subset=\"training\",\n",
    "    seed=123,\n",
    "    image_size=(img_height, img_width),\n",
    "    batch_size=batch_size\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "a22c7dd8",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-06-10T10:59:56.398264Z",
     "iopub.status.busy": "2024-06-10T10:59:56.397955Z",
     "iopub.status.idle": "2024-06-10T11:00:47.128548Z",
     "shell.execute_reply": "2024-06-10T11:00:47.127571Z"
    },
    "papermill": {
     "duration": 50.739326,
     "end_time": "2024-06-10T11:00:47.130954",
     "exception": false,
     "start_time": "2024-06-10T10:59:56.391628",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 174425 files belonging to 2 classes.\n",
      "Using 69770 files for validation.\n"
     ]
    }
   ],
   "source": [
    "val_ds = tf.keras.preprocessing.image_dataset_from_directory(\n",
    "    dataset_path,\n",
    "    validation_split=0.4,\n",
    "    subset=\"validation\",\n",
    "    seed=123,\n",
    "    image_size=(img_height, img_width),\n",
    "    batch_size=batch_size\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "99dd2403",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-06-10T11:00:47.142456Z",
     "iopub.status.busy": "2024-06-10T11:00:47.142120Z",
     "iopub.status.idle": "2024-06-10T11:00:47.146731Z",
     "shell.execute_reply": "2024-06-10T11:00:47.145886Z"
    },
    "papermill": {
     "duration": 0.012254,
     "end_time": "2024-06-10T11:00:47.148544",
     "exception": false,
     "start_time": "2024-06-10T11:00:47.136290",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['FAKE', 'REAL']\n"
     ]
    }
   ],
   "source": [
    "# Check class names\n",
    "class_names = train_ds.class_names\n",
    "print(class_names)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "fca7e05a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-06-10T11:00:47.159698Z",
     "iopub.status.busy": "2024-06-10T11:00:47.159453Z",
     "iopub.status.idle": "2024-06-10T11:00:47.169455Z",
     "shell.execute_reply": "2024-06-10T11:00:47.168815Z"
    },
    "papermill": {
     "duration": 0.017713,
     "end_time": "2024-06-10T11:00:47.171316",
     "exception": false,
     "start_time": "2024-06-10T11:00:47.153603",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "test_ds = val_ds.take(int(len(val_ds) * 0.5))\n",
    "val_ds = val_ds.skip(int(len(val_ds) * 0.5))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "2c18eeb1",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-06-10T11:00:47.183104Z",
     "iopub.status.busy": "2024-06-10T11:00:47.182858Z",
     "iopub.status.idle": "2024-06-10T11:00:47.813427Z",
     "shell.execute_reply": "2024-06-10T11:00:47.812645Z"
    },
    "papermill": {
     "duration": 0.638727,
     "end_time": "2024-06-10T11:00:47.815720",
     "exception": false,
     "start_time": "2024-06-10T11:00:47.176993",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "model= ResNet34()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "7b285d9f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-06-10T11:00:47.827494Z",
     "iopub.status.busy": "2024-06-10T11:00:47.827186Z",
     "iopub.status.idle": "2024-06-10T11:00:47.839084Z",
     "shell.execute_reply": "2024-06-10T11:00:47.838436Z"
    },
    "papermill": {
     "duration": 0.019723,
     "end_time": "2024-06-10T11:00:47.840876",
     "exception": false,
     "start_time": "2024-06-10T11:00:47.821153",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "model.compile(optimizer='adam',\n",
