{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.12.12","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"gpu","dataSources":[{"sourceId":129543,"databundleVersionId":15525987,"sourceType":"competition"}],"dockerImageVersionId":31260,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"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":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true,"execution":{"iopub.status.busy":"2026-02-01T04:26:21.996952Z","iopub.execute_input":"2026-02-01T04:26:21.997237Z","iopub.status.idle":"2026-02-01T04:26:25.431476Z","shell.execute_reply.started":"2026-02-01T04:26:21.997179Z","shell.execute_reply":"2026-02-01T04:26:25.430471Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"import os\nimport sys\nimport subprocess\nimport numpy as np\nimport pandas as pd\nimport torch\nimport torch.nn as nn\nimport torch.nn.functional as F\nfrom torch.utils.data import Dataset, DataLoader\nfrom torchvision import transforms\nfrom PIL import Image\nfrom tqdm.auto import tqdm\nfrom sklearn.preprocessing import LabelEncoder\n\n# --- 1. AUTO-DETECT DATA PATH -----------------------------------------------------------\ndef find_data_root(filename='train.csv'):\n    # Common locations for Kaggle and local environments\n    search_paths = [\n        '.', \n        '/kaggle/input/round-2-jaguar-reidentification-challenge',\n        '/kaggle/input/jaguar-reidentification-challenge',\n        '../input/round-2-jaguar-reidentification-challenge',\n        '../input/jaguar-reid',\n        '../input'\n    ]\n    \n    for path in search_paths:\n        candidate = os.path.join(path, filename)\n        if os.path.exists(candidate):\n            return path\n            \n    # Recursive search if standard paths fail (slower but robust)\n    print(\"⚠️ Standard paths failed, searching recursively...\")\n    for root, dirs, files in os.walk('/kaggle/input'):\n        if filename in files:\n            return root\n    return None\n\ndata_root = find_data_root()\n\nif data_root is None:\n    raise FileNotFoundError(\"❌ CRITICAL ERROR: Could not find 'train.csv'. Please make sure the dataset is added to your notebook.\")\nelse:\n    print(f\"✅ Found dataset in: {data_root}\")\n\n# --- 2. CONFIGURATION -------------------------------------------------------------------\n# Install timm if not present\ntry:\n    import timm\nexcept ImportError:\n    subprocess.check_call([sys.executable, \"-m\", \"pip\", \"install\", \"timm\"])\n    import timm\n\nCONFIG = {\n    \"data_dir\": data_root,\n    \"train_dir\": os.path.join(data_root, \"train\"),\n    \"test_dir\": os.path.join(data_root, \"test\"),\n    \"train_csv\": \"train.csv\",\n    \"test_csv\": \"test.csv\",\n    \"model_name\": \"resnet50\", \n    \"img_size\": 224,\n    \"batch_size\": 32,\n    \"epochs\": 5,           \n    \"lr\": 3e-4,\n    \"device\": torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\"),\n    \"num_workers\": 2\n}\n\nprint(f\"🚀 Running on {CONFIG['device']} with {CONFIG['model_name']}...\")\n\n# --- 3. DATASET CLASS -------------------------------------------------------------------\nclass JaguarDataset(Dataset):\n    def __init__(self, df, img_dir, transform=None, is_test=False):\n        self.df = df\n        self.img_dir = img_dir\n        self.transform = transform\n        self.is_test = is_test\n        if self.is_test:\n            self.image_names = df['filename'].values\n            self.labels = None\n        else:\n            self.image_names = df['filename'].values\n            self.labels = df['label_idx'].values\n\n    def __len__(self):\n        return len(self.image_names)\n\n    def __getitem__(self, idx):\n        img_name = self.image_names[idx]\n        img_path = os.path.join(self.img_dir, img_name)\n        \n        try:\n            image = Image.open(img_path).convert(\"RGB\")\n        except Exception as e:\n            # Fallback (black image) so training doesn't crash on one missing file\n            image = Image.new('RGB', (224, 224))\n            \n        if self.transform:\n            image = self.transform(image)\n            \n        if self.is_test:\n            return image, img_name\n        else:\n            label = torch.tensor(self.labels[idx], dtype=torch.long)\n            return image, label\n\n# --- 4. PREPARATION ---------------------------------------------------------------------\n# Load Metadata\ntrain_csv_path = os.path.join(CONFIG['data_dir'], CONFIG['train_csv'])\ntest_csv_path = os.path.join(CONFIG['data_dir'], CONFIG['test_csv'])\n\ntrain_df = pd.read_csv(train_csv_path)\ntest_df_pairs = pd.read_csv(test_csv_path)\n\n# Encode Labels\nle = LabelEncoder()\ntrain_df['label_idx'] = le.fit_transform(train_df['ground_truth'])\nnum_classes = len(le.classes_)\nprint(f\"📋 Loaded {len(train_df)} train images. Classes: {num_classes}\")\n\n# Transforms\ntrain_transforms = transforms.Compose([\n    transforms.Resize((CONFIG['img_size'], CONFIG['img_size'])),\n    transforms.RandomHorizontalFlip(),\n    transforms.ToTensor(),\n    transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])\n])\n\ntest_transforms = transforms.Compose([\n    transforms.Resize((CONFIG['img_size'], CONFIG['img_size'])),\n    transforms.ToTensor(),\n    transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])\n])\n\n# DataLoaders\ntrain_dataset = JaguarDataset(train_df, CONFIG['train_dir'], transform=train_transforms)\ntrain_loader = DataLoader(train_dataset, batch_size=CONFIG['batch_size'], shuffle=True, num_workers=CONFIG['num_workers'])\n\n# --- 5. MODEL DEFINITION ----------------------------------------------------------------\nclass JaguarReIDModel(nn.Module):\n    def __init__(self, model_name, num_classes, pretrained=True):\n        super().__init__()\n        self.backbone = timm.create_model(model_name, pretrained=pretrained, num_classes=0)\n        in_features = self.backbone.num_features\n        self.head = nn.Linear(in_features, num_classes)\n        \n    def forward(self, x, return_embeddings=False):\n        features = self.backbone(x)\n        if return_embeddings:\n            return F.normalize(features, p=2, dim=1)\n        return self.head(features)\n\nmodel = JaguarReIDModel(CONFIG['model_name'], num_classes).to(CONFIG['device'])\noptimizer = torch.optim.Adam(model.parameters(), lr=CONFIG['lr'])\ncriterion = nn.CrossEntropyLoss()\n\n# --- 6. TRAINING LOOP -------------------------------------------------------------------\nprint(\"🏋️ Starting Training...\")\nmodel.train()\nfor epoch in range(CONFIG['epochs']):\n    total_loss = 0\n    pbar = tqdm(train_loader, desc=f\"Epoch {epoch+1}/{CONFIG['epochs']}\", leave=False)\n    for images, labels in pbar:\n        images, labels = images.to(CONFIG['device']), labels.to(CONFIG['device'])\n        optimizer.zero_grad()\n        outputs = model(images)\n        loss = criterion(outputs, labels)\n        loss.backward()\n        optimizer.step()\n        total_loss += loss.item()\n        pbar.set_postfix({'loss': f\"{loss.item():.4f}\"})\n    print(f\"Epoch {epoch+1} - Avg Loss: {total_loss/len(train_loader):.4f}\")\n\n# --- 7. INFERENCE & SUBMISSION ----------------------------------------------------------\nprint(\"🔍 Generating Submission...\")\nmodel.eval()\n\n# Unique test images\nunique_test_images = sorted(list(set(test_df_pairs['query_image']) | set(test_df_pairs['gallery_image'])))\ntest_img_df = pd.DataFrame({'filename': unique_test_images})\ntest_dataset = JaguarDataset(test_img_df, CONFIG['test_dir'], transform=test_transforms, is_test=True)\ntest_loader = DataLoader(test_dataset, batch_size=CONFIG['batch_size']*2, shuffle=False, num_workers=CONFIG['num_workers'])\n\nembeddings = {}\nwith torch.no_grad():\n    for images, names in tqdm(test_loader, desc=\"Extracting Features\"):\n        images = images.to(CONFIG['device'])\n        feats = model(images, return_embeddings=True).cpu().numpy()\n        for name, feat in zip(names, feats):\n            embeddings[name] = feat\n\nsimilarities = []\nfor _, row in tqdm(test_df_pairs.iterrows(), total=len(test_df_pairs), desc=\"Computing Pairs\"):\n    sim = np.dot(embeddings[row['query_image']], embeddings[row['gallery_image']])\n    similarities.append((sim + 1) / 2) # Norm to [0, 1]\n\nsubmission = pd.DataFrame({'row_id': test_df_pairs['row_id'], 'similarity': similarities})\nsubmission['similarity'] = submission['similarity'].clip(0.0, 1.0)\nsubmission.to_csv('submission.csv', index=False)\nprint(f\"\\n✅ SUCCESS! Saved 'submission.csv' ({len(submission)} rows)\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-02-01T04:29:27.883516Z","iopub.execute_input":"2026-02-01T04:29:27.883835Z","iopub.status.idle":"2026-02-01T04:54:02.331836Z","shell.execute_reply.started":"2026-02-01T04:29:27.883809Z","shell.execute_reply":"2026-02-01T04:54:02.330806Z"}},"outputs":[],"execution_count":null}]}