{
"cells": [
{
"cell_type": "code",
"execution_count": 1,
"metadata": {
"id": "h3_RFpXiPQR6"
},
"outputs": [],
"source": [
"from tensorflow.keras import Model\n",
"from tensorflow.keras import Input\n",
"from tensorflow.keras.layers import Conv2D, ReLU, ELU, LeakyReLU, Dropout, Dense, MaxPooling2D, Flatten, BatchNormalization\n",
"from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
"from tensorflow.keras.optimizers import Adam\n",
"from tensorflow.keras.callbacks import ModelCheckpoint, TensorBoard, EarlyStopping\n",
"from tensorflow.keras.utils import plot_model\n",
"from tensorflow.keras.models import load_model\n",
"from sklearn.metrics import classification_report\n",
"\n",
"import numpy as np\n",
"import matplotlib.pyplot as plt\n"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {
"id": "ADU3Hu_TAFvG"
},
"outputs": [],
"source": [
"IMG_WIDTH = 256"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {
"id": "NyTzubIUjeCR"
},
"outputs": [],
"source": [
"def get_datagen(use_default_augmentation=True, **kwargs):\n",
" kwargs.update({'rescale': 1./255})\n",
" if use_default_augmentation:\n",
" kwargs.update({\n",
" 'rotation_range': 15,\n",
" 'zoom_range': 0.2,\n",
" 'brightness_range': (0.8, 1.2),\n",
" 'channel_shift_range': 30,\n",
" 'horizontal_flip': True,\n",
" })\n",
" return ImageDataGenerator(**kwargs)"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {
"id": "nrH7Fz6EM4mk"
},
"outputs": [],
"source": [
"def get_train_data_generator(\n",
" train_data_dir, \n",
" batch_size, \n",
" validation_split=None, \n",
" use_default_augmentation=True,\n",
" augmentations=None\n",
"):\n",
" if not augmentations:\n",
" augmentations = {}\n",
"\n",
" train_datagen = get_datagen(\n",
" use_default_augmentation=use_default_augmentation,\n",
" validation_split=validation_split if validation_split else 0.0,\n",
" **augmentations\n",
" )\n",
" \n",
" train_generator = train_datagen.flow_from_directory(\n",
" directory=train_data_dir,\n",
" target_size=(IMG_WIDTH, IMG_WIDTH),\n",
" batch_size=batch_size,\n",
" class_mode='binary',\n",
" subset='training',\n",
" )\n",
"\n",
" validation_generator = None\n",
"\n",
" if validation_split:\n",
" validation_generator = train_datagen.flow_from_directory(\n",
" directory=train_data_dir,\n",
" target_size=(IMG_WIDTH, IMG_WIDTH),\n",
" batch_size=batch_size,\n",
" class_mode='binary',\n",
" subset='validation'\n",
" )\n",
"\n",
" return train_generator, validation_generator"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {
"id": "6G7tVf0wNHvd"
},
"outputs": [],
"source": [
"def get_test_data_generator(test_data_dir, batch_size, shuffle=False):\n",
" test_datagen = get_datagen(use_default_augmentation=False)\n",
" return test_datagen.flow_from_directory(\n",
" directory=test_data_dir,\n",
" target_size=(IMG_WIDTH, IMG_WIDTH),\n",
" batch_size=batch_size,\n",
" class_mode='binary',\n",
" shuffle=shuffle\n",
" )"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {
"id": "kSWlbqMv-TK4"
},
"outputs": [],
"source": [
"def activation_layer(ip, activation, *args):\n",
" return {'relu': ReLU(*args)(ip),\n",
" 'elu': ELU(*args)(ip),\n",
" 'lrelu': LeakyReLU(*args)(ip)}[activation]"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {
"id": "CvF1f4Y28oPM"
},
"outputs": [],
"source": [
"def conv2D(ip,\n",
" filters,\n",
" kernel_size,\n",
" activation,\n",
" padding='same',\n",
" pool_size=(2, 2)):\n",
" layer = Conv2D(filters,\n",
" kernel_size=kernel_size,\n",
" padding=padding)(ip)\n",
"\n",
" layer = activation_layer(layer, activation=activation)\n",
"\n",
" layer = BatchNormalization()(layer)\n",
"\n",
" return MaxPooling2D(pool_size=pool_size, padding=padding)(layer)"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {
