{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"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"},"kaggle":{"accelerator":"nvidiaTeslaT4","dataSources":[{"sourceId":20270,"databundleVersionId":1222630,"sourceType":"competition"},{"sourceId":13312689,"sourceType":"datasetVersion","datasetId":8439135},{"sourceId":267011918,"sourceType":"kernelVersion"}],"dockerImageVersionId":31154,"isInternetEnabled":false,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import sys\nsys.path.append(\"../input/tez-lib/tez-main\")","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true,"execution":{"iopub.status.busy":"2025-10-10T16:53:24.164397Z","iopub.execute_input":"2025-10-10T16:53:24.164644Z","iopub.status.idle":"2025-10-10T16:53:24.168529Z","shell.execute_reply.started":"2025-10-10T16:53:24.164626Z","shell.execute_reply":"2025-10-10T16:53:24.167650Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"import tez\nfrom tez import Tez, TezConfig\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\nimport math","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-10T16:53:24.170199Z","iopub.execute_input":"2025-10-10T16:53:24.170487Z","iopub.status.idle":"2025-10-10T16:53:24.184883Z","shell.execute_reply.started":"2025-10-10T16:53:24.170465Z","shell.execute_reply":"2025-10-10T16:53:24.184156Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"class args:\n    batch_size = 64#16\n    image_size = 384 #64","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-10T16:53:24.186128Z","iopub.execute_input":"2025-10-10T16:53:24.186354Z","iopub.status.idle":"2025-10-10T16:53:24.202521Z","shell.execute_reply.started":"2025-10-10T16:53:24.186339Z","shell.execute_reply":"2025-10-10T16:53:24.201916Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"def sigmoid(x):\n    return 1 / (1 + math.exp(-x))","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-10T16:53:24.203201Z","iopub.execute_input":"2025-10-10T16:53:24.203364Z","iopub.status.idle":"2025-10-10T16:53:24.217908Z","shell.execute_reply.started":"2025-10-10T16:53:24.203351Z","shell.execute_reply":"2025-10-10T16:53:24.217254Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"class MelanomaDataset:\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 = image.astype(np.float32) / 255.0   \n        image = np.transpose(image, (2, 0, 1)).astype(np.float32)\n        \n        features = self.dense_features[item, :].astype(np.float32)\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":"2025-10-10T17:04:56.741208Z","iopub.execute_input":"2025-10-10T17:04:56.741539Z","iopub.status.idle":"2025-10-10T17:04:56.749679Z","shell.execute_reply.started":"2025-10-10T17:04:56.741510Z","shell.execute_reply":"2025-10-10T17:04:56.748779Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"class MelanomaModel(nn.Module):\n    def __init__(self):\n        super().__init__()        \n        self.model = timm.create_model(\"resnet50\", pretrained=False, num_classes=0)#resnet50#resnet101#eca_nfnet_l1#resnest101e\n\n        \n        self.dropout = nn.Dropout(0.5)# increase dropout\n\n        self.out = nn.Linear(2066, 1)\n\n        \n        self.step_scheduler_after = \"epoch\"\n\n\n    def monitor_metrics(self, outputs, targets, loss):\n        return {\"bce_loss\": loss}\n\n    def optimizer_scheduler(self):\n        opt = torch.optim.AdamW(self.parameters(), lr=2.5e-05, weight_decay=0.01)\n        sch = torch.optim.lr_scheduler.CosineAnnealingWarmRestarts(\n            opt, T_0=10, T_mult=1, eta_min=1e-6, last_epoch=-1\n        )\n        return opt,sch\n\n    def forward(self, image, features, targets=None):\n\n        x = self.model(image)\n        x = self.dropout(x)\n        x = torch.cat([x, features], dim=1)\n        x = self.dropout(x)\n        x = self.out(x)\n\n        if targets is not None:\n            targets = targets.view(-1, 1).float()\n            loss = nn.BCEWithLogitsLoss()(x, targets)\n            metrics = self.monitor_metrics(x, targets, loss)\n            return x, loss, metrics\n        return x, 0, {}\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-10T16:57:04.725021Z","iopub.execute_input":"2025-10-10T16:57:04.725611Z","iopub.status.idle":"2025-10-10T16:57:04.732465Z","shell.execute_reply.started":"2025-10-10T16:57:04.725589Z","shell.execute_reply":"2025-10-10T16:57:04.731643Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"test_aug = albumentations.Compose(\n    [\n        albumentations.Resize(args.image_size, args.image_size, p=1),\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":"2025-10-10T16:53:24.250609Z","iopub.execute_input":"2025-10-10T16:53:24.250934Z","iopub.status.idle":"2025-10-10T16:53:24.267459Z","shell.execute_reply.started":"2025-10-10T16:53:24.250913Z","shell.execute_reply":"2025-10-10T16:53:24.266751Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"\ntest_df = pd.read_csv(\"/kaggle/input/siim-isic-melanoma-classification/test.csv\")\n\n# --- handle missing categorical columns ---\n# If 'diagnosis' is missing, create it with default 'unknown'\nif \"diagnosis\" not in test_df.columns:\n    test_df[\"diagnosis\"] = \"unknown\"\n\n# Fill missing site info\ntest_df[\"anatom_site_general_challenge\"] = test_df[\"anatom_site_general_challenge\"].fillna(\"unknown\")\n\n# --- one-hot encode ---\ntest_df = pd.get_dummies(\n    test_df,\n    columns=[\"diagnosis\", \"anatom_site_general_challenge\"],\n    dtype=np.uint8\n)\n\n# --- encode sex as 1/0 ---\ntest_df[\"sex\"] = test_df[\"sex\"].map({\"male\": 1, \"female\": 0}).fillna(0)\n\n# --- handle numeric missing values ---\nif \"age_approx\" in test_df.columns:\n    test_df[\"age_approx\"] = test_df[\"age_approx\"].fillna(test_df[\"age_approx\"].median())\n\n# --- ensure same feature columns as training ---\nexpected_features= [\n    'image_name','sex', 'age_approx',\n    'diagnosis_atypical melanocytic proliferation', 'diagnosis_cafe-au-lait macule',\n    'diagnosis_lentigo NOS', 'diagnosis_lichenoid keratosis',\n    'diagnosis_melanoma', 'diagnosis_nevus', 'diagnosis_seborrheic keratosis',\n    'diagnosis_solar lentigo', 'diagnosis_unknown',\n    'anatom_site_general_challenge_head/neck', 'anatom_site_general_challenge_lower extremity',\n    'anatom_site_general_challenge_oral/genital', 'anatom_site_general_challenge_palms/soles',\n    'anatom_site_general_challenge_torso', 'anatom_site_general_challenge_unknown',\n    'anatom_site_general_challenge_upper extremity'\n]\n\n\nfor col in expected_features:\n    if col not in test_df.columns:\n        test_df[col] = 0\n\n# Ensure column order matches training\ntest_df = test_df[expected_features]\n\n\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-10T17:03:05.086086Z","iopub.execute_input":"2025-10-10T17:03:05.086887Z","iopub.status.idle":"2025-10-10T17:03:05.117115Z","shell.execute_reply.started":"2025-10-10T17:03:05.086861Z","shell.execute_reply":"2025-10-10T17:03:05.116267Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"test_df.head()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-10T17:03:07.756084Z","iopub.execute_input":"2025-10-10T17:03:07.756541Z","iopub.status.idle":"2025-10-10T17:03:07.768681Z","shell.execute_reply.started":"2025-10-10T17:03:07.756519Z","shell.execute_reply":"2025-10-10T17:03:07.767787Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"super_final_predictions = []\ndf_test = test_df\ntest_img_paths = [f\"/kaggle/input/siim-isic-melanoma-classification/jpeg/test/{x}.jpg\" for x in df_test[\"image_name\"].values]\nfor i in range(5):\n    i=0\n    model = MelanomaModel()\n    model = Tez(model)\n    model.load(f\"/kaggle/input/aditya-23bcs10007-melanoma-classification-training/model_f{i}.bin\", weights_only=True)\n    \n    \n    dense_features = [\n        'sex', 'age_approx', 'diagnosis_atypical melanocytic proliferation', 'diagnosis_cafe-au-lait macule', 'diagnosis_lentigo NOS', 'diagnosis_lichenoid keratosis', 'diagnosis_melanoma', 'diagnosis_nevus', 'diagnosis_seborrheic keratosis', 'diagnosis_solar lentigo', 'diagnosis_unknown', 'anatom_site_general_challenge_head/neck', 'anatom_site_general_challenge_lower extremity', 'anatom_site_general_challenge_oral/genital', 'anatom_site_general_challenge_palms/soles', 'anatom_site_general_challenge_torso', 'anatom_site_general_challenge_unknown', 'anatom_site_general_challenge_upper extremity'\n    ]\n    \n    test_dataset = MelanomaDataset(\n        image_paths=test_img_paths,\n        dense_features=df_test[dense_features].values,\n        targets=np.ones(len(test_img_paths)),\n        augmentations=test_aug,\n    )\n    test_predictions = model.predict(test_dataset, batch_size=2*args.batch_size, n_jobs=-1)\n    \n    final_test_predictions = []\n    for preds in tqdm(test_predictions):\n        final_test_predictions.extend(preds.ravel().tolist())\n    \n    # final_test_predictions = [sigmoid(x) * 100 for x in final_test_predictions]\n    super_final_predictions.append(final_test_predictions)\n\nsuper_final_predictions = np.mean(np.column_stack(super_final_predictions), axis=1)\ndf_test[\"target\"] = super_final_predictions\ndf_test = df_test[[\"image_name\", \"target\"]]\ndf_test.to_csv(\"submission.csv\", index=False)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-10T17:06:15.861383Z","iopub.execute_input":"2025-10-10T17:06:15.862409Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"df=pd.read_csv(\"/kaggle/working/submission.csv\")\ndf","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-10T16:54:28.767307Z","iopub.status.idle":"2025-10-10T16:54:28.767615Z","shell.execute_reply.started":"2025-10-10T16:54:28.767464Z","shell.execute_reply":"2025-10-10T16:54:28.767478Z"}},"outputs":[],"execution_count":null}]}