{"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":"nvidiaTeslaT4","dataSources":[{"sourceId":16880,"databundleVersionId":858837,"sourceType":"competition"},{"sourceId":924245,"sourceType":"datasetVersion","datasetId":464091}],"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"## Importing Required libraries","metadata":{"id":"dFXIv9qNpKzt","tags":[]}},{"cell_type":"code","source":"!pip install -U --upgrade tensorflow","metadata":{"execution":{"iopub.status.busy":"2024-05-02T02:26:04.065908Z","iopub.execute_input":"2024-05-02T02:26:04.066218Z","iopub.status.idle":"2024-05-02T02:27:13.552845Z","shell.execute_reply.started":"2024-05-02T02:26:04.066182Z","shell.execute_reply":"2024-05-02T02:27:13.551785Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import sys\nimport sklearn\nimport tensorflow as tf\n\nimport cv2\nimport pandas as pd\nimport numpy as np\n\nimport plotly.graph_objs as go\nfrom plotly.offline import iplot\nfrom matplotlib import pyplot as plt","metadata":{"id":"TFSU3FCOpKzu","execution":{"iopub.status.busy":"2024-05-02T02:27:13.554853Z","iopub.execute_input":"2024-05-02T02:27:13.555217Z","iopub.status.idle":"2024-05-02T02:27:21.447822Z","shell.execute_reply.started":"2024-05-02T02:27:13.555182Z","shell.execute_reply":"2024-05-02T02:27:21.447025Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"tf.test.is_gpu_available()","metadata":{"execution":{"iopub.status.busy":"2024-05-02T02:27:21.44883Z","iopub.execute_input":"2024-05-02T02:27:21.449328Z","iopub.status.idle":"2024-05-02T02:27:21.902607Z","shell.execute_reply.started":"2024-05-02T02:27:21.449302Z","shell.execute_reply":"2024-05-02T02:27:21.901112Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"tf.__version__","metadata":{"execution":{"iopub.status.busy":"2024-05-02T02:27:21.90461Z","iopub.execute_input":"2024-05-02T02:27:21.90491Z","iopub.status.idle":"2024-05-02T02:27:21.917524Z","shell.execute_reply.started":"2024-05-02T02:27:21.904884Z","shell.execute_reply":"2024-05-02T02:27:21.916679Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import matplotlib.pyplot as plt\n\nplt.rc('font', size=14)\nplt.rc('axes', labelsize=14, titlesize=14)\nplt.rc('legend', fontsize=14)\nplt.rc('xtick', labelsize=10)\nplt.rc('ytick', labelsize=10)","metadata":{"id":"8d4TH3NbpKzx","execution":{"iopub.status.busy":"2024-05-02T02:27:21.918549Z","iopub.execute_input":"2024-05-02T02:27:21.918835Z","iopub.status.idle":"2024-05-02T02:27:21.927931Z","shell.execute_reply.started":"2024-05-02T02:27:21.918803Z","shell.execute_reply":"2024-05-02T02:27:21.92717Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Data Visualisation","metadata":{"id":"NL3Ht4wC9b3n"}},{"cell_type":"code","source":"import os\n\ndef get_data():\n    return pd.read_csv('../input/deepfake-faces/metadata.csv')","metadata":{"id":"jfv9PxSB4tM8","execution":{"iopub.status.busy":"2024-05-02T02:27:21.928964Z","iopub.execute_input":"2024-05-02T02:27:21.929234Z","iopub.status.idle":"2024-05-02T02:27:21.938291Z","shell.execute_reply.started":"2024-05-02T02:27:21.929212Z","shell.execute_reply":"2024-05-02T02:27:21.937438Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"meta=get_data()\nmeta.head()","metadata":{"id":"tDW7BRph9ehF","outputId":"97de18b5-0a37-4302-8804-8a16a7d2ed2f","execution":{"iopub.status.busy":"2024-05-02T02:27:21.939385Z","iopub.execute_input":"2024-05-02T02:27:21.939917Z","iopub.status.idle":"2024-05-02T02:27:22.14885Z","shell.execute_reply.started":"2024-05-02T02:27:21.939888Z","shell.execute_reply":"2024-05-02T02:27:22.147886Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"meta.shape","metadata":{"id":"n7FSdDifbZxn","outputId":"5451a127-405a-4c0b-a197-c920b796adbb","execution":{"iopub.status.busy":"2024-05-02T02:27:22.149966Z","iopub.execute_input":"2024-05-02T02:27:22.150277Z","iopub.status.idle":"2024-05-02T02:27:22.155734Z","shell.execute_reply.started":"2024-05-02T02:27:22.150252Z","shell.execute_reply":"2024-05-02T02:27:22.154872Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"len(meta[meta.label=='FAKE']),len(meta[meta.label=='REAL'])","metadata":{"id":"_FJcz2IthxVG","outputId":"274c3f65-7acb-4f99-8aa9-a5b2a23bf06a","execution":{"iopub.status.busy":"2024-05-02T02:27:22.156821Z","iopub.execute_input":"2024-05-02T02:27:22.157072Z","iopub.status.idle":"2024-05-02T02:27:22.20722Z","shell.execute_reply.started":"2024-05-02T02:27:22.15705Z","shell.execute_reply":"2024-05-02T02:27:22.206403Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"real_df = meta[meta[\"label\"] == \"REAL\"]\nfake_df = meta[meta[\"label\"] == \"FAKE\"]\nsample_size = 8000\n\nreal_df = real_df.sample(sample_size, random_state=42)\nfake_df = fake_df.sample(sample_size, random_state=42)\n\nsample_meta = pd.concat([real_df, fake_df])","metadata":{"id":"IgMfzY-PjjtH","execution":{"iopub.status.busy":"2024-05-02T02:27:22.210499Z","iopub.execute_input":"2024-05-02T02:27:22.210973Z","iopub.status.idle":"2024-05-02T02:27:22.261635Z","shell.execute_reply.started":"2024-05-02T02:27:22.21095Z","shell.execute_reply":"2024-05-02T02:27:22.260961Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from sklearn.model_selection import train_test_split\n\nTrain_set, Test_set = train_test_split(sample_meta,test_size=0.2,random_state=42,stratify=sample_meta['label'])\nTrain_set, Val_set  = train_test_split(Train_set,test_size=0.3,random_state=42,stratify=Train_set['label'])","metadata":{"id":"5eB86S6K-T5Z","execution":{"iopub.status.busy":"2024-05-02T02:27:22.262477Z","iopub.execute_input":"2024-05-02T02:27:22.262755Z","iopub.status.idle":"2024-05-02T02:27:22.434526Z","shell.execute_reply.started":"2024-05-02T02:27:22.262721Z","shell.execute_reply":"2024-05-02T02:27:22.433556Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"Train_set.shape,Val_set.shape,Test_set.shape","metadata":{"id":"8p-TONijb4qA","outputId":"56d0b529-9d81-4019-d8fa-618c8cdba90f","execution":{"iopub.status.busy":"2024-05-02T02:27:22.435841Z","iopub.execute_input":"2024-05-02T02:27:22.436553Z","iopub.status.idle":"2024-05-02T02:27:22.443185Z","shell.execute_reply.started":"2024-05-02T02:27:22.436498Z","shell.execute_reply":"2024-05-02T02:27:22.442195Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"y = dict()\n\ny[0] = []\ny[1] = []\n\nfor set_name in (np.array(Train_set['label']), np.array(Val_set['label']), np.array(Test_set['label'])):\n    y[0].append(np.sum(set_name == 'REAL'))\n    y[1].append(np.sum(set_name == 'FAKE'))\n\ntrace0 = go.Bar(\n    x=['Train Set', 'Validation Set', 'Test Set'],\n    y=y[0],\n    name='REAL',\n    marker=dict(color='#33cc33'),\n    opacity=0.7\n)\ntrace1 = go.Bar(\n    x=['Train Set', 'Validation Set', 'Test Set'],\n    y=y[1],\n    name='FAKE',\n    marker=dict(color='#ff3300'),\n    opacity=0.7\n)\n\ndata = [trace0, trace1]\nlayout = go.Layout(\n    title='Count of classes in each set',\n    xaxis={'title': 'Set'},\n    yaxis={'title': 'Count'}\n)\n\nfig = go.Figure(data, layout)\niplot(fig)","metadata":{"id":"hzNGtCWd-mTk","outputId":"5178c3ed-cbba-4f99-99d0-26bde11a5dab","execution":{"iopub.status.busy":"2024-05-02T02:27:22.4445Z","iopub.execute_input":"2024-05-02T02:27:22.444858Z","iopub.status.idle":"2024-05-02T02:27:24.190891Z","shell.execute_reply.started":"2024-05-02T02:27:22.444824Z","shell.execute_reply":"2024-05-02T02:27:24.189976Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"plt.figure(figsize=(15,15))\nfor cur,i in enumerate(Train_set.index[25:50]):\n    plt.subplot(5,5,cur+1)\n    plt.xticks([])\n    plt.yticks([])\n    plt.grid(False)\n    \n    plt.imshow(cv2.imread('../input/deepfake-faces/faces_224/'+Train_set.loc[i,'videoname'][:-4]+'.jpg'))\n    \n    if(Train_set.loc[i,'label']=='FAKE'):\n        plt.xlabel('FAKE Image')\n    else:\n        plt.xlabel('REAL Image')\n        \nplt.show()","metadata":{"id":"VR7Uly2fcUYi","outputId":"c1f47a82-ef4f-4bcd-b51c-d4738142fc0f","execution":{"iopub.status.busy":"2024-05-02T02:27:24.192082Z","iopub.execute_input":"2024-05-02T02:27:24.192385Z","iopub.status.idle":"2024-05-02T02:27:26.338458Z","shell.execute_reply.started":"2024-05-02T02:27:24.19236Z","shell.execute_reply":"2024-05-02T02:27:26.337266Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Modelling","metadata":{"id":"dOvN_divkl-N"}},{"cell_type":"markdown","source":"### Custom CNN Architecture","metadata":{"id":"oid44Xx-pKz6"}},{"cell_type":"code","source":"def retreive_dataset(set_name):\n    images,labels=[],[]\n    for (img, imclass) in zip(set_name['videoname'], set_name['label']):\n        images.append(cv2.imread('../input/deepfake-faces/faces_224/'+img[:-4]+'.jpg'))\n        if(imclass=='FAKE'):\n            labels.append(1)\n        else:\n            labels.append(0)\n    \n    return np.array(images),np.array(labels)","metadata":{"id":"Hz0ZdQ_fgHhG","execution":{"iopub.status.busy":"2024-05-02T02:27:26.339643Z","iopub.execute_input":"2024-05-02T02:27:26.339954Z","iopub.status.idle":"2024-05-02T02:27:26.34671Z","shell.execute_reply.started":"2024-05-02T02:27:26.339926Z","shell.execute_reply":"2024-05-02T02:27:26.345875Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"X_train,y_train=retreive_dataset(Train_set)\nX_val,y_val=retreive_dataset(Val_set)\nX_test,y_test=retreive_dataset(Test_set)","metadata":{"id":"zeAGRcAbguKU","execution":{"iopub.status.busy":"2024-05-02T02:27:26.347959Z","iopub.execute_input":"2024-05-02T02:27:26.348566Z","iopub.status.idle":"2024-05-02T02:30:10.436866Z","shell.execute_reply.started":"2024-05-02T02:27:26.348536Z","shell.execute_reply":"2024-05-02T02:30:10.436032Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from functools import partial\n\ntf.random.set_seed(42) \nDefaultConv2D = partial(tf.keras.layers.Conv2D, kernel_size=3, padding=\"same\",\n                        activation=\"relu\", kernel_initializer=\"he_normal\")\n\nmodel = tf.keras.Sequential([\n    DefaultConv2D(filters=64, kernel_size=7, input_shape=[224, 224, 3]),\n    tf.keras.layers.MaxPool2D(),\n    DefaultConv2D(filters=128),\n    DefaultConv2D(filters=128),\n    tf.keras.layers.MaxPool2D(),\n    tf.keras.layers.Flatten(),\n    tf.keras.layers.Dense(units=128, activation=\"relu\",\n                          kernel_initializer=\"he_normal\"),\n    tf.keras.layers.Dropout(0.5),\n    tf.keras.layers.Dense(units=64, activation=\"relu\",\n                          kernel_initializer=\"he_normal\"),\n    tf.keras.layers.Dropout(0.5),\n    tf.keras.layers.Dense(units=1, activation=\"sigmoid\")\n])","metadata":{"id":"34upiak4pKz6","execution":{"iopub.status.busy":"2024-05-02T02:30:10.438251Z","iopub.execute_input":"2024-05-02T02:30:10.43887Z","iopub.status.idle":"2024-05-02T02:30:10.825624Z","shell.execute_reply.started":"2024-05-02T02:30:10.438834Z","shell.execute_reply":"2024-05-02T02:30:10.824662Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"model.compile(loss=\"binary_crossentropy\", optimizer=\"nadam\",\n              