"id": "d-4--jRd-bz1"
},
"outputs": [],
"source": [
"def fully_connected_layer(ip,\n",
" hidden_activation,\n",
" dropout):\n",
" layer = Dense(16)(ip)\n",
" layer = activation_layer(layer, hidden_activation, *[0.1,])\n",
" return Dropout(rate=dropout)(layer)"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {
"id": "1Cp48aFy_k4G"
},
"outputs": [],
"source": [
"def build_model(ip=Input(shape=(IMG_WIDTH, IMG_WIDTH, 3)),\n",
" activation='relu',\n",
" dropout=0.5,\n",
" hidden_activation='lrelu'):\n",
" \n",
" layer = conv2D(ip, filters=8, kernel_size=(3, 3), activation=activation)\n",
"\n",
" layer = conv2D(layer, filters=8, kernel_size=(5, 5), activation=activation)\n",
"\n",
" layer = conv2D(layer, filters=16, kernel_size=(5, 5), activation=activation)\n",
"\n",
" layer = conv2D(layer, filters=16, kernel_size=(5, 5), activation=activation, pool_size=(4, 4))\n",
"\n",
" layer = Flatten()(layer)\n",
" layer = Dropout(rate=dropout)(layer)\n",
"\n",
" layer = fully_connected_layer(layer, hidden_activation=hidden_activation, dropout=dropout)\n",
"\n",
" op_layer = Dense(1, activation='sigmoid')(layer)\n",
"\n",
" model = Model(ip, op_layer)\n",
"\n",
" return model"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {
"id": "ZdoMu0LbDGMC"
},
"outputs": [],
"source": [
"def evaluate_model(model, test_data_dir, batch_size):\n",
" data = get_test_data_generator(test_data_dir, batch_size)\n",
" return model.evaluate(data)\n",
"\n",
"\n",
"def predict(model, data, steps=None, threshold=0.5):\n",
" predictions = model.predict(data, steps=steps, verbose=1)\n",
" return predictions, np.where(predictions >= threshold, 1, 0)\n",
"\n",
"\n",
"def save_model_history(history, filename):\n",
" with open(filename, 'wb') as f:\n",
" pickle.dump(history.history, f)"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {
"id": "fuXsZWxke_ic"
},
"outputs": [],
"source": [
"def get_activation_model(model, conv_idx):\n",
" conv_layers = [layer for layer in model.layers if 'conv' in layer.name]\n",
" selected_layers = [layer for index, layer in enumerate(conv_layers) if index in conv_idx]\n",
" activation_model = Model(\n",
" inputs=model.inputs,\n",
" outputs=[layer.output for layer in selected_layers]\n",
" )\n",
" return activation_model"
]
},
{
"cell_type": "code",
"execution_count": 15,
"metadata": {
"id": "W0Sda_34HCzQ"
},
"outputs": [],
"source": [
"def train_model(model,\n",
" train_data_dir,\n",
" validation_split=None,\n",
" batch_size=32,\n",
" use_default_augmentation=True,\n",
" augmentations=None,\n",
" epochs=30,\n",
" lr=1e-3,\n",
" loss='binary_crossentropy',\n",
" compile=True,\n",
" lr_decay=True,\n",
" decay_rate=0.10,\n",
" decay_limit=1e-6,\n",
" checkpoint=True,\n",
" stop_early=True,\n",
" monitor='val_accuracy',\n",
" mode='max',\n",
" patience=20,\n",
" tensorboard=True,\n",
" loss_curve=True):\n",
" \n",
"\n",
" train_generator, validation_generator = get_train_data_generator(\n",
" train_data_dir=train_data_dir,\n",
" batch_size=batch_size,\n",
" validation_split=validation_split,\n",
" use_default_augmentation=use_default_augmentation,\n",
" augmentations=augmentations\n",
" )\n",
"\n",
" callbacks = []\n",
" if checkpoint:\n",
" filepath = f'mesonet_trained.hdf5'\n",
" model_checkpoint = ModelCheckpoint(\n",
" filepath, monitor='val_accuracy', verbose=1,\n",
" save_best_only=True\n",
" )\n",
" callbacks.append(model_checkpoint)\n",
"\n",
" if stop_early:\n",
" callbacks.append(\n",
" EarlyStopping(\n",
" monitor=monitor,\n",
" mode=mode,\n",
" patience=patience,\n",
" verbose=1\n",
" )\n",
" )\n",
"\n",
"\n",
" history = model.fit(\n",
" train_generator,\n",
" epochs=epochs,\n",
" verbose=1,\n",
" callbacks=callbacks,\n",
" validation_data=validation_generator,\n",
