{"metadata":{"kernelspec":{"name":"python3","display_name":"Python 3","language":"python"},"language_info":{"name":"python","version":"3.11.13","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"colab":{"machine_shape":"hm","gpuType":"T4"},"accelerator":"GPU","kaggle":{"accelerator":"gpu","dataSources":[{"sourceId":20270,"databundleVersionId":1222630,"isSourceIdPinned":false,"sourceType":"competition"},{"sourceId":25383,"databundleVersionId":2684322,"isSourceIdPinned":false,"sourceType":"competition"},{"sourceId":9797223,"sourceType":"datasetVersion","datasetId":982170},{"sourceId":267279633,"sourceType":"kernelVersion"}],"dockerImageVersionId":31090,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import sys\nsys.path.append(\"../input/tez-lib/\")\nimport random\nimport numpy as np\nimport torch\nimport os\n\ndef seed_everything(seed):\n    random.seed(seed)\n    os.environ['PYTHONHASHSEED'] = str(seed)\n    np.random.seed(seed)\n    torch.manual_seed(seed)\n    torch.cuda.manual_seed(seed)\n    torch.backends.cudnn.deterministic = True\n    \nseed_everything(42)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-11T13:42:07.612702Z","iopub.execute_input":"2025-10-11T13:42:07.612892Z","iopub.status.idle":"2025-10-11T13:42:11.691556Z","shell.execute_reply.started":"2025-10-11T13:42:07.612868Z","shell.execute_reply":"2025-10-11T13:42:11.690646Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"import torch.nn as nn\nfrom tez import Tez, TezConfig\nimport tez\nimport albumentations\nimport pandas as pd\nimport cv2\nimport numpy as np\nimport timm\nimport torch.nn as nn\nfrom sklearn import metrics\nimport torch\nfrom tez.callbacks import EarlyStopping\nfrom tqdm import tqdm","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-11T13:42:11.693259Z","iopub.execute_input":"2025-10-11T13:42:11.694323Z","iopub.status.idle":"2025-10-11T13:42:22.715697Z","shell.execute_reply.started":"2025-10-11T13:42:11.694295Z","shell.execute_reply":"2025-10-11T13:42:22.715021Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"class args:\n    batch_size = 10\n    image_size = 384\n    epochs = 1 #10\n    fold = 0","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-11T13:42:22.716454Z","iopub.execute_input":"2025-10-11T13:42:22.716856Z","iopub.status.idle":"2025-10-11T13:42:22.721395Z","shell.execute_reply.started":"2025-10-11T13:42:22.716837Z","shell.execute_reply":"2025-10-11T13:42:22.720565Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"import torch.nn as nn\n\nfrom tez import Tez, TezConfig\nimport tez\nimport albumentations\nimport pandas as pd\nimport cv2\nimport numpy as np\nimport timm\nimport torch.nn as nn\nfrom sklearn import metrics\nimport torch\nfrom tez.callbacks import EarlyStopping\nfrom tqdm import tqdm","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-11T13:42:22.722196Z","iopub.execute_input":"2025-10-11T13:42:22.72277Z","iopub.status.idle":"2025-10-11T13:42:22.739477Z","shell.execute_reply.started":"2025-10-11T13:42:22.722743Z","shell.execute_reply":"2025-10-11T13:42:22.738647Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"class args:\n    batch_size = 8\n    image_size = 384\n    epochs = 1 #10\n    fold = 0","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-11T13:42:22.741537Z","iopub.execute_input":"2025-10-11T13:42:22.741817Z","iopub.status.idle":"2025-10-11T13:42:22.756004Z","shell.execute_reply.started":"2025-10-11T13:42:22.741798Z","shell.execute_reply":"2025-10-11T13:42:22.755061Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"class CustomDataset:\n    def __init__(self, image_paths, dense_features, targets, augmentations):\n        self.image_paths = image_paths\n        self.dense_features = dense_features\n        self.targets = targets\n        self.augmentations = augmentations\n        \n    def __len__(self):\n        return len(self.image_paths)\n    \n    def __getitem__(self, item):\n        image = cv2.imread(self.image_paths[item])\n        image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)\n        \n        if self.augmentations is not None:\n            augmented = self.augmentations(image=image)\n            image = augmented[\"image\"]\n            \n        image = np.transpose(image, (2, 0, 1)).astype(np.float32)\n        \n        features = self.dense_features[item, :]\n        targets = self.targets[item]\n        \n        return {\n            \"image\": torch.tensor(image, dtype=torch.float),\n            \"features\": torch.tensor(features, dtype=torch.float),\n            \"targets\": torch.tensor(targets, dtype=torch.float),\n        }","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-11T13:42:22.756944Z","iopub.execute_input":"2025-10-11T13:42:22.757283Z","iopub.status.idle":"2025-10-11T13:42:22.772339Z","shell.execute_reply.started":"2025-10-11T13:42:22.757231Z","shell.execute_reply":"2025-10-11T13:42:22.771589Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"class CustomModel(nn.Module):\n    def __init__(self):\n        super().__init__()        \n        self.model = timm.create_model(\"resnext50_32x4d\", pretrained=True, in_chans=3)\n        \n        self.dropout = nn.Dropout(0.4)\n        self.out = nn.Linear(1000, 1)\n        \n        self.step_scheduler_after = \"epoch\"\n        \n    def monitor_metrics(self, outputs, targets, loss):\n        valid_binaryloss = loss\n        if str(valid_binaryloss) == 'nan':\n            valid_binaryloss = float('inf')\n        return {\"binaryloss\": valid_binaryloss}\n    \n    def optimizer_scheduler(self):\n        opt = torch.optim.AdamW(self.parameters(), lr=3e-05, weight_decay=0.015)\n        sch = torch.optim.lr_scheduler.CosineAnnealingWarmRestarts(\n            opt, T_0=10, T_mult=1, eta_min=5e-7, last_epoch=-1\n        )\n        return opt, sch\n    \n    def forward(self, image, features, targets=None):\n        x = self.model(image)\n        x = self.dropout(x)\n        x = self.out(x)\n        \n        if targets is not None:\n            loss = nn.BCEWithLogitsLoss()(x, targets.view(-1, 1).type_as(x))\n            metrics = self.monitor_metrics(x, targets, loss)\n            return x, loss, metrics\n        return x, 0, {}","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-11T13:42:22.773331Z","iopub.execute_input":"2025-10-11T13:42:22.773683Z","iopub.status.idle":"2025-10-11T13:42:22.790577Z","shell.execute_reply.started":"2025-10-11T13:42:22.773663Z","shell.execute_reply":"2025-10-11T13:42:22.789718Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"train_aug = albumentations.Compose(\n    [\n        albumentations.LongestMaxSize(args.image_size, p=1),\n        albumentations.PadIfNeeded(args.image_size, args.image_size, p=1, border_mode=0),\n        \n        albumentations.HorizontalFlip(p=0.5),\n        albumentations.VerticalFlip(p=0.15),\n        albumentations.Rotate(limit=180, p=0.5),\n        albumentations.ShiftScaleRotate(\n            shift_limit=0.12, scale_limit=0.12, rotate_limit=45, p=0.5\n        ),\n        \n        albumentations.HueSaturationValue(\n            hue_shift_limit=0.25, sat_shift_limit=0.25, val_shift_limit=0.25, p=0.5\n        ),\n        albumentations.RandomBrightnessContrast(\n            brightness_limit=(-0.15, 0.15), contrast_limit=(-0.15, 0.15), p=0.5\n        ),\n        albumentations.Normalize(\n            mean=[0.485, 0.456, 0.406],\n            std=[0.229, 0.224, 0.225],\n            max_pixel_value=255.0,\n            p=1.0,\n        ),\n    ],\n    p=1.0,\n)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-11T13:42:22.791487Z","iopub.execute_input":"2025-10-11T13:42:22.791805Z","iopub.status.idle":"2025-10-11T13:42:22.814873Z","shell.execute_reply.started":"2025-10-11T13:42:22.791778Z","shell.execute_reply":"2025-10-11T13:42:22.813934Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"valid_aug = albumentations.Compose(\n    [\n        albumentations.LongestMaxSize(args.image_size, p=1),\n        albumentations.PadIfNeeded(args.image_size, args.image_size, p=1, border_mode=0),\n        albumentations.Normalize(\n            mean=[0.485, 0.456, 0.406],\n            std=[0.229, 0.224, 0.225],\n            max_pixel_value=255.0,\n            p=1.0,\n        ),\n    ],\n    