metrics=[\"accuracy\"])\nmodel.summary()","metadata":{"execution":{"iopub.status.busy":"2024-05-02T02:30:10.826784Z","iopub.execute_input":"2024-05-02T02:30:10.827056Z","iopub.status.idle":"2024-05-02T02:30:10.861104Z","shell.execute_reply.started":"2024-05-02T02:30:10.827031Z","shell.execute_reply":"2024-05-02T02:30:10.860246Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"history = model.fit(X_train, y_train, epochs=5,batch_size=64,\n                    validation_data=(X_val, y_val))","metadata":{"id":"KZbWeIBYpKz6","outputId":"deb6f56a-7b93-4241-a1bd-b210c0f2d426","execution":{"iopub.status.busy":"2024-05-02T02:30:10.862247Z","iopub.execute_input":"2024-05-02T02:30:10.862585Z","iopub.status.idle":"2024-05-02T02:34:02.196564Z","shell.execute_reply.started":"2024-05-02T02:30:10.862556Z","shell.execute_reply":"2024-05-02T02:34:02.195582Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"score = model.evaluate(X_test, y_test)","metadata":{"id":"6HDDr4uehast","execution":{"iopub.status.busy":"2024-05-02T02:34:02.198079Z","iopub.execute_input":"2024-05-02T02:34:02.203824Z","iopub.status.idle":"2024-05-02T02:34:23.94181Z","shell.execute_reply.started":"2024-05-02T02:34:02.203793Z","shell.execute_reply":"2024-05-02T02:34:23.940605Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# plot model performance\nacc = history.history['accuracy']\nval_acc = history.history['val_accuracy']\nloss = history.history['loss']\nval_loss = history.history['val_loss']\nepochs_range = range(1, len(history.epoch) + 1)\n\nplt.figure(figsize=(15,5))\n\nplt.subplot(1, 2, 1)\nplt.plot(epochs_range, acc, label='Train Set')\nplt.plot(epochs_range, val_acc, label='Val Set')\nplt.legend(loc=\"best\")\nplt.xlabel('Epochs')\nplt.ylabel('Accuracy')\nplt.title('Model Accuracy')\n\nplt.subplot(1, 2, 2)\nplt.plot(epochs_range, loss, label='Train Set')\nplt.plot(epochs_range, val_loss, label='Val Set')\nplt.legend(loc=\"best\")\nplt.xlabel('Epochs')\nplt.ylabel('Loss')\nplt.title('Model Loss')\n\nplt.tight_layout()\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2024-05-02T02:34:23.943627Z","iopub.execute_input":"2024-05-02T02:34:23.944032Z","iopub.status.idle":"2024-05-02T02:34:24.539045Z","shell.execute_reply.started":"2024-05-02T02:34:23.943996Z","shell.execute_reply":"2024-05-02T02:34:24.538149Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Pretrained Models for Transfer Learning","metadata":{"id":"hqxnSBJ3pKz8"}},{"cell_type":"code","source":"train_set_raw=tf.data.Dataset.from_tensor_slices((X_train,y_train))\nvalid_set_raw=tf.data.Dataset.from_tensor_slices((X_val,y_val))\ntest_set_raw=tf.data.Dataset.from_tensor_slices((X_test,y_test))","metadata":{"execution":{"iopub.status.busy":"2024-05-02T02:34:24.540443Z","iopub.execute_input":"2024-05-02T02:34:24.540728Z","iopub.status.idle":"2024-05-02T02:34:29.816055Z","shell.execute_reply.started":"2024-05-02T02:34:24.540703Z","shell.execute_reply":"2024-05-02T02:34:29.815081Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"tf.keras.backend.clear_session()  # extra code – resets layer name counter\n\nbatch_size = 32\npreprocess = tf.keras.applications.xception.preprocess_input\ntrain_set = train_set_raw.map(lambda X, y: (preprocess(tf.cast(X, tf.float32)), y))\ntrain_set = train_set.shuffle(1000, seed=42).batch(batch_size).prefetch(1)\nvalid_set = valid_set_raw.map(lambda X, y: (preprocess(tf.cast(X, tf.float32)), y)).batch(batch_size)\ntest_set = test_set_raw.map(lambda X, y: (preprocess(tf.cast(X, tf.float32)), y)).batch(batch_size)","metadata":{"id":"Bnz0n9XApKz9","execution":{"iopub.status.busy":"2024-05-02T02:34:29.817378Z","iopub.execute_input":"2024-05-02T02:34:29.817703Z","iopub.status.idle":"2024-05-02T02:34:31.541891Z","shell.execute_reply.started":"2024-05-02T02:34:29.817674Z","shell.execute_reply":"2024-05-02T02:34:31.541119Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# extra code – displays the first 9 images in the first batch of valid_set\n\nplt.figure(figsize=(12, 12))\nfor X_batch, y_batch in valid_set.take(1):\n    for index in range(9):\n        plt.subplot(3, 3, index + 1)\n        plt.imshow((X_batch[index] + 1) / 2)  # rescale to 0–1 for imshow()\n        if(y_batch[index]==1):\n            classt='FAKE'\n        else:\n            classt='REAL'\n        plt.title(f\"Class: {classt}\")\n        plt.axis(\"off\")\n\nplt.show()","metadata":{"id":"ZL3c3i4opKz9","outputId":"38847d8d-8822-41a3-cfb2-27479aa5debe","execution":{"iopub.status.busy":"2024-05-02T02:34:31.54292Z","iopub.execute_input":"2024-05-02T02:34:31.543191Z","iopub.status.idle":"2024-05-02T02:34:33.165646Z","shell.execute_reply.started":"2024-05-02T02:34:31.543167Z","shell.execute_reply":"2024-05-02T02:34:33.164795Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"data_augmentation = tf.keras.Sequential([\n    tf.keras.layers.RandomFlip(mode=\"horizontal\", seed=42),\n    tf.keras.layers.RandomRotation(factor=0.05, seed=42),\n    tf.keras.layers.RandomContrast(factor=0.2, seed=42)\n])","metadata":{"id":"Ib0cA8Y1pKz9","execution":{"iopub.status.busy":"2024-05-02T02:34:33.166706Z","iopub.execute_input":"2024-05-02T02:34:33.166959Z","iopub.status.idle":"2024-05-02T02:34:33.182542Z","shell.execute_reply.started":"2024-05-02T02:34:33.166936Z","shell.execute_reply":"2024-05-02T02:34:33.181538Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# extra code – displays the same first 9 images, after augmentation\n\nplt.figure(figsize=(12, 12))\nfor X_batch, y_batch in valid_set.take(1):\n    X_batch_augmented = data_augmentation(X_batch, training=True)\n    for index in range(9):\n        plt.subplot(3, 3, index + 1)\n        # We must rescale the images to the 0-1 range for imshow(), and also\n        # clip the result to that range, because data augmentation may\n        # make some values go out of bounds (e.g., RandomContrast in this case).