" steps_per_epoch=train_generator.samples // batch_size,\n",
" validation_steps=validation_generator.samples // batch_size if validation_generator else None,\n",
" )\n",
"\n",
" return history"
]
},
{
"cell_type": "code",
"execution_count": 14,
"metadata": {
"id": "aZtXFYPDoLZp"
},
"outputs": [
{
"ename": "OSError",
"evalue": "Unable to open file (file signature not found)",
"output_type": "error",
"traceback": [
"\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
"\u001b[0;31mOSError\u001b[0m Traceback (most recent call last)",
"Input \u001b[0;32mIn [14]\u001b[0m, in \u001b[0;36m<cell line: 1>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0m model_exp \u001b[38;5;241m=\u001b[39m \u001b[43mload_model\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43m/Users/jarvis/pymycod/Deepfakes/Meso_4.ipynb\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m)\u001b[49m\n",
"File \u001b[0;32m/opt/anaconda3/envs/tensor/lib/python3.8/site-packages/keras/utils/traceback_utils.py:67\u001b[0m, in \u001b[0;36mfilter_traceback.<locals>.error_handler\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m 65\u001b[0m \u001b[38;5;28;01mexcept\u001b[39;00m \u001b[38;5;167;01mException\u001b[39;00m \u001b[38;5;28;01mas\u001b[39;00m e: \u001b[38;5;66;03m# pylint: disable=broad-except\u001b[39;00m\n\u001b[1;32m 66\u001b[0m filtered_tb \u001b[38;5;241m=\u001b[39m _process_traceback_frames(e\u001b[38;5;241m.\u001b[39m__traceback__)\n\u001b[0;32m---> 67\u001b[0m \u001b[38;5;28;01mraise\u001b[39;00m e\u001b[38;5;241m.\u001b[39mwith_traceback(filtered_tb) \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m\n\u001b[1;32m 68\u001b[0m \u001b[38;5;28;01mfinally\u001b[39;00m:\n\u001b[1;32m 69\u001b[0m \u001b[38;5;28;01mdel\u001b[39;00m filtered_tb\n",
"File \u001b[0;32m/opt/anaconda3/envs/tensor/lib/python3.8/site-packages/h5py/_hl/files.py:507\u001b[0m, in \u001b[0;36mFile.__init__\u001b[0;34m(self, name, mode, driver, libver, userblock_size, swmr, rdcc_nslots, rdcc_nbytes, rdcc_w0, track_order, fs_strategy, fs_persist, fs_threshold, fs_page_size, page_buf_size, min_meta_keep, min_raw_keep, locking, **kwds)\u001b[0m\n\u001b[1;32m 502\u001b[0m fapl \u001b[38;5;241m=\u001b[39m make_fapl(driver, libver, rdcc_nslots, rdcc_nbytes, rdcc_w0,\n\u001b[1;32m 503\u001b[0m locking, page_buf_size, min_meta_keep, min_raw_keep, \u001b[38;5;241m*\u001b[39m\u001b[38;5;241m*\u001b[39mkwds)\n\u001b[1;32m 504\u001b[0m fcpl \u001b[38;5;241m=\u001b[39m make_fcpl(track_order\u001b[38;5;241m=\u001b[39mtrack_order, fs_strategy\u001b[38;5;241m=\u001b[39mfs_strategy,\n\u001b[1;32m 505\u001b[0m fs_persist\u001b[38;5;241m=\u001b[39mfs_persist, fs_threshold\u001b[38;5;241m=\u001b[39mfs_threshold,\n\u001b[1;32m 506\u001b[0m fs_page_size\u001b[38;5;241m=\u001b[39mfs_page_size)\n\u001b[0;32m--> 507\u001b[0m fid \u001b[38;5;241m=\u001b[39m make_fid(name, mode, userblock_size, fapl, fcpl, swmr\u001b[38;5;241m=\u001b[39mswmr)\n\u001b[1;32m 509\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28misinstance\u001b[39m(libver, \u001b[38;5;28mtuple\u001b[39m):\n\u001b[1;32m 510\u001b[0m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_libver \u001b[38;5;241m=\u001b[39m libver\n",
