{"metadata":{"kernelspec":{"name":"python3","display_name":"Python 3","language":"python"},"language_info":{"codemirror_mode":{"name":"ipython","version":3},"file_extension":".py","mimetype":"text/x-python","name":"python","nbconvert_exporter":"python","pygments_lexer":"ipython3","version":"3.10.12"},"kaggle":{"accelerator":"gpu","dataSources":[{"sourceId":5441,"databundleVersionId":38425,"sourceType":"competition"}],"dockerImageVersionId":30636,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"Kaggle link: https://www.kaggle.com/c/dogs-vs-cats-redux-kernels-edition","metadata":{}},{"cell_type":"markdown","source":"#### Kaggle notebook: https://www.kaggle.com/code/paulabal/dogs-vs-cats-cnn\n#### Wandb logs: https://wandb.ai/lefugdot/cats-vs-dogs","metadata":{}},{"cell_type":"code","source":"from kaggle_secrets import UserSecretsClient\nuser_secrets = UserSecretsClient()\nsecret = user_secrets.get_secret(\"wandb-key\")\n\nimport wandb\nwandb.login(key=secret)","metadata":{"execution":{"iopub.status.busy":"2024-01-14T21:14:02.777671Z","iopub.execute_input":"2024-01-14T21:14:02.778076Z","iopub.status.idle":"2024-01-14T21:14:03.023248Z","shell.execute_reply.started":"2024-01-14T21:14:02.778044Z","shell.execute_reply":"2024-01-14T21:14:03.022171Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# This Python 3 environment comes with many helpful analytics libraries installed\n# It is defined by the kaggle/python Docker image: https://github.com/kaggle/docker-python\n# For example, here's several helpful packages to load\n\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n\n# Input data files are available in the read-only \"../input/\" directory\n# For example, running this (by clicking run or pressing Shift+Enter) will list all files under the input directory\n\nimport os\nfor dirname, _, filenames in os.walk('/kaggle/input'):\n    for filename in filenames:\n        print(os.path.join(dirname, filename))\n\n# You can write up to 20GB to the current directory (/kaggle/working/) that gets preserved as output when you create a version using \"Save & Run All\" \n# You can also write temporary files to /kaggle/temp/, but they won't be saved outside of the current session","metadata":{"_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","execution":{"iopub.status.busy":"2024-01-14T19:36:33.879785Z","iopub.execute_input":"2024-01-14T19:36:33.880132Z","iopub.status.idle":"2024-01-14T19:36:34.177401Z","shell.execute_reply.started":"2024-01-14T19:36:33.880106Z","shell.execute_reply":"2024-01-14T19:36:34.176483Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Import everything needed","metadata":{}},{"cell_type":"code","source":"import zipfile\nimport glob\nfrom PIL import Image\nimport matplotlib.pyplot as plt\n\nfrom sklearn.model_selection import train_test_split\nimport torch\nimport torch.nn as nn\nimport torch.nn.functional as F\nimport torch.optim as optim\nfrom torch.utils.data import DataLoader, Dataset\nfrom torchvision import datasets, transforms\n\nimport wandb \nwandb.init(project='cats-vs-dogs')\n\nnp.random.seed(0)\ntorch.manual_seed(0)\ntorch.cuda.manual_seed(0)","metadata":{"execution":{"iopub.status.busy":"2024-01-14T19:36:37.441350Z","iopub.execute_input":"2024-01-14T19:36:37.441920Z","iopub.status.idle":"2024-01-14T19:37:11.089542Z","shell.execute_reply.started":"2024-01-14T19:36:37.441886Z","shell.execute_reply":"2024-01-14T19:37:11.088675Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Unzip datasets","metadata":{}},{"cell_type":"code","source":"train_dir = 'train'\ntest_dir = 'test'\nwith zipfile.ZipFile('/kaggle/input/dogs-vs-cats-redux-kernels-edition/train.zip') as train_zip:\n    train_zip.extractall('')\n    \nwith zipfile.ZipFile('/kaggle/input/dogs-vs-cats-redux-kernels-edition/test.zip') as test_zip:\n    test_zip.extractall('')\ntrain_list = glob.glob(os.path.join(train_dir,'*.jpg'))\ntest_list = glob.glob(os.path.join(test_dir, '*.jpg'))\nprint(f\"Train Data: {len(train_list)}\")\nprint(f\"Test Data: {len(test_list)}\")","metadata":{"execution":{"iopub.status.busy":"2024-01-14T19:37:17.478950Z","iopub.execute_input":"2024-01-14T19:37:17.479337Z","iopub.status.idle":"2024-01-14T19:37:33.427468Z","shell.execute_reply.started":"2024-01-14T19:37:17.479306Z","shell.execute_reply":"2024-01-14T19:37:33.425909Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"labels = [path.split('/')[-1].split('.')