\n        plt.imshow(np.clip((X_batch_augmented[index] + 1) / 2, 0, 1))\n        if(y_batch[index]==1):\n            classt='FAKE'\n        else:\n            classt='REAL'\n        plt.title(f\"Class: {classt}\")\n        plt.axis(\"off\")\n\nplt.show()","metadata":{"id":"w6GH5_vupKz-","outputId":"eeb2c924-2f4f-4aa1-bea9-951bebef4bf0","execution":{"iopub.status.busy":"2024-05-02T02:34:33.183685Z","iopub.execute_input":"2024-05-02T02:34:33.183949Z","iopub.status.idle":"2024-05-02T02:34:36.48032Z","shell.execute_reply.started":"2024-05-02T02:34:33.183926Z","shell.execute_reply":"2024-05-02T02:34:36.479361Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"tf.random.set_seed(42)  # extra code – ensures reproducibility\nbase_model = tf.keras.applications.xception.Xception(weights=\"imagenet\",\n                                                     include_top=False)\navg = tf.keras.layers.GlobalAveragePooling2D()(base_model.output)\noutput = tf.keras.layers.Dense(1, activation=\"sigmoid\")(avg)\nmodel = tf.keras.Model(inputs=base_model.input, outputs=output)","metadata":{"id":"lRyCgvaKpKz-","outputId":"a825e173-8b1d-4217-a1c4-5491b49c3e82","execution":{"iopub.status.busy":"2024-05-02T02:34:36.481551Z","iopub.execute_input":"2024-05-02T02:34:36.481882Z","iopub.status.idle":"2024-05-02T02:34:38.080029Z","shell.execute_reply.started":"2024-05-02T02:34:36.481853Z","shell.execute_reply":"2024-05-02T02:34:38.079274Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for layer in base_model.layers:\n    layer.trainable = False","metadata":{"id":"KBlyG6ElpKz-","execution":{"iopub.status.busy":"2024-05-02T02:34:38.086521Z","iopub.execute_input":"2024-05-02T02:34:38.086788Z","iopub.status.idle":"2024-05-02T02:34:38.09396Z","shell.execute_reply.started":"2024-05-02T02:34:38.086764Z","shell.execute_reply":"2024-05-02T02:34:38.093055Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"optimizer = tf.keras.optimizers.SGD(learning_rate=0.1, momentum=0.9)\nmodel.compile(loss=\"binary_crossentropy\", optimizer=optimizer,\n              metrics=[\"accuracy\"])\nhistory = model.fit(train_set, validation_data=valid_set, epochs=3)","metadata":{"id":"GGxK2yPcpKz-","outputId":"6b64214a-e104-4b6c-9b7a-3388fc9aa15f","execution":{"iopub.status.busy":"2024-05-02T02:34:38.09509Z","iopub.execute_input":"2024-05-02T02:34:38.095386Z","iopub.status.idle":"2024-05-02T02:37:14.637703Z","shell.execute_reply.started":"2024-05-02T02:34:38.095363Z","shell.execute_reply":"2024-05-02T02:37:14.636868Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for indices in zip(range(33), range(33, 66), range(66, 99), range(99, 132)):\n    for idx in indices:\n        print(f\"{idx:3}: {base_model.layers[idx].name:22}\", end=\"\")\n    print()","metadata":{"id":"GvGMiJMLpKz-","outputId":"91f2c96c-c058-45e0-e428-66fa6076ad56","execution":{"iopub.status.busy":"2024-05-02T02:37:14.63883Z","iopub.execute_input":"2024-05-02T02:37:14.639114Z","iopub.status.idle":"2024-05-02T02:37:14.66783Z","shell.execute_reply.started":"2024-05-02T02:37:14.639088Z","shell.execute_reply":"2024-05-02T02:37:14.666887Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"model.evaluate(test_set)","metadata":{"execution":{"iopub.status.busy":"2024-05-02T02:37:14.668838Z","iopub.execute_input":"2024-05-02T02:37:14.669075Z","iopub.status.idle":"2024-05-02T02:37:26.331271Z","shell.execute_reply.started":"2024-05-02T02:37:14.669051Z","shell.execute_reply":"2024-05-02T02:37:26.33043Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for layer in base_model.layers[56:]:\n    layer.trainable = True\n\noptimizer = tf.keras.optimizers.SGD(learning_rate=0.01, momentum=0.9)\nmodel.compile(loss=\"binary_crossentropy\", optimizer=optimizer,\n              metrics=[\"accuracy\"])\nhistory = model.fit(train_set, validation_data=valid_set, epochs=10)","metadata":{"id":"GEUNGlhvpKz_","outputId":"c622a91d-f634-4443-b87e-8d46defdb578","execution":{"iopub.status.busy":"2024-05-02T02:37:26.332332Z","iopub.execute_input":"2024-05-02T02:37:26.332612Z","iopub.status.idle":"2024-05-02T02:51:52.822384Z","shell.execute_reply.started":"2024-05-02T02:37:26.332588Z","shell.execute_reply":"2024-05-02T02:51:52.821447Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# plot model performance\nacc = history.history['accuracy']\nval_acc = history.history['val_accuracy']\nloss = history.history['loss']\nval_loss = history.history['val_loss']\nepochs_range = range(1, len(history.epoch) + 1)\n\nplt.figure(figsize=(15,5))\n\nplt.subplot(1, 2, 1)\nplt.plot(epochs_range, acc, label='Train Set')\nplt.plot(epochs_range, val_acc, label='Val Set')\nplt.legend(loc=\"best\")\nplt.xlabel('Epochs')\nplt.ylabel('Accuracy')\nplt.title('Model Accuracy')\n\nplt.subplot(1, 