"File \u001b[0;32m/opt/anaconda3/envs/tensor/lib/python3.8/site-packages/h5py/_hl/files.py:220\u001b[0m, in \u001b[0;36mmake_fid\u001b[0;34m(name, mode, userblock_size, fapl, fcpl, swmr)\u001b[0m\n\u001b[1;32m 218\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m swmr \u001b[38;5;129;01mand\u001b[39;00m swmr_support:\n\u001b[1;32m 219\u001b[0m flags \u001b[38;5;241m|\u001b[39m\u001b[38;5;241m=\u001b[39m h5f\u001b[38;5;241m.\u001b[39mACC_SWMR_READ\n\u001b[0;32m--> 220\u001b[0m fid \u001b[38;5;241m=\u001b[39m \u001b[43mh5f\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mopen\u001b[49m\u001b[43m(\u001b[49m\u001b[43mname\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mflags\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mfapl\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mfapl\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m 221\u001b[0m \u001b[38;5;28;01melif\u001b[39;00m mode \u001b[38;5;241m==\u001b[39m \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mr+\u001b[39m\u001b[38;5;124m'\u001b[39m:\n\u001b[1;32m 222\u001b[0m fid \u001b[38;5;241m=\u001b[39m h5f\u001b[38;5;241m.\u001b[39mopen(name, h5f\u001b[38;5;241m.\u001b[39mACC_RDWR, fapl\u001b[38;5;241m=\u001b[39mfapl)\n",
"File \u001b[0;32mh5py/_objects.pyx:54\u001b[0m, in \u001b[0;36mh5py._objects.with_phil.wrapper\u001b[0;34m()\u001b[0m\n",
"File \u001b[0;32mh5py/_objects.pyx:55\u001b[0m, in \u001b[0;36mh5py._objects.with_phil.wrapper\u001b[0;34m()\u001b[0m\n",
"File \u001b[0;32mh5py/h5f.pyx:106\u001b[0m, in \u001b[0;36mh5py.h5f.open\u001b[0;34m()\u001b[0m\n",
"\u001b[0;31mOSError\u001b[0m: Unable to open file (file signature not found)"
]
}
],
"source": [
"model_exp = load_model('/Users/jarvis/pymycod/Deepfakes/Meso_4.ipynb')"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "IC4V7HflFZC2",
"outputId": "c86ee986-baeb-449c-b024-d6ad7ecdaf2a"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Found 1945 images belonging to 2 classes.\n",
"31/31 [==============================] - 7s 235ms/step - loss: 0.0998 - accuracy: 0.9625\n"
]
},
{
"data": {
"text/plain": [
"[0.09982584416866302, 0.9624678492546082]"
]
},
"execution_count": 124,
"metadata": {
"tags": []
},
"output_type": "execute_result"
}
],
"source": [
"evaluate_model(model_exp, 'data/test', 64)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "fG6V0lOzFeg2",
"outputId": "db1fec59-cd05-440f-9ae6-acaea21db0ea"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Found 1945 images belonging to 2 classes.\n",
"31/31 [==============================] - 7s 235ms/step\n",
" precision recall f1-score support\n",
"\n",
" 0 0.96 0.94 0.95 773\n",
" 1 0.96 0.97 0.97 1172\n",
"\n",
" accuracy 0.96 1945\n",
" macro avg 0.96 0.96 0.96 1945\n",
"weighted avg 0.96 0.96 0.96 1945\n",
"\n"
]
}
],
"source": [
"print(get_classification_report(model_exp, 'data/test'))"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "cjHCL4fiF29I"
},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "_z00HkaPG76d",
"outputId": "f3d07ca5-29dd-4191-ef04-e6ef04bf592a"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Found 1945 images belonging to 2 classes.\n",
"31/31 [==============================] - 8s 241ms/step - loss: 0.2321 - accuracy: 0.9080\n"
]
},
{
"data": {
"text/plain": [
"[0.23209184408187866, 0.9079691767692566]"
]
},
"execution_count": 129,
"metadata": {
"tags": []
},
"output_type": "execute_result"
}
],
"source": [
"evaluate_model(model_exp, 'data/test', 64)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "v3ZQN53bHBEf",
"outputId": "14bd2bd7-0c83-4e52-ee12-054a4e52ca9f"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Found 1945 images belonging to 2 classes.\n",
"31/31 [==============================] - 7s 234ms/step\n",
" precision recall f1-score support\n",
"\n",
" 0 0.90 0.87 0.88 773\n",
" 1 0.91 0.93 0.92 1172\n",
"\n",
" accuracy 0.91 1945\n",
" macro avg 0.91 0.90 0.90 1945\n",
"weighted avg 0.91 0.91 0.91 1945\n",
"\n"
]
}
],
"source": [
"print(get_classification_report(model_exp, 'data/test'))"
]
}
],
"metadata": {
"accelerator": "GPU",
"colab": {
"collapsed_sections": [],
"name": "Meso-4.ipynb",
"provenance": []
},
"kernelspec": {
"display_name": "Python 3",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.8.13"
}
},
"nbformat": 4,
"nbformat_minor": 0
}