image-forgery-detector.ipynb

{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.10.13","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"none","dataSources":[{"sourceId":115146,"sourceType":"datasetVersion","datasetId":59500},{"sourceId":2065466,"sourceType":"datasetVersion","datasetId":1237999}],"dockerImageVersionId":30664,"isInternetEnabled":false,"language":"python","sourceType":"notebook","isGpuEnabled":false}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import os\nimport numpy as np\nimport matplotlib.pyplot as plt\nimport tensorflow as tf\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.metrics import confusion_matrix\nfrom tensorflow.keras.utils import to_categorical\nfrom keras.models import Sequential\nfrom keras.layers import Dense, Flatten, Conv2D, MaxPool2D, Dropout\nfrom keras.optimizers import Adam\nfrom tensorflow.keras.preprocessing.image import ImageDataGenerator\nfrom keras.callbacks import EarlyStopping\nfrom PIL import Image, ImageChops, ImageEnhance\nimport os\nimport itertools","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","execution":{"iopub.status.busy":"2024-03-04T07:37:33.264526Z","iopub.execute_input":"2024-03-04T07:37:33.265120Z","iopub.status.idle":"2024-03-04T07:37:49.381435Z","shell.execute_reply.started":"2024-03-04T07:37:33.265086Z","shell.execute_reply":"2024-03-04T07:37:49.380243Z"},"trusted":true},"execution_count":1,"outputs":[{"name":"stderr","text":"2024-03-04 07:37:35.371364: 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\n2024-03-04 07:37:35.371499: 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\n2024-03-04 07:37:35.519151: 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","output_type":"stream"}]},{"cell_type":"code","source":"def convert_to_ela_image(path, quality):\n    temp_filename = 'temp_file_name.jpg'\n    ela_filename = 'temp_ela.png'\n    \n    image = Image.open(path).convert('RGB')\n    image.save(temp_filename, 'JPEG', quality = quality)\n    temp_image = Image.open(temp_filename)\n    \n    ela_image = ImageChops.difference(image, temp_image)\n    \n    extrema = ela_image.getextrema()\n    max_diff = max([ex[1] for ex in extrema])\n    if max_diff == 0:\n        max_diff = 1\n    scale = 255.0 / max_diff\n    \n    ela_image = ImageEnhance.Brightness(ela_image).enhance(scale)\n    \n    return ela_image","metadata":{"execution":{"iopub.status.busy":"2024-03-04T07:37:49.383767Z","iopub.execute_input":"2024-03-04T07:37:49.384963Z","iopub.status.idle":"2024-03-04T07:37:49.393847Z","shell.execute_reply.started":"2024-03-04T07:37:49.384900Z","shell.execute_reply":"2024-03-04T07:37:49.392564Z"},"trusted":true},"execution_count":2,"outputs":[]},{"cell_type":"code","source":"def compute_ela_cv(path, quality):\n    temp_filename = 'temp_file_name.jpg'\n    SCALE = 15\n    orig_img = cv2.imread(path)\n    orig_img = cv2.cvtColor(orig_img, cv2.COLOR_BGR2RGB)\n    \n    cv2.imwrite(temp_filename, orig_img, [cv2.IMWRITE_JPEG_QUALITY, quality])\n\n    # read compressed image\n    compressed_img = cv2.imread(temp_filename)\n\n    # get absolute difference between img1 and img2 and multiply by scale\n    diff = SCALE * cv2.absdiff(orig_img, compressed_img)\n    return diff","metadata":{"execution":{"iopub.status.busy":"2024-03-04T07:37:49.395715Z","iopub.execute_input":"2024-03-04T07:37:49.396205Z","iopub.status.idle":"2024-03-04T07:37:49.404005Z","shell.execute_reply.started":"2024-03-04T07:37:49.396173Z","shell.execute_reply":"2024-03-04T07:37:49.402981Z"},"trusted":true},"execution_count":3,"outputs":[]},{"cell_type":"code","source":"image_size = (128, 128)\ndef prepare_image(image_path):\n    return np.array(convert_to_ela_image(image_path, 91).resize(image_size)).flatten() / 255.0","metadata":{"execution":{"iopub.status.busy":"2024-03-04T07:37:49.405570Z","iopub.execute_input":"2024-03-04T07:37:49.405920Z","iopub.status.idle":"2024-03-04T07:37:49.414629Z","shell.execute_reply.started":"2024-03-04T07:37:49.405892Z","shell.execute_reply":"2024-03-04T07:37:49.413612Z"},"trusted":true},"execution_count":4,"outputs":[]},{"cell_type":"code","source":"X = [] # ELA converted images\nY = [] # 0 for fake, 1 for real","metadata":{"execution":{"iopub.status.busy":"2024-03-04T07:37:49.417522Z","iopub.execute_input":"2024-03-04T07:37:49.417871Z","iopub.status.idle":"2024-03-04T07:37:49.424252Z","shell.execute_reply.started":"2024-03-04T07:37:49.417842Z","shell.execute_reply":"2024-03-04T07:37:49.423432Z"},"trusted":true},"execution_count":5,"outputs":[]},{"cell_type":"code","source":"import random\npath = '../input/casia-dataset/CASIA2/Au/'\nfor dirname, _, filenames in os.walk(path):\n    for filename in filenames:\n        if filename.endswith('jpg') or filename.endswith('png'):\n            full_path = os.path.join(dirname, filename)\n            X.append(prepare_image(full_path))\n            Y.append(1)\n            if len(Y) % 500 == 0:\n                print(f'Processing {len(Y)} images')\n\nrandom.shuffle(X)\nX = X[:2100]\nY = Y[:2100]\nprint(len(X), len(Y))","metadata":{"execution":{"iopub.status.busy":"2024-03-04T07:37:49.425257Z","iopub.execute_input":"2024-03-04T07:37:49.425558Z","iopub.status.idle":"2024-03-04T07:40:29.271315Z","shell.execute_reply.started":"2024-03-04T07:37:49.425528Z","shell.execute_reply":"2024-03-04T07:40:29.270098Z"},"trusted":true},"execution_count":6,"outputs":[{"name":"stdout","text":"Processing 500 images\nProcessing 1000 images\nProcessing 1500 images\nProcessing 2000 images\nProcessing 2500 images\nProcessing 3000 images\nProcessing 3500 images\nProcessing 4000 images\nProcessing 4500 images\nProcessing 5000 images\nProcessing 5500 images\nProcessing 6000 images\nProcessing 6500 images\nProcessing 7000 images\n2100 2100\n","output_type":"stream"}]},{"cell_type":"code","source":"path = '../input/casia-dataset/CASIA2/Tp/'\nfor dirname, _, filenames in os.walk(path):\n    for filename in filenames:\n        if filename.endswith('jpg') or filename.endswith('png'):\n            full_path = os.path.join(dirname, filename)\n            X.append(prepare_image(full_path))\n            Y.append(0)\n            if len(Y) % 500 == 0:\n                print(f'Processing {len(Y)} images')\n\nprint(len(X), len(Y))","metadata":{"execution":{"iopub.status.busy":"2024-03-04T07:40:29.272632Z","iopub.execute_input":"2024-03-04T07:40:29.272980Z","iopub.status.idle":"2024-03-04T07:41:23.318801Z","shell.execute_reply.started":"2024-03-04T07:40:29.272951Z","shell.execute_reply":"2024-03-04T07:41:23.316406Z"},"trusted":true},"execution_count":7,"outputs":[{"name":"stdout","text":"Processing 2500 images\nProcessing 3000 images\nProcessing 3500 images\nProcessing 4000 images\n4164 4164\n","output_type":"stream"}]},{"cell_type":"code","source":"X = np.array(X)\nY = to_categorical(Y, 2)\nX = X.reshape(-1, 128, 128, 3)","metadata":{"execution":{"iopub.status.busy":"2024-03-04T07:41:23.320039Z","iopub.execute_input":"2024-03-04T07:41:23.320788Z","iopub.status.idle":"2024-03-04T07:41:24.143762Z","shell.execute_reply.started":"2024-03-04T07:41:23.320752Z","shell.execute_reply":"2024-03-04T07:41:24.142433Z"},"trusted":true},"execution_count":8,"outputs":[]},{"cell_type":"code","source":"X_train, X_val, Y_train, Y_val = train_test_split(X, Y, test_size = 0.2, random_state=5)\nX = X.reshape(-1,1,1,1)\nprint(len(X_train), len(Y_train))\nprint(len(X_val), len(Y_val))","metadata":{"execution":{"iopub.status.busy":"2024-03-04T07:41:24.145368Z","iopub.execute_input":"2024-03-04T07:41:24.145713Z","iopub.status.idle":"2024-03-04T07:41:24.826040Z","shell.execute_reply.started":"2024-03-04T07:41:24.145686Z","shell.execute_reply":"2024-03-04T07:41:24.824899Z"},"trusted":true},"execution_count":9,"outputs":[{"name":"stdout","text":"3331 3331\n833 833\n","output_type":"stream"}]},{"cell_type":"code","source":"def build_model():\n    model = Sequential()\n    model.add(Conv2D(filters = 32, kernel_size = (5, 5), padding = 'valid', activation = 'relu', input_shape = (128, 128, 3)))\n    model.add(Conv2D(filters = 32, kernel_size = (5, 5), padding = 