2, 2)\nplt.plot(epochs_range, loss, label='Train Set')\nplt.plot(epochs_range, val_loss, label='Val Set')\nplt.legend(loc=\"best\")\nplt.xlabel('Epochs')\nplt.ylabel('Loss')\nplt.title('Model Loss')\n\nplt.tight_layout()\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2024-05-02T02:51:52.841828Z","iopub.execute_input":"2024-05-02T02:51:52.842161Z","iopub.status.idle":"2024-05-02T02:51:53.437049Z","shell.execute_reply.started":"2024-05-02T02:51:52.842107Z","shell.execute_reply":"2024-05-02T02:51:53.436166Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"model.evaluate(test_set)","metadata":{"execution":{"iopub.status.busy":"2024-05-02T02:51:53.438254Z","iopub.execute_input":"2024-05-02T02:51:53.438539Z","iopub.status.idle":"2024-05-02T02:52:04.844025Z","shell.execute_reply.started":"2024-05-02T02:51:53.438515Z","shell.execute_reply":"2024-05-02T02:52:04.843165Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"model.save('xception_deepfake_image.h5')","metadata":{"execution":{"iopub.status.busy":"2024-05-02T02:52:04.845286Z","iopub.execute_input":"2024-05-02T02:52:04.845635Z","iopub.status.idle":"2024-05-02T02:52:05.323885Z","shell.execute_reply.started":"2024-05-02T02:52:04.845603Z","shell.execute_reply":"2024-05-02T02:52:05.323044Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Deep Fake Video Classification","metadata":{}},{"cell_type":"markdown","source":"## Importing required libraries","metadata":{}},{"cell_type":"code","source":"!pip install -U --upgrade tensorflow","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:15:55.166025Z","iopub.execute_input":"2024-05-02T04:15:55.166387Z","iopub.status.idle":"2024-05-02T04:17:07.516969Z","shell.execute_reply.started":"2024-05-02T04:15:55.166358Z","shell.execute_reply":"2024-05-02T04:17:07.515997Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# from tensorflow_docs.vis import embed\nfrom tensorflow import keras\n#from imutils import paths\n\nimport matplotlib.pyplot as plt\nimport tensorflow as tf\nimport pandas as pd\nimport numpy as np\nimport imageio\nimport cv2\nimport os","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:07.518914Z","iopub.execute_input":"2024-05-02T04:17:07.51933Z","iopub.status.idle":"2024-05-02T04:17:14.485907Z","shell.execute_reply.started":"2024-05-02T04:17:07.5193Z","shell.execute_reply":"2024-05-02T04:17:14.48455Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Data Visualisation","metadata":{}},{"cell_type":"code","source":"DATA_FOLDER = '../input/deepfake-detection-challenge'\nTRAIN_SAMPLE_FOLDER = 'train_sample_videos'\nTEST_FOLDER = 'test_videos'\n\nprint(f\"Train samples: {len(os.listdir(os.path.join(DATA_FOLDER, TRAIN_SAMPLE_FOLDER)))}\")\nprint(f\"Test samples: {len(os.listdir(os.path.join(DATA_FOLDER, TEST_FOLDER)))}\")","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:14.487272Z","iopub.execute_input":"2024-05-02T04:17:14.487957Z","iopub.status.idle":"2024-05-02T04:17:14.716135Z","shell.execute_reply.started":"2024-05-02T04:17:14.487924Z","shell.execute_reply":"2024-05-02T04:17:14.715068Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_sample_metadata = pd.read_json('../input/deepfake-detection-challenge/train_sample_videos/metadata.json').T\ntrain_sample_metadata.head()","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:14.718293Z","iopub.execute_input":"2024-05-02T04:17:14.718581Z","iopub.status.idle":"2024-05-02T04:17:14.846545Z","shell.execute_reply.started":"2024-05-02T04:17:14.718556Z","shell.execute_reply":"2024-05-02T04:17:14.845666Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_sample_metadata.groupby('label')['label'].count().plot(figsize=(15, 5), kind='bar', title='Distribution of Labels in the Training Set')\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:14.847788Z","iopub.execute_input":"2024-05-02T04:17:14.848158Z","iopub.status.idle":"2024-05-02T04:17:15.164822Z","shell.execute_reply.started":"2024-05-02T04:17:14.84812Z","shell.execute_reply":"2024-05-02T04:17:15.163979Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_sample_metadata.shape","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:15.165834Z","iopub.execute_input":"2024-05-02T04:17:15.166072Z","iopub.status.idle":"2024-05-02T04:17:15.171916Z","shell.execute_reply.started":"2024-05-02T04:17:15.16605Z","shell.execute_reply":"2024-05-02T04:17:15.170932Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Few fake videos","metadata":{}},{"cell_type":"code","source":"fake_train_sample_video = list(train_sample_metadata.loc[train_sample_metadata.label=='FAKE'].sample(3).index)\nfake_train_sample_video","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:15.173016Z","iopub.execute_input":"2024-05-02T04:17:15.173297Z","iopub.status.idle":"2024-05-02T04:17:15.185633Z","shell.execute_reply.started":"2024-05-02T04:17:15.173264Z","shell.execute_reply":"2024-05-02T04:17:15.184768Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def display_image_from_video(video_path):\n    '''\n    input: video_path - path for video\n    process:\n    1. perform a video capture from the video\n    2. read the image\n    3. display the image\n    '''\n    capture_image = cv2.VideoCapture(video_path) \n    ret, frame = capture_image.read()\n    fig = plt.figure(figsize=(10,10))\n    ax = fig.add_subplot(111)\n    frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)\n    