    "              loss='binary_crossentropy',  # Binary cross-entropy for binary classification\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "50d03739",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-06-10T11:00:47.851927Z",
     "iopub.status.busy": "2024-06-10T11:00:47.851627Z",
     "iopub.status.idle": "2024-06-10T14:42:56.354460Z",
     "shell.execute_reply": "2024-06-10T14:42:56.353334Z"
    },
    "papermill": {
     "duration": 13328.716324,
     "end_time": "2024-06-10T14:42:56.562218",
     "exception": false,
     "start_time": "2024-06-10T11:00:47.845894",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/28\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2024-06-10 11:02:16.163256: E external/local_xla/xla/service/slow_operation_alarm.cc:65] Trying algorithm eng0{} for conv (f32[64,3,7,7]{3,2,1,0}, u8[0]{0}) custom-call(f32[256,3,301,301]{3,2,1,0}, f32[256,64,153,153]{3,2,1,0}), window={size=7x7 stride=2x2 pad=5_5x5_5}, dim_labels=bf01_oi01->bf01, custom_call_target=\"__cudnn$convBackwardFilter\", backend_config={\"conv_result_scale\":1,\"activation_mode\":\"kNone\",\"side_input_scale\":0,\"leakyrelu_alpha\":0} is taking a while...\n",
      "2024-06-10 11:02:22.053983: E external/local_xla/xla/service/slow_operation_alarm.cc:133] The operation took 6.890903373s\n",
      "Trying algorithm eng0{} for conv (f32[64,3,7,7]{3,2,1,0}, u8[0]{0}) custom-call(f32[256,3,301,301]{3,2,1,0}, f32[256,64,153,153]{3,2,1,0}), window={size=7x7 stride=2x2 pad=5_5x5_5}, dim_labels=bf01_oi01->bf01, custom_call_target=\"__cudnn$convBackwardFilter\", backend_config={\"conv_result_scale\":1,\"activation_mode\":\"kNone\",\"side_input_scale\":0,\"leakyrelu_alpha\":0} is taking a while...\n",
      "2024-06-10 11:02:26.011161: E external/local_xla/xla/service/slow_operation_alarm.cc:65] Trying algorithm eng0{} for conv (f32[64,3,7,7]{3,2,1,0}, u8[0]{0}) custom-call(f32[256,3,301,301]{3,2,1,0}, f32[256,64,153,153]{3,2,1,0}), window={size=7x7 stride=2x2 pad=5_5x5_5}, dim_labels=bf01_oi01->bf01, custom_call_target=\"__cudnn$convBackwardFilter\", backend_config={\"conv_result_scale\":1,\"activation_mode\":\"kNone\",\"side_input_scale\":0,\"leakyrelu_alpha\":0} is taking a while...\n",
      "2024-06-10 11:02:33.066723: E external/local_xla/xla/service/slow_operation_alarm.cc:133] The operation took 8.05566451s\n",
      "Trying algorithm eng0{} for conv (f32[64,3,7,7]{3,2,1,0}, u8[0]{0}) custom-call(f32[256,3,301,301]{3,2,1,0}, f32[256,64,153,153]{3,2,1,0}), window={size=7x7 stride=2x2 pad=5_5x5_5}, dim_labels=bf01_oi01->bf01, custom_call_target=\"__cudnn$convBackwardFilter\", backend_config={\"conv_result_scale\":1,\"activation_mode\":\"kNone\",\"side_input_scale\":0,\"leakyrelu_alpha\":0} is taking a while...\n",
      "WARNING: All log messages before absl::InitializeLog() is called are written to STDERR\n",
      "I0000 00:00:1718017389.705501      78 device_compiler.h:186] Compiled cluster using XLA!  This line is logged at most once for the lifetime of the process.\n",
      "W0000 00:00:1718017389.774181      78 graph_launch.cc:671] Fallback to op-by-op mode because memset node breaks graph update\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m408/409\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m2s\u001b[0m 2s/step - accuracy: 0.5697 - loss: 2.5148"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2024-06-10 11:20:49.751447: E external/local_xla/xla/service/slow_operation_alarm.cc:65] Trying algorithm eng0{} for conv (f32[64,3,7,7]{3,2,1,0}, u8[0]{0}) custom-call(f32[207,3,301,301]{3,2,1,0}, f32[207,64,153,153]{3,2,1,0}), window={size=7x7 stride=2x2 pad=5_5x5_5}, dim_labels=bf01_oi01->bf01, custom_call_target=\"__cudnn$convBackwardFilter\", backend_config={\"conv_result_scale\":1,\"activation_mode\":\"kNone\",\"side_input_scale\":0,\"leakyrelu_alpha\":0} is taking a while...\n",