[0] for path in train_list]","metadata":{"execution":{"iopub.status.busy":"2024-01-14T19:37:36.187041Z","iopub.execute_input":"2024-01-14T19:37:36.187900Z","iopub.status.idle":"2024-01-14T19:37:36.207891Z","shell.execute_reply.started":"2024-01-14T19:37:36.187868Z","shell.execute_reply":"2024-01-14T19:37:36.206816Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Plot random image with their label","metadata":{}},{"cell_type":"code","source":"random_idx = np.random.randint(1, len(train_list), size=9)\nfig, axes = plt.subplots(3, 3, figsize=(16, 12))\n\nfor idx, ax in enumerate(axes.ravel()):\n    img = Image.open(train_list[idx])\n    ax.set_title(labels[idx])\n    ax.imshow(img)","metadata":{"execution":{"iopub.status.busy":"2024-01-14T19:37:38.433949Z","iopub.execute_input":"2024-01-14T19:37:38.434833Z","iopub.status.idle":"2024-01-14T19:37:40.916108Z","shell.execute_reply.started":"2024-01-14T19:37:38.434796Z","shell.execute_reply":"2024-01-14T19:37:40.914809Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Use Sklearn to split the training data","metadata":{}},{"cell_type":"code","source":"train_list, valid_list = train_test_split(train_list, \n                                          test_size=0.2,\n                                          stratify=labels,\n                                          random_state=0)","metadata":{"execution":{"iopub.status.busy":"2024-01-14T19:37:44.715604Z","iopub.execute_input":"2024-01-14T19:37:44.716349Z","iopub.status.idle":"2024-01-14T19:37:44.751276Z","shell.execute_reply.started":"2024-01-14T19:37:44.716298Z","shell.execute_reply":"2024-01-14T19:37:44.750343Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(f\"Train Data: {len(train_list)}\")\nprint(f\"Validation Data: {len(valid_list)}\")\nprint(f\"Test Data: {len(test_list)}\")","metadata":{"execution":{"iopub.status.busy":"2024-01-14T19:37:46.765557Z","iopub.execute_input":"2024-01-14T19:37:46.765948Z","iopub.status.idle":"2024-01-14T19:37:46.773100Z","shell.execute_reply.started":"2024-01-14T19:37:46.765917Z","shell.execute_reply":"2024-01-14T19:37:46.772156Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"We will discuss this in more detail in a near future...","metadata":{}},{"cell_type":"code","source":"train_transforms = transforms.Compose([\n    transforms.RandomResizedCrop(224),\n    transforms.RandomHorizontalFlip(),\n    transforms.ToTensor(),  \n    transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])  \n])\n\ntest_transforms = transforms.Compose([\n    transforms.RandomResizedCrop(224),\n    transforms.RandomHorizontalFlip(),\n    transforms.ToTensor(),  \n    transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])  \n])","metadata":{"execution":{"iopub.status.busy":"2024-01-14T19:37:47.740143Z","iopub.execute_input":"2024-01-14T19:37:47.740540Z","iopub.status.idle":"2024-01-14T19:37:47.748993Z","shell.execute_reply.started":"2024-01-14T19:37:47.740508Z","shell.execute_reply":"2024-01-14T19:37:47.748005Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Define the dataset using PIL to read image","metadata":{}},{"cell_type":"code","source":"class CatsDogsDataset(Dataset):\n    def __init__(self, file_list, transform=None):\n        self.file_list = file_list\n        self.transform = transform\n        self.filelength = len(file_list)\n\n    def __len__(self):\n        return self.filelength\n\n    def __getitem__(self, idx):\n        img_path = self.file_list[idx]\n        img = Image.open(img_path)\n        img_transformed = self.transform(img)\n        label = img_path.split(\"/\")[-1].split(\".