'valid', activation = 'relu', input_shape = (128, 128, 3)))\n    model.add(MaxPool2D(pool_size = (2, 2)))\n    model.add(Dropout(0.25))\n    model.add(Flatten())\n    model.add(Dense(256, activation = 'relu'))\n    model.add(Dropout(0.5))\n    model.add(Dense(2, activation = 'softmax'))\n    return model","metadata":{"execution":{"iopub.status.busy":"2024-03-04T07:41:24.827346Z","iopub.execute_input":"2024-03-04T07:41:24.827688Z","iopub.status.idle":"2024-03-04T07:41:24.836063Z","shell.execute_reply.started":"2024-03-04T07:41:24.827632Z","shell.execute_reply":"2024-03-04T07:41:24.834922Z"},"trusted":true},"execution_count":10,"outputs":[]},{"cell_type":"code","source":"model = build_model()\nmodel.summary()","metadata":{"execution":{"iopub.status.busy":"2024-03-04T07:41:24.837634Z","iopub.execute_input":"2024-03-04T07:41:24.838064Z","iopub.status.idle":"2024-03-04T07:41:25.102090Z","shell.execute_reply.started":"2024-03-04T07:41:24.838023Z","shell.execute_reply":"2024-03-04T07:41:25.100957Z"},"trusted":true},"execution_count":11,"outputs":[{"name":"stderr","text":"/opt/conda/lib/python3.10/site-packages/keras/src/layers/convolutional/base_conv.py:99: UserWarning: Do not pass an `input_shape`/`input_dim` argument to a layer. When using Sequential models, prefer using an `Input(shape)` object as the first layer in the model instead.\n  super().__init__(\n","output_type":"stream"},{"output_type":"display_data","data":{"text/plain":"\u001b[1mModel: \"sequential\"\u001b[0m\n","text/html":"<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"sequential\"</span>\n</pre>\n"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n┃\u001b[1m \u001b[0m\u001b[1mLayer (type)                   \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape          \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m      Param #\u001b[0m\u001b[1m \u001b[0m┃\n┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n│ conv2d (\u001b[38;5;33mConv2D\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m124\u001b[0m, \u001b[38;5;34m124\u001b[0m, \u001b[38;5;34m32\u001b[0m)   │         \u001b[38;5;34m2,432\u001b[0m │\n├─────────────────────────────────┼────────────────────────┼───────────────┤\n│ conv2d_1 (\u001b[38;5;33mConv2D\u001b[0m)               │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m120\u001b[0m, \u001b[38;5;34m120\u001b[0m, \u001b[38;5;34m32\u001b[0m)   │        \u001b[38;5;34m25,632\u001b[0m │\n├─────────────────────────────────┼────────────────────────┼───────────────┤\n│ max_pooling2d (\u001b[38;5;33mMaxPooling2D\u001b[0m)    │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m60\u001b[0m, \u001b[38;5;34m60\u001b[0m, \u001b[38;5;34m32\u001b[0m)     │             \u001b[38;5;34m0\u001b[0m │\n├─────────────────────────────────┼────────────────────────┼───────────────┤\n│ dropout (\u001b[38;5;33mDropout\u001b[0m)               │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m60\u001b[0m, \u001b[38;5;34m60\u001b[0m, \u001b[38;5;34m32\u001b[0m)     │             \u001b[38;5;34m0\u001b[0m │\n├─────────────────────────────────┼────────────────────────┼───────────────┤\n│ flatten (\u001b[38;5;33mFlatten\u001b[0m)               │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m115200\u001b[0m)         │             \u001b[38;5;34m0\u001b[0m │\n├─────────────────────────────────┼────────────────────────┼───────────────┤\n│ dense (\u001b[38;5;33mDense\u001b[0m)                   │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │    \u001b[38;5;34m29,491,456\u001b[0m │\n├─────────────────────────────────┼────────────────────────┼───────────────┤\n│ dropout_1 (\u001b[38;5;33mDropout\u001b[0m)             │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n├─────────────────────────────────┼────────────────────────┼───────────────┤\n│ dense_1 (\u001b[38;5;33mDense\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m2\u001b[0m)              │           \u001b[38;5;34m514\u001b[0m │\n└─────────────────────────────────┴────────────────────────┴───────────────┘\n","text/html":"<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n┃<span style=\"font-weight: bold\"> Layer (type)                    </span>┃<span style=\"font-weight: bold\"> Output Shape           </span>┃<span style=\"font-weight: bold\">       Param # </span>┃\n┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n│ conv2d (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Conv2D</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">124</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">124</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>)   │         <span style=\"color: #00af00; text-decoration-color: #00af00\">2,432</span> │\n├─────────────────────────────────┼────────────────────────┼───────────────┤\n│ conv2d_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Conv2D</span>)               │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">120</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">120</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>)   │        <span style=\"color: #00af00; text-decoration-color: #00af00\">25,632</span> │\n├─────────────────────────────────┼────────────────────────┼───────────────┤\n│ max_pooling2d (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">MaxPooling2D</span>)    │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">60</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">60</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>)     │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n├─────────────────────────────────┼────────────────────────┼───────────────┤\n│ dropout (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>)               │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">60</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">60</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>)     │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n├─────────────────────────────────┼────────────────────────┼───────────────┤\n│ flatten (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Flatten</span>)               │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">115200</span>)         │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n├─────────────────────────────────┼────────────────────────┼───────────────┤\n│ dense (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                   │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)            │    <span style=\"color: #00af00; text-decoration-color: #00af00\">29,491,456</span> │\n├─────────────────────────────────┼────────────────────────┼───────────────┤\n│ dropout_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>)             │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n├─────────────────────────────────┼────────────────────────┼───────────────┤\n│ dense_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">2</span>)              │           <span style=\"color: #00af00; text-decoration-color: #00af00\">514</span> │\n└─────────────────────────────────┴────────────────────────┴───────────────┘\n</pre>\n"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"\u001b[1m Total params: \u001b[0m\u001b[38;5;34m29,520,034\u001b[0m (112.61 MB)\n","text/html":"<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">29,520,034</span> (112.61 MB)\n</pre>\n"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m29,520,034\u001b[0m (112.61 MB)\n","text/html":"<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">29,520,034</span> (112.61 MB)\n</pre>\n"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n","text/html":"<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n</pre>\n"},"metadata":{}}]},{"cell_type":"code","source":"from keras import optimizers\nfrom keras.optimizers import Adam\n\n\nmodel.compile(loss='categorical_crossentropy', optimizer='Nadam', metrics=['accuracy'])\nepochs = 24\nbatch_size = 32\ninit_lr = 1e-4\noptimizer = Adam(learning_rate=init_lr)\n","metadata":{"execution":{"iopub.status.busy":"2024-03-04T07:41:25.103891Z","iopub.execute_input":"2024-03-04T07:41:25.104297Z","iopub.status.idle":"2024-03-04T07:41:25.123758Z","shell.execute_reply.started":"2024-03-04T07:41:25.104241Z","shell.execute_reply":"2024-03-04T07:41:25.122714Z"},"trusted":true},"execution_count":12,"outputs":[]},{"cell_type":"code","source":"early_stopping = EarlyStopping(monitor = 'val_acc',\n                              min_delta = 0,\n                              patience = 2,\n                              verbose = 0,\n                              mode = 'auto')","metadata":{"execution":{"iopub.status.busy":"2024-03-04T07:41:25.125634Z","iopub.execute_input":"2024-03-04T07:41:25.126742Z","iopub.status.idle":"2024-03-04T07:41:25.132410Z","shell.execute_reply.started":"2024-03-04T07:41:25.126695Z","shell.execute_reply":"2024-03-04T07:41:25.131304Z"},"trusted":true},"execution_count":13,"outputs":[]},{"cell_type":"code","source":"from tensorflow import keras\nfrom tensorflow.keras.preprocessing.image import ImageDataGenerator\nimport numpy as np\n\n# Assuming you have defined and compiled your model before this point\n# model = ...\n\nx_train2 = np.array(X_train, copy=True) \ny_train2 = np.array(Y_train, copy=True) \n\ndatagen = ImageDataGenerator(\n    featurewise_center=True,\n    featurewise_std_normalization=True,\n    rotation_range=10,\n    fill_mode='nearest',\n    validation_split=0.2\n)\n\ndatagen.fit(x_train2)  # Use x_train2 instead of X_train\n\n# Early stopping callback\nearly_stopping = keras.callbacks.EarlyStopping(monitor='val_loss', min_delta=0, patience=30, verbose=0, mode='min')\n\n# Fit the model using the fit method\nhistory = model.fit(datagen.flow(x_train2, y_train2, batch_size=32, subset='training'),\n                    epochs=epochs,\n                    validation_data=datagen.flow(x_train2, y_train2, batch_size=32, subset='validation'),\n                    verbose=1,\n                    callbacks=[early_stopping])\n","metadata":{"execution":{"iopub.status.busy":"2024-03-04T07:41:25.138460Z","iopub.execute_input":"2024-03-04T07:41:25.139235Z","iopub.status.idle":"2024-03-04T08:57:51.878215Z","shell.execute_reply.started":"2024-03-04T07:41:25.139199Z","shell.execute_reply":"2024-03-04T08:57:51.877265Z"},"trusted":true},"execution_count":14,"outputs":[{"name":"stdout","text":"Epoch 1/24\n","output_type":"stream"},{"name":"stderr","text":"/opt/conda/lib/python3.10/site-packages/keras/src/trainers/data_adapters/py_dataset_adapter.py:122: UserWarning: Your `PyDataset` class should call `super().__init__(**kwargs)` in its constructor. `**kwargs` can include `workers`, `use_multiprocessing`, `max_queue_size`. Do not pass these arguments to `fit()`, as they will be ignored.\n  self._warn_if_super_not_called()\n","output_type":"stream"},{"name":"stdout","text":"\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m167s\u001b[0m 2s/step - accuracy: 0.7326 - loss: 3.3061 - val_accuracy: 0.8453 - val_loss: 0.4766\nEpoch 2/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m201s\u001b[0m 2s/step - accuracy: 0.8461 - loss: 0.3987 - val_accuracy: 0.8904 - val_loss: 0.3244\nEpoch 3/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m202s\u001b[0m 2s/step - accuracy: 0.8993 - loss: 0.2960 - val_accuracy: 0.8814 - val_loss: 0.3116\nEpoch 4/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m202s\u001b[0m 2s/step - accuracy: 0.8984 - loss: 0.2665 - val_accuracy: 0.8904 - val_loss: 0.3082\nEpoch 5/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m202s\u001b[0m 2s/step - accuracy: 0.9246 - loss: 0.2119 - val_accuracy: 0.9009 - val_loss: 0.3044\nEpoch 6/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m164s\u001b[0m 2s/step - accuracy: 0.9381 - loss: 0.2041 - val_accuracy: 0.8994 - val_loss: 0.2783\nEpoch 7/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m164s\u001b[0m 2s/step - accuracy: 0.9462 - loss: 0.1489 - val_accuracy: 0.9054 - val_loss: 0.2944\nEpoch 8/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m201s\u001b[0m 2s/step - accuracy: 0.9529 - loss: 0.1449 - val_accuracy: 0.8979 - val_loss: 0.2833\nEpoch 9/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m163s\u001b[0m 2s/step - accuracy: 0.9502 - loss: 0.1307 - val_accuracy: 0.9024 - val_loss: 0.3187\nEpoch 10/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m201s\u001b[0m 2s/step - accuracy: 0.9604 - loss: 0.1184 - val_accuracy: 0.8889 - val_loss: 0.3325\nEpoch 11/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m162s\u001b[0m 2s/step - accuracy: 0.9570 - loss: 0.1121 - val_accuracy: 0.9024 - val_loss: 0.2786\nEpoch 12/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m202s\u001b[0m 2s/step - accuracy: 0.9650 - loss: 0.0987 - val_accuracy: 0.9024 - val_loss: 0.4633\nEpoch 13/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m202s\u001b[0m 2s/step - accuracy: 0.8575 - loss: 0.6464 - val_accuracy: 0.8919 - val_loss: 0.3338\nEpoch 14/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m163s\u001b[0m 2s/step - accuracy: 0.9377 - loss: 0.1802 - val_accuracy: 0.8754 - val_loss: 0.3425\nEpoch 15/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m202s\u001b[0m 2s/step - accuracy: 0.9411 - loss: 0.1789 - val_accuracy: 0.9054 - val_loss: 0.3208\nEpoch 16/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m203s\u001b[0m 2s/step - accuracy: 0.9444 - loss: 0.1459 - val_accuracy: 0.9024 - val_loss: 0.3441\nEpoch 17/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m202s\u001b[0m 2s/step - accuracy: 0.9446 - loss: 0.1443 - val_accuracy: 0.8994 - val_loss: 0.3104\nEpoch 18/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m203s\u001b[0m 2s/step - accuracy: 0.9661 - loss: 0.0950 - val_accuracy: 0.8829 - val_loss: 0.3622\nEpoch 19/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m201s\u001b[0m 2s/step - accuracy: 0.9703 - loss: 0.0929 - val_accuracy: 0.8829 - val_loss: 0.3103\nEpoch 20/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m202s\u001b[0m 2s/step - accuracy: 0.9715 - loss: 0.0837 - val_accuracy: 0.8799 - val_loss: 0.4041\nEpoch 21/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m202s\u001b[0m 2s/step - accuracy: 0.9698 - loss: 0.0839 - val_accuracy: 0.9024 - val_loss: 0.3399\nEpoch 22/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m202s\u001b[0m 2s/step - accuracy: 0.9737 - loss: 0.0735 - val_accuracy: 0.9099 - val_loss: 0.4256\nEpoch 23/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m164s\u001b[0m 2s/step - accuracy: 0.9745 - loss: 0.0786 - val_accuracy: 0.8994 - val_loss: 0.3880\nEpoch 24/24\n\u001b[1m84/84\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m202s\u001b[0m 2s/step - accuracy: 0.9786 - loss: 0.0606 - val_accuracy: 0.8784 - val_loss: 0.4705\n","output_type":"stream"}]},{"cell_type":"code","source":"model.compile(optimizer = optimizer, loss = 'binary_crossentropy', metrics = ['accuracy'])","metadata":{"execution":{"iopub.status.busy":"2024-03-04T08:57:51.879470Z","iopub.execute_input":"2024-03-04T08:57:51.880140Z","iopub.status.idle":"2024-03-04T08:57:51.888260Z","shell.execute_reply.started":"2024-03-04T08:57:51.880107Z","shell.execute_reply":"2024-03-04T08:57:51.887147Z"},"trusted":true},"execution_count":15,"outputs":[]},{"cell_type":"code","source":"hist = model.fit(X_train,\n                 Y_train,\n                 batch_size = batch_size,\n                 epochs = epochs,\n                validation_data = (X_val, Y_val),\n                callbacks = [early_stopping])","metadata":{"execution":{"iopub.status.busy":"2024-03-04T08:57:51.892091Z","iopub.execute_input":"2024-03-04T08:57:51.892449Z","iopub.status.idle":"2024-03-04T10:16:43.439277Z","shell.execute_reply.started":"2024-03-04T08:57:51.892418Z","shell.execute_reply":"2024-03-04T10:16:43.437799Z"},"trusted":true},"execution_count":16,"outputs":[{"name":"stdout","text":"Epoch 1/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m192s\u001b[0m 2s/step - accuracy: 0.5591 - loss: 0.7323 - val_accuracy: 0.7575 - val_loss: 0.5544\nEpoch 2/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m201s\u001b[0m 2s/step - accuracy: 0.7283 - loss: 0.5599 - val_accuracy: 