{"cells":[{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"cell_type":"code","source":"package_path = '../input/pytorch-image-models/pytorch-image-models-master'\nimport sys; sys.path.append(package_path)\n\nfrom glob import glob\nfrom sklearn.model_selection import GroupKFold, StratifiedKFold\nimport cv2\nfrom skimage import io\nimport torch\nfrom torch import nn\nimport os\nfrom datetime import datetime\nimport time\nimport random\nimport cv2\nimport torchvision\nfrom torchvision import transforms\nimport pandas as pd\nimport numpy as np\nfrom tqdm import tqdm\n\nimport matplotlib.pyplot as plt\nfrom torch.utils.data import Dataset,DataLoader\nfrom torch.utils.data.sampler import SequentialSampler, RandomSampler\nfrom  torch.cuda.amp import autocast, GradScaler\n\nimport sklearn\nimport warnings\nimport joblib\nfrom sklearn.metrics import roc_auc_score, log_loss\nfrom sklearn import metrics\nimport warnings\nimport cv2\nimport pydicom\nimport timm\nfrom scipy.