ax.imshow(frame)","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:15.186842Z","iopub.execute_input":"2024-05-02T04:17:15.187177Z","iopub.status.idle":"2024-05-02T04:17:15.195438Z","shell.execute_reply.started":"2024-05-02T04:17:15.187146Z","shell.execute_reply":"2024-05-02T04:17:15.194603Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for video_file in fake_train_sample_video:\n    display_image_from_video(os.path.join(DATA_FOLDER, TRAIN_SAMPLE_FOLDER, video_file))","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:15.196952Z","iopub.execute_input":"2024-05-02T04:17:15.19733Z","iopub.status.idle":"2024-05-02T04:17:18.029833Z","shell.execute_reply.started":"2024-05-02T04:17:15.197301Z","shell.execute_reply":"2024-05-02T04:17:18.028854Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Few Real Videos","metadata":{}},{"cell_type":"code","source":"real_train_sample_video = list(train_sample_metadata.loc[train_sample_metadata.label=='REAL'].sample(3).index)\nreal_train_sample_video","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:18.033319Z","iopub.execute_input":"2024-05-02T04:17:18.03361Z","iopub.status.idle":"2024-05-02T04:17:18.042128Z","shell.execute_reply.started":"2024-05-02T04:17:18.033586Z","shell.execute_reply":"2024-05-02T04:17:18.041048Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for video_file in real_train_sample_video:\n    display_image_from_video(os.path.join(DATA_FOLDER, TRAIN_SAMPLE_FOLDER, video_file))","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:18.043121Z","iopub.execute_input":"2024-05-02T04:17:18.043389Z","iopub.status.idle":"2024-05-02T04:17:20.409151Z","shell.execute_reply.started":"2024-05-02T04:17:18.043366Z","shell.execute_reply":"2024-05-02T04:17:20.408224Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Videos with same original","metadata":{}},{"cell_type":"code","source":"train_sample_metadata['original'].value_counts()[0:5]","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:20.410262Z","iopub.execute_input":"2024-05-02T04:17:20.410545Z","iopub.status.idle":"2024-05-02T04:17:20.422913Z","shell.execute_reply.started":"2024-05-02T04:17:20.410521Z","shell.execute_reply":"2024-05-02T04:17:20.421844Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def display_image_from_video_list(video_path_list, video_folder=TRAIN_SAMPLE_FOLDER):\n    '''\n    input: video_path_list - path for video\n    process:\n    0. for each video in the video path list\n        1. perform a video capture from the video\n        2. read the image\n        3. display the image\n    '''\n    plt.figure()\n    fig, ax = plt.subplots(2,3,figsize=(16,8))\n    # we only show images extracted from the first 6 videos\n    for i, video_file in enumerate(video_path_list[0:6]):\n        video_path = os.path.join(DATA_FOLDER, video_folder,video_file)\n        capture_image = cv2.VideoCapture(video_path) \n        ret, frame = capture_image.read()\n        frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)\n        ax[i//3, i%3].imshow(frame)\n        ax[i//3, i%3].set_title(f\"Video: {video_file}\")\n        ax[i//3, i%3].axis('on')\n","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:20.42439Z","iopub.execute_input":"2024-05-02T04:17:20.425318Z","iopub.status.idle":"2024-05-02T04:17:20.434119Z","shell.execute_reply.started":"2024-05-02T04:17:20.425281Z","shell.execute_reply":"2024-05-02T04:17:20.433236Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"same_original_fake_train_sample_video = list(train_sample_metadata.loc[train_sample_metadata.original=='atvmxvwyns.mp4'].index)\ndisplay_image_from_video_list(same_original_fake_train_sample_video)","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:20.435236Z","iopub.execute_input":"2024-05-02T04:17:20.435566Z","iopub.status.idle":"2024-05-02T04:17:24.553188Z","shell.execute_reply.started":"2024-05-02T04:17:20.435535Z","shell.execute_reply":"2024-05-02T04:17:24.552234Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Test video files","metadata":{}},{"cell_type":"code","source":"test_videos = pd.DataFrame(list(os.listdir(os.path.join(DATA_FOLDER, TEST_FOLDER))), columns=['video'])","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:24.554663Z","iopub.execute_input":"2024-05-02T04:17:24.555038Z","iopub.status.idle":"2024-05-02T04:17:24.561211Z","shell.execute_reply.started":"2024-05-02T04:17:24.555007Z","shell.execute_reply":"2024-05-02T04:17:24.560181Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"test_videos.head()","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:24.562457Z","iopub.execute_input":"2024-05-02T04:17:24.562779Z","iopub.status.idle":"2024-05-02T04:17:24.576775Z","shell.execute_reply.started":"2024-05-02T04:17:24.562756Z","shell.execute_reply":"2024-05-02T04:17:24.575729Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"display_image_from_video(os.path.join(DATA_FOLDER, TEST_FOLDER, test_videos.iloc[2].video))","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:24.577729Z","iopub.execute_input":"2024-05-02T04:17:24.578015Z","iopub.status.idle":"2024-05-02T04:17:25.492875Z","shell.execute_reply.started":"2024-05-02T04:17:24.577991Z","shell.execute_reply":"2024-05-02T04:17:25.49192Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Play video files","metadata":{}},{"cell_type":"code","source":"fake_videos = list(train_sample_metadata.loc[train_sample_metadata.label=='FAKE'].index)","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:25.494157Z","iopub.execute_input":"2024-05-02T04:17:25.494541Z","iopub.status.idle":"2024-05-02T04:17:25.501164Z","shell.execute_reply.started":"2024-05-02T04:17:25.494507Z","shell.execute_reply":"2024-05-02T04:17:25.50012Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from IPython.display import HTML\nfrom base64 import b64encode\n\ndef play_video(video_file, subset=TRAIN_SAMPLE_FOLDER):\n    '''\n    Display video\n    param: video_file - the name of the video file to display\n    param: subset - the folder where the video file is located (can be TRAIN_SAMPLE_FOLDER or TEST_Folder)\n    '''\n    video_url = open(os.path.join(DATA_FOLDER, subset,video_file),'rb').read()\n    data_url = \"data:video/mp4;base64,\" + b64encode(video_url).decode()\n    return HTML(\"\"\"<video width=500 controls><source src=\"%s\" type=\"video/mp4\"></video>\"\"\" % data_url)\n\nplay_video(fake_videos[10])","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:25.502547Z","iopub.execute_input":"2024-05-02T04:17:25.503184Z","iopub.status.idle":"2024-05-02T04:17:25.892727Z","shell.execute_reply.started":"2024-05-02T04:17:25.503143Z","shell.execute_reply":"2024-05-02T04:17:25.891152Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Modelling","metadata":{}},{"cell_type":"markdown","source":"### A CNN-RNN Architecture","metadata":{}},{"cell_type":"code","source":"IMG_SIZE = 224\nBATCH_SIZE = 64\nEPOCHS = 10\n\nMAX_SEQ_LENGTH = 20\nNUM_FEATURES = 2048","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:25.894058Z","iopub.execute_input":"2024-05-02T04:17:25.894654Z","iopub.status.idle":"2024-05-02T04:17:25.902984Z","shell.execute_reply.started":"2024-05-02T04:17:25.894597Z","shell.execute_reply":"2024-05-02T04:17:25.899209Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def crop_center_square(frame):\n    y, x = frame.shape[0:2]\n    min_dim = min(y, x)\n    start_x = (x // 2) - (min_dim // 2)\n    start_y = (y // 2) - (min_dim // 2)\n    return frame[start_y : start_y + min_dim, start_x : start_x + min_dim]\n\n\ndef load_video(path, max_frames=0, resize=(IMG_SIZE, IMG_SIZE)):\n    cap = cv2.VideoCapture(path)\n    frames = []\n    try:\n        while True:\n            ret, frame = cap.read()\n            if not ret:\n                break\n            frame = crop_center_square(frame)\n            frame = cv2.resize(frame, resize)\n            frame = frame[:, :, [2, 1, 0]]\n            frames.append(frame)\n\n            if len(frames) == max_frames:\n                break\n    finally:\n        cap.release()\n    return np.array(frames)","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:25.904653Z","iopub.execute_input":"2024-05-02T04:17:25.904985Z","iopub.status.idle":"2024-05-02T04:17:25.917304Z","shell.execute_reply.started":"2024-05-02T04:17:25.90495Z","shell.execute_reply":"2024-05-02T04:17:25.916467Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def build_feature_extractor():\n    feature_extractor = keras.applications.InceptionV3(\n        weights=\"imagenet\",\n        include_top=False,\n        pooling=\"avg\",\n        input_shape=(IMG_SIZE, IMG_SIZE, 3),\n    )\n    preprocess_input = keras.applications.inception_v3.preprocess_input\n\n    inputs = keras.Input((IMG_SIZE, IMG_SIZE, 3))\n    preprocessed = preprocess_input(inputs)\n\n    outputs = feature_extractor(preprocessed)\n    return keras.Model(inputs, outputs, name=\"feature_extractor\")\n\n\nfeature_extractor = build_feature_extractor()","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:25.918342Z","iopub.execute_input":"2024-05-02T04:17:25.918614Z","iopub.status.idle":"2024-05-02T04:17:29.313492Z","shell.execute_reply.started":"2024-05-02T04:17:25.918592Z","shell.execute_reply":"2024-05-02T04:17:29.312712Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def prepare_all_videos(df, root_dir):\n    num_samples = len(df)\n    video_paths = list(df.index)\n    labels = df[\"label\"].values\n    labels = np.array(labels=='FAKE').astype(int)\n\n    # `frame_masks` and `frame_features` are what we will feed to our sequence model.\n    # `frame_masks` will contain a bunch of booleans denoting if a timestep is\n    # masked with padding or not.\n    frame_masks = np.zeros(shape=(num_samples, MAX_SEQ_LENGTH), dtype=\"bool\")\n    frame_features = np.zeros(\n        shape=(num_samples, MAX_SEQ_LENGTH, NUM_FEATURES), dtype=\"float32\"\n    )\n\n    # For each video.\n    for idx, path in enumerate(video_paths):\n        # Gather all its frames and add a batch dimension.\n        frames = load_video(os.path.join(root_dir, path))\n        frames = frames[None, ...]\n\n        # Initialize placeholders to store the masks and features of the current video.\n        temp_frame_mask = np.zeros(shape=(1, MAX_SEQ_LENGTH,), dtype=\"bool\")\n        temp_frame_features = np.zeros(\n            shape=(1, MAX_SEQ_LENGTH, NUM_FEATURES), dtype=\"float32\"\n        )\n\n        # Extract features from the frames of the current video.