0.8403 - val_loss: 0.3981\nEpoch 3/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m202s\u001b[0m 2s/step - accuracy: 0.8307 - loss: 0.4044 - val_accuracy: 0.8812 - val_loss: 0.3186\nEpoch 4/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m202s\u001b[0m 2s/step - accuracy: 0.8675 - loss: 0.3436 - val_accuracy: 0.8908 - val_loss: 0.2868\nEpoch 5/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m189s\u001b[0m 2s/step - accuracy: 0.8875 - loss: 0.2981 - val_accuracy: 0.8992 - val_loss: 0.2619\nEpoch 6/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m203s\u001b[0m 2s/step - accuracy: 0.8951 - loss: 0.2841 - val_accuracy: 0.9076 - val_loss: 0.2464\nEpoch 7/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m190s\u001b[0m 2s/step - accuracy: 0.9079 - loss: 0.2535 - val_accuracy: 0.9088 - val_loss: 0.2372\nEpoch 8/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m202s\u001b[0m 2s/step - accuracy: 0.9141 - loss: 0.2440 - val_accuracy: 0.9088 - val_loss: 0.2313\nEpoch 9/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m190s\u001b[0m 2s/step - accuracy: 0.9183 - loss: 0.2172 - val_accuracy: 0.9052 - val_loss: 0.2232\nEpoch 10/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m201s\u001b[0m 2s/step - accuracy: 0.9244 - loss: 0.2042 - val_accuracy: 0.9088 - val_loss: 0.2136\nEpoch 11/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m204s\u001b[0m 2s/step - accuracy: 0.9316 - loss: 0.1764 - val_accuracy: 0.9112 - val_loss: 0.2092\nEpoch 12/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m190s\u001b[0m 2s/step - accuracy: 0.9436 - loss: 0.1674 - val_accuracy: 0.9112 - val_loss: 0.2025\nEpoch 13/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m202s\u001b[0m 2s/step - accuracy: 0.9323 - loss: 0.1879 - val_accuracy: 0.9196 - val_loss: 0.1953\nEpoch 14/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m201s\u001b[0m 2s/step - accuracy: 0.9359 - loss: 0.1775 - val_accuracy: 0.9016 - val_loss: 0.2202\nEpoch 15/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m202s\u001b[0m 2s/step - accuracy: 0.9472 - loss: 0.1429 - val_accuracy: 0.9016 - val_loss: 0.2162\nEpoch 16/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m189s\u001b[0m 2s/step - accuracy: 0.9540 - loss: 0.1333 - val_accuracy: 0.9208 - val_loss: 0.1759\nEpoch 17/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m202s\u001b[0m 2s/step - accuracy: 0.9581 - loss: 0.1266 - val_accuracy: 0.9136 - val_loss: 0.1933\nEpoch 18/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m201s\u001b[0m 2s/step - accuracy: 0.9646 - loss: 0.1053 - val_accuracy: 0.9256 - val_loss: 0.1748\nEpoch 19/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m188s\u001b[0m 2s/step - accuracy: 0.9718 - loss: 0.0944 - val_accuracy: 0.9136 - val_loss: 0.1841\nEpoch 20/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m201s\u001b[0m 2s/step - accuracy: 0.9731 - loss: 0.0852 - val_accuracy: 0.9280 - val_loss: 0.1667\nEpoch 21/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m188s\u001b[0m 2s/step - accuracy: 0.9751 - loss: 0.0793 - val_accuracy: 0.9088 - val_loss: 0.2121\nEpoch 22/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m201s\u001b[0m 2s/step - accuracy: 0.9762 - loss: 0.0755 - val_accuracy: 0.9244 - val_loss: 0.1728\nEpoch 23/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m187s\u001b[0m 2s/step - accuracy: 0.9779 - loss: 0.0708 - val_accuracy: 0.9256 - val_loss: 0.1761\nEpoch 24/24\n\u001b[1m105/105\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m202s\u001b[0m 2s/step - accuracy: 0.9812 - loss: 0.0640 - val_accuracy: 0.9280 - val_loss: 0.1666\n","output_type":"stream"}]},{"cell_type":"code","source":"model.save('model_casia_run1.h5')","metadata":{"execution":{"iopub.status.busy":"2024-03-04T10:16:43.441911Z","iopub.execute_input":"2024-03-04T10:16:43.442371Z","iopub.status.idle":"2024-03-04T10:16:44.072762Z","shell.execute_reply.started":"2024-03-04T10:16:43.442330Z","shell.execute_reply":"2024-03-04T10:16:44.071618Z"},"trusted":true},"execution_count":17,"outputs":[]}]}