\")[0]\n        label = 1 if label == \"dog\" else 0\n        \n        base_name = os.path.basename(img_path)\n        filename, _ = os.path.splitext(base_name)\n        return img_transformed, label, filename","metadata":{"execution":{"iopub.status.busy":"2024-01-14T19:37:49.168483Z","iopub.execute_input":"2024-01-14T19:37:49.168886Z","iopub.status.idle":"2024-01-14T19:37:49.178048Z","shell.execute_reply.started":"2024-01-14T19:37:49.168854Z","shell.execute_reply":"2024-01-14T19:37:49.177097Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_data = CatsDogsDataset(train_list, transform=train_transforms)\nvalid_data = CatsDogsDataset(valid_list, transform=test_transforms)\ntest_data = CatsDogsDataset(test_list, transform=test_transforms)","metadata":{"execution":{"iopub.status.busy":"2024-01-14T19:37:50.787690Z","iopub.execute_input":"2024-01-14T19:37:50.788139Z","iopub.status.idle":"2024-01-14T19:37:50.795257Z","shell.execute_reply.started":"2024-01-14T19:37:50.788104Z","shell.execute_reply":"2024-01-14T19:37:50.794153Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Create dataloader, you can modify the batch size if needed","metadata":{}},{"cell_type":"code","source":"batch_size = 32\ntrain_loader = DataLoader(dataset=train_data, batch_size=batch_size, shuffle=True)\nvalid_loader = DataLoader(dataset=valid_data, batch_size=batch_size, shuffle=False)\ntest_loader = DataLoader(dataset=test_data, batch_size=batch_size, shuffle=False)","metadata":{"execution":{"iopub.status.busy":"2024-01-14T19:37:51.851465Z","iopub.execute_input":"2024-01-14T19:37:51.851843Z","iopub.status.idle":"2024-01-14T19:37:51.859278Z","shell.execute_reply.started":"2024-01-14T19:37:51.851809Z","shell.execute_reply":"2024-01-14T19:37:51.858263Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"A CNN for Cats / Dogs classification. ","metadata":{}},{"cell_type":"code","source":"class CatsOrDogsCNN(nn.Module):\n    def __init__(self):\n        super(CatsOrDogsCNN, self).__init__()\n        self.conv1 = nn.Conv2d(3, 32, 3, padding=1)\n        self.bn1 = nn.BatchNorm2d(32)\n        self.conv2 = nn.Conv2d(32, 64, 3, padding=1)\n        self.bn2 = nn.BatchNorm2d(64)\n        self.conv3 = nn.Conv2d(64, 128, 3, padding=1)\n        self.bn3 = nn.BatchNorm2d(128)\n        self.conv4 = nn.Conv2d(128, 256, 3, padding=1)\n        self.bn4 = nn.BatchNorm2d(256)\n        self.conv5 = nn.Conv2d(256, 512, 3, padding=1)\n        self.bn5 = nn.BatchNorm2d(512)\n        self.pool = nn.MaxPool2d(2, 2)\n        self.fc1 = nn.Linear(512 * 7 * 7, 2048)\n        self.dropout1 = nn.Dropout(0.5)\n        self.fc2 = nn.Linear(2048, 1024)\n        self.dropout2 = nn.Dropout(0.5)\n        self.fc3 = nn.Linear(1024, 512)\n        self.dropout3 = nn.Dropout(0.5)\n        self.fc4 = nn.Linear(512, 2)\n\n    def forward(self, x):\n        x = self.pool(F.relu(self.bn1(self.conv1(x))))\n        x = self.pool(F.relu(self.bn2(self.conv2(x))))\n        x = self.pool(F.relu(self.bn3(self.conv3(x))))\n        x = self.pool(F.relu(self.bn4(self.conv4(x))))\n        x = self.pool(F.relu(self.bn5(self.conv5(x))))\n        x = x.view(-1, 512 * 7 * 7)\n        x = self.dropout1(F.relu(self.fc1(x)))\n        x = self.dropout2(F.relu(self.fc2(x)))\n        x = self.dropout3(F.relu(self.fc3(x)))\n        x = self.fc4(x)\n        return x","metadata":{"execution":{"iopub.status.busy":"2024-01-14T19:37:53.433293Z","iopub.execute_input":"2024-01-14T19:37:53.433944Z","iopub.status.idle":"2024-01-14T19:37:53.449726Z","shell.execute_reply.started":"2024-01-14T19:37:53.433908Z","shell.execute_reply":"2024-01-14T19:37:53.448578Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def calculate_accuracy(y_pred, y_true):\n    _, predicted = torch.max(y_pred.data, 1)\n    total = y_true.size(0)\n    correct = (predicted == y_true).sum().item()\n    return correct / total","metadata":{"execution":{"iopub.status.busy":"2024-01-14T19:37:55.700269Z","iopub.execute_input":"2024-01-14T19:37:55.700951Z","iopub.status.idle":"2024-01-14T19:37:55.707179Z","shell.execute_reply.started":"2024-01-14T19:37:55.700915Z","shell.execute_reply":"2024-01-14T19:37:55.706049Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Training loop and logging to wandb","metadata":{}},{"cell_type":"code","source":"device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\nmodel = CatsOrDogsCNN().to(device)\n\nlr = 0.001\nmomentum = 0.9\nepochs = 50\n\ncriterion = nn.CrossEntropyLoss()\noptimizer = optim.SGD(model.parameters(), lr=lr, momentum=momentum)\n\nfor epoch in range(epochs):\n    training_loss = 0.0 \n    training_accuracy = 0.0\n\n    for i,data in enumerate(train_loader):\n        input, labels = data[0].to(device), data[1].to(device)\n        optimizer.zero_grad()\n\n        output = model(input)\n        loss = criterion(output, labels)\n        loss.backward()\n        optimizer.step()\n\n        accuracy = calculate_accuracy(output, labels)\n\n        training_loss += loss.item()\n        training_accuracy += accuracy\n\n\n    average_loss = training_loss / len(train_loader)\n    average_accuracy = training_accuracy / len(train_loader)\n    wandb.log({ 'loss': average_loss, 'accuracy': average_accuracy})\n\n    print(f'Epoch: {epoch + 1}, Loss: {average_loss}, Accuracy: {average_accuracy}')\n\n","metadata":{"execution":{"iopub.status.busy":"2024-01-14T19:37:57.051365Z","iopub.execute_input":"2024-01-14T19:37:57.051743Z","iopub.status.idle":"2024-01-14T20:57:29.684348Z","shell.execute_reply.started":"2024-01-14T19:37:57.051706Z","shell.execute_reply":"2024-01-14T20:57:29.683333Z"},"collapsed":true,"jupyter":{"outputs_hidden":true},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# torch.save(model.state_dict(), 'model_state_dict.pth')","metadata":{"execution":{"iopub.status.busy":"2024-01-14T21:08:07.769105Z","iopub.execute_input":"2024-01-14T21:08:07.769527Z","iopub.status.idle":"2024-01-14T21:08:08.294005Z","shell.execute_reply.started":"2024-01-14T21:08:07.769497Z","shell.execute_reply":"2024-01-14T21:08:08.292937Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Validation & Testing loops","metadata":{}},{"cell_type":"code","source":"with torch.no_grad():\n    for i, data in enumerate(valid_loader):\n        input, labels = data[0].to(device), data[1].to(device)\n        output = model(input)\n        loss = criterion(output, labels)\n        accuracy = calculate_accuracy(output, labels)\n\n        wandb.log({\"validation_loss\": loss.item(), \"validation_accuracy\": accuracy })\n        print(f'Validation Loss: {loss.item()}, Validation Accuracy: {accuracy}')","metadata":{"execution":{"iopub.status.busy":"2024-01-14T21:12:27.734638Z","iopub.execute_input":"2024-01-14T21:12:27.735567Z","iopub.status.idle":"2024-01-14T21:12:45.486516Z","shell.execute_reply.started":"2024-01-14T21:12:27.735532Z","shell.execute_reply":"2024-01-14T21:12:45.484829Z"},"collapsed":true,"jupyter":{"outputs_hidden":true},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Save the testing results","metadata":{}},{"cell_type":"code","source":"results = []\n\nfor i, data in enumerate(test_loader):\n    input, _, filename = data\n    input = input.to(device)\n    output = model(input)\n    probabilities = F.softmax(output.data, dim=1)\n    dog_probabilities = probabilities[:, 1]\n    for file, prob in zip(filename, dog_probabilities):\n        results.append({'id': file, 'label': prob.item()})","metadata":{"execution":{"iopub.status.busy":"2024-01-14T21:10:02.354301Z","iopub.execute_input":"2024-01-14T21:10:02.354980Z","iopub.status.idle":"2024-01-14T21:10:46.994917Z","shell.execute_reply.started":"2024-01-14T21:10:02.354945Z","shell.execute_reply":"2024-01-14T21:10:46.993840Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Save the results to a submission file. ","metadata":{}},{"cell_type":"code","source":"import csv\n\nwith open('submission.csv', 'w', newline='') as csvfile:\n    fieldnames = ['id', 'label']\n    writer = csv.DictWriter(csvfile, fieldnames=fieldnames)\n    writer.writeheader()\n    for result in results:\n        writer.writerow(result)","metadata":{"execution":{"iopub.status.busy":"2024-01-14T21:10:56.754910Z","iopub.execute_input":"2024-01-14T21:10:56.755661Z","iopub.status.idle":"2024-01-14T21:10:56.822500Z","shell.execute_reply.started":"2024-01-14T21:10:56.755629Z","shell.execute_reply":"2024-01-14T21:10:56.821294Z"},"trusted":true},"execution_count":null,"outputs":[]}]}