\n        for i, batch in enumerate(frames):\n            video_length = batch.shape[0]\n            length = min(MAX_SEQ_LENGTH, video_length)\n            for j in range(length):\n                temp_frame_features[i, j, :] = feature_extractor.predict(\n                    batch[None, j, :]\n                )\n            temp_frame_mask[i, :length] = 1  # 1 = not masked, 0 = masked\n\n        frame_features[idx,] = temp_frame_features.squeeze()\n        frame_masks[idx,] = temp_frame_mask.squeeze()\n\n    return (frame_features, frame_masks), labels","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:29.314595Z","iopub.execute_input":"2024-05-02T04:17:29.314906Z","iopub.status.idle":"2024-05-02T04:17:29.325063Z","shell.execute_reply.started":"2024-05-02T04:17:29.31488Z","shell.execute_reply":"2024-05-02T04:17:29.324063Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from sklearn.model_selection import train_test_split\n\nTrain_set, Test_set = train_test_split(train_sample_metadata,test_size=0.1,random_state=42,stratify=train_sample_metadata['label'])\n\nprint(Train_set.shape, Test_set.shape )","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:29.329013Z","iopub.execute_input":"2024-05-02T04:17:29.329367Z","iopub.status.idle":"2024-05-02T04:17:30.204928Z","shell.execute_reply.started":"2024-05-02T04:17:29.329334Z","shell.execute_reply":"2024-05-02T04:17:30.203897Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_data, train_labels = prepare_all_videos(Train_set, \"train\")\ntest_data, test_labels = prepare_all_videos(Test_set, \"test\")\n\nprint(f\"Frame features in train set: {train_data[0].shape}\")\nprint(f\"Frame masks in train set: {train_data[1].shape}\")","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:30.206425Z","iopub.execute_input":"2024-05-02T04:17:30.207118Z","iopub.status.idle":"2024-05-02T04:17:30.290672Z","shell.execute_reply.started":"2024-05-02T04:17:30.207078Z","shell.execute_reply":"2024-05-02T04:17:30.289738Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## The sequence model","metadata":{}},{"cell_type":"code","source":"frame_features_input = keras.Input((MAX_SEQ_LENGTH, NUM_FEATURES))\nmask_input = keras.Input((MAX_SEQ_LENGTH,), dtype=\"bool\")\n\n# Refer to the following tutorial to understand the significance of using `mask`:\n# https://keras.io/api/layers/recurrent_layers/gru/\nx = keras.layers.GRU(16, return_sequences=True)(\n    frame_features_input, mask=mask_input\n)\nx = keras.layers.GRU(8)(x)\nx = keras.layers.Dropout(0.4)(x)\nx = keras.layers.Dense(8, activation=\"relu\")(x)\noutput = keras.layers.Dense(1, activation=\"sigmoid\")(x)\n\nmodel = keras.Model([frame_features_input, mask_input], output)\n\nmodel.compile(loss=\"binary_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])\nmodel.summary()","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:30.291925Z","iopub.execute_input":"2024-05-02T04:17:30.294348Z","iopub.status.idle":"2024-05-02T04:17:30.637462Z","shell.execute_reply.started":"2024-05-02T04:17:30.294319Z","shell.execute_reply":"2024-05-02T04:17:30.636429Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"checkpoint = keras.callbacks.ModelCheckpoint('./', save_weights_only=True, save_best_only=True)\nhistory = model.fit(\n        [train_data[0], train_data[1]],\n        train_labels,\n        validation_data=([test_data[0], test_data[1]],test_labels),\n        callbacks=[checkpoint],\n        epochs=EPOCHS,\n        batch_size=8\n    )","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:17:30.638976Z","iopub.execute_input":"2024-05-02T04:17:30.639642Z","iopub.status.idle":"2024-05-02T04:17:31.02327Z","shell.execute_reply.started":"2024-05-02T04:17:30.639602Z","shell.execute_reply":"2024-05-02T04:17:31.021998Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Inference","metadata":{}},{"cell_type":"code","source":"def prepare_single_video(frames):\n    frames = frames[None, ...]\n    frame_mask = np.zeros(shape=(1, MAX_SEQ_LENGTH,), dtype=\"bool\")\n    frame_features = np.zeros(shape=(1, MAX_SEQ_LENGTH, NUM_FEATURES), dtype=\"float32\")\n\n    for i, batch in enumerate(frames):\n        video_length = batch.shape[0]\n        length = min(MAX_SEQ_LENGTH, video_length)\n        for j in range(length):\n            frame_features[i, j, :] = feature_extractor.predict(batch[None, j, :])\n        frame_mask[i, :length] = 1  # 1 = not masked, 0 = masked\n\n    return frame_features, frame_mask\n\ndef sequence_prediction(path):\n    frames = load_video(os.path.join(DATA_FOLDER, TEST_FOLDER,path))\n    frame_features, frame_mask = prepare_single_video(frames)\n    return model.predict([frame_features, frame_mask])[0]\n    \n# This utility is for visualization.\n# Referenced from:\n# https://www.tensorflow.org/hub/tutorials/action_recognition_with_tf_hub\ndef to_gif(images):\n    converted_images = images.astype(np.uint8)\n    imageio.mimsave(\"animation.gif\", converted_images, fps=10)\n    return embed.embed_file(\"animation.gif\")\n\n\ntest_video = np.random.choice(test_videos[\"video\"].values.tolist())\nprint(f\"Test video path: {test_video}\")\n\nif(sequence_prediction(test_video)>=0.5):\n    print(f'The predicted class of the video is FAKE')\nelse:\n    print(f'The predicted class of the video is REAL')\n\nplay_video(test_video,TEST_FOLDER)","metadata":{"execution":{"iopub.status.busy":"2024-05-02T04:18:33.99496Z","iopub.execute_input":"2024-05-02T04:18:33.995911Z","iopub.status.idle":"2024-05-02T04:18:45.44372Z","shell.execute_reply.started":"2024-05-02T04:18:33.995874Z","shell.execute_reply":"2024-05-02T04:18:45.440786Z"},"trusted":true},"execution_count":null,"outputs":[]}]}