      "2024-06-10 11:20:55.101066: E external/local_xla/xla/service/slow_operation_alarm.cc:133] The operation took 6.349796444s\n",
      "Trying algorithm eng0{} for conv (f32[64,3,7,7]{3,2,1,0}, u8[0]{0}) custom-call(f32[207,3,301,301]{3,2,1,0}, f32[207,64,153,153]{3,2,1,0}), window={size=7x7 stride=2x2 pad=5_5x5_5}, dim_labels=bf01_oi01->bf01, custom_call_target=\"__cudnn$convBackwardFilter\", backend_config={\"conv_result_scale\":1,\"activation_mode\":\"kNone\",\"side_input_scale\":0,\"leakyrelu_alpha\":0} is taking a while...\n",
      "2024-06-10 11:20:58.698806: E external/local_xla/xla/service/slow_operation_alarm.cc:65] Trying algorithm eng0{} for conv (f32[64,3,7,7]{3,2,1,0}, u8[0]{0}) custom-call(f32[207,3,301,301]{3,2,1,0}, f32[207,64,153,153]{3,2,1,0}), window={size=7x7 stride=2x2 pad=5_5x5_5}, dim_labels=bf01_oi01->bf01, custom_call_target=\"__cudnn$convBackwardFilter\", backend_config={\"conv_result_scale\":1,\"activation_mode\":\"kNone\",\"side_input_scale\":0,\"leakyrelu_alpha\":0} is taking a while...\n",
      "2024-06-10 11:21:03.115727: E external/local_xla/xla/service/slow_operation_alarm.cc:133] The operation took 5.417018829s\n",
      "Trying algorithm eng0{} for conv (f32[64,3,7,7]{3,2,1,0}, u8[0]{0}) custom-call(f32[207,3,301,301]{3,2,1,0}, f32[207,64,153,153]{3,2,1,0}), window={size=7x7 stride=2x2 pad=5_5x5_5}, dim_labels=bf01_oi01->bf01, custom_call_target=\"__cudnn$convBackwardFilter\", backend_config={\"conv_result_scale\":1,\"activation_mode\":\"kNone\",\"side_input_scale\":0,\"leakyrelu_alpha\":0} is taking a while...\n",
      "W0000 00:00:1718018494.191483      78 graph_launch.cc:671] Fallback to op-by-op mode because memset node breaks graph update\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m409/409\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 3s/step - accuracy: 0.5698 - loss: 2.5110"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "W0000 00:00:1718018569.397062      78 graph_launch.cc:671] Fallback to op-by-op mode because memset node breaks graph update\n",
      "W0000 00:00:1718018711.812674      77 graph_launch.cc:671] Fallback to op-by-op mode because memset node breaks graph update\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m409/409\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1465s\u001b[0m 3s/step - accuracy: 0.5700 - loss: 2.5072 - val_accuracy: 0.7031 - val_loss: 0.5771\n",
      "Epoch 2/28\n",
      "\u001b[1m409/409\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1158s\u001b[0m 3s/step - accuracy: 0.8239 - loss: 0.3674 - val_accuracy: 0.8169 - val_loss: 0.3987\n",
      "Epoch 3/28\n",
      "\u001b[1m409/409\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1165s\u001b[0m 3s/step - accuracy: 0.8992 - loss: 0.2202 - val_accuracy: 0.8693 - val_loss: 0.2879\n",
      "Epoch 4/28\n",
      "\u001b[1m409/409\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1164s\u001b[0m 3s/step - accuracy: 0.9180 - loss: 0.1809 - val_accuracy: 0.7939 - val_loss: 0.4151\n",
      "Epoch 5/28\n",
      "\u001b[1m409/409\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1163s\u001b[0m 3s/step - accuracy: 0.9274 - loss: 0.1591 - val_accuracy: 0.8292 - val_loss: 0.3729\n",