p=1.0,\n)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-11T13:42:22.815916Z","iopub.execute_input":"2025-10-11T13:42:22.816292Z","iopub.status.idle":"2025-10-11T13:42:22.824157Z","shell.execute_reply.started":"2025-10-11T13:42:22.816266Z","shell.execute_reply":"2025-10-11T13:42:22.823296Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"df = pd.read_csv(\"/kaggle/input/10148-aditya-kfold-melanoma-my/train_5folds.csv\")\ndf.head()\n\ni = 0\nprint(f'training fold: {i} start')\nargs.fold = 0\ndf_train = df[df.kfold != args.fold].reset_index(drop=True)\ndf_valid = df[df.kfold == args.fold].reset_index(drop=True)\n\ndense_features = [\n    'age_approx'\n]","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-11T13:42:22.824933Z","iopub.execute_input":"2025-10-11T13:42:22.825208Z","iopub.status.idle":"2025-10-11T13:42:22.937657Z","shell.execute_reply.started":"2025-10-11T13:42:22.825178Z","shell.execute_reply":"2025-10-11T13:42:22.936863Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"train_img_paths = [f\"/kaggle/input/siim-isic-melanoma-classification/jpeg/train/{x}.jpg\" for x in df_train[\"image_name\"].values]\nvalid_img_paths = [f\"/kaggle/input/siim-isic-melanoma-classification/jpeg/train/{x}.jpg\" for x in df_valid[\"image_name\"].values]","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-11T13:43:56.715769Z","iopub.execute_input":"2025-10-11T13:43:56.71613Z","iopub.status.idle":"2025-10-11T13:43:56.728093Z","shell.execute_reply.started":"2025-10-11T13:43:56.716108Z","shell.execute_reply":"2025-10-11T13:43:56.727317Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"train_dataset = CustomDataset(\n    image_paths=train_img_paths,\n    dense_features=df_train[dense_features].values,\n    targets=df_train.target.values,\n    augmentations=train_aug,\n)\n\nvalid_dataset = CustomDataset(\n    image_paths=valid_img_paths,\n    dense_features=df_valid[dense_features].values,\n    targets=df_valid.target.values,\n    augmentations=valid_aug,\n)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-11T13:44:09.484578Z","iopub.execute_input":"2025-10-11T13:44:09.485179Z","iopub.status.idle":"2025-10-11T13:44:09.491413Z","shell.execute_reply.started":"2025-10-11T13:44:09.485155Z","shell.execute_reply":"2025-10-11T13:44:09.490765Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"model = CustomModel()\nmodel = Tez(model)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-11T13:44:12.985421Z","iopub.execute_input":"2025-10-11T13:44:12.985751Z","iopub.status.idle":"2025-10-11T13:44:15.081229Z","shell.execute_reply.started":"2025-10-11T13:44:12.985728Z","shell.execute_reply":"2025-10-11T13:44:15.080507Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"config = TezConfig(\n    training_batch_size=args.batch_size,\n    validation_batch_size=2 * args.batch_size,\n    epochs=args.epochs,\n    step_scheduler_after=\"epoch\",\n    step_scheduler_metric=\"valid_binaryloss\",\n    fp16=True,\n    val_strategy=\"batch\",\n    val_steps=900,\n)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-11T13:44:23.705851Z","iopub.execute_input":"2025-10-11T13:44:23.706424Z","iopub.status.idle":"2025-10-11T13:44:23.710686Z","shell.execute_reply.started":"2025-10-11T13:44:23.706401Z","shell.execute_reply":"2025-10-11T13:44:23.70975Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"es = EarlyStopping(\n    monitor=\"valid_binaryloss\",\n    model_path=f\"model_f{args.fold}.bin\",\n    patience=4,\n    mode=\"min\",\n    save_weights_only=True,\n)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-11T13:44:26.441527Z","iopub.execute_input":"2025-10-11T13:44:26.442138Z","iopub.status.idle":"2025-10-11T13:44:26.446198Z","shell.execute_reply.started":"2025-10-11T13:44:26.442114Z","shell.execute_reply":"2025-10-11T13:44:26.445542Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"model.fit(\n    train_dataset,\n    valid_dataset=valid_dataset,\n    callbacks=[es],\n    config=config,\n)\n\nprint(f'training fold: {i} complete')","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-11T13:44:29.259693Z","iopub.execute_input":"2025-10-11T13:44:29.260332Z","iopub.status.idle":"2025-10-11T14:32:01.146187Z","shell.execute_reply.started":"2025-10-11T13:44:29.26031Z","shell.execute_reply":"2025-10-11T14:32:01.145318Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null}]}