      "Epoch 6/28\n",
      "\u001b[1m409/409\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1223s\u001b[0m 3s/step - accuracy: 0.9333 - loss: 0.1445 - val_accuracy: 0.8929 - val_loss: 0.2235\n",
      "Epoch 7/28\n",
      "\u001b[1m409/409\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1221s\u001b[0m 3s/step - accuracy: 0.9397 - loss: 0.1316 - val_accuracy: 0.8989 - val_loss: 0.2583\n",
      "Epoch 8/28\n",
      "\u001b[1m409/409\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1222s\u001b[0m 3s/step - accuracy: 0.9466 - loss: 0.1177 - val_accuracy: 0.8464 - val_loss: 0.3369\n",
      "Epoch 9/28\n",
      "\u001b[1m409/409\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1163s\u001b[0m 3s/step - accuracy: 0.9520 - loss: 0.1075 - val_accuracy: 0.8867 - val_loss: 0.3307\n",
      "Epoch 10/28\n",
      "\u001b[1m409/409\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1223s\u001b[0m 3s/step - accuracy: 0.9567 - loss: 0.0982 - val_accuracy: 0.8933 - val_loss: 0.3080\n",
      "Epoch 11/28\n",
      "\u001b[1m409/409\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1163s\u001b[0m 3s/step - accuracy: 0.9627 - loss: 0.0863 - val_accuracy: 0.9148 - val_loss: 0.2334\n"
     ]
    }
   ],
   "source": [
    "epochs = 28\n",
    "early_stopping = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=5, restore_best_weights=True)\n",
    "\n",
    "history = model.fit(\n",
    "    train_ds,\n",
    "    validation_data=val_ds,\n",
    "    epochs=epochs,\n",
    "    callbacks=[early_stopping]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "8d8fe67c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-06-10T14:42:57.291526Z",
     "iopub.status.busy": "2024-06-10T14:42:57.291177Z",
     "iopub.status.idle": "2024-06-10T14:42:57.772946Z",
     "shell.execute_reply": "2024-06-10T14:42:57.772002Z"
    },
    "papermill": {
     "duration": 0.845871,
     "end_time": "2024-06-10T14:42:57.774900",
     "exception": false,
     "start_time": "2024-06-10T14:42:56.929029",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot training & validation accuracy values\n",
    "plt.figure(figsize=(12, 4))\n",
    "plt.subplot(1, 2, 1)\n",
    "plt.plot(history.history['accuracy'])\n",
    "plt.plot(history.history['val_accuracy'])\n",
    "plt.title('Model accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Validation'], loc='upper left')\n",
    "\n",
    "# Plot training & validation loss values\n",
    "plt.subplot(1, 2, 2)\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Validation'], loc='upper left')\n",
    "\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kaggle": {
   "accelerator": "nvidiaTeslaT4",
   "dataSources": [
    {
     "datasetId": 3531487,
     "sourceId": 6156382,
     "sourceType": "datasetVersion"
    }
   ],
   "dockerImageVersionId": 30733,
   "isGpuEnabled": true,
   "isInternetEnabled": false,
   "language": "python",
   "sourceType": "notebook"
  },
  "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": "3.10.13"
  },
  "papermill": {
   "default_parameters": {},
   "duration": 13647.159367,
   "end_time": "2024-06-10T14:43:01.937002",
   "environment_variables": {},
   "exception": null,
   "input_path": "__notebook__.ipynb",
   "output_path": "__notebook__.ipynb",
   "parameters": {},
   "start_time": "2024-06-10T10:55:34.777635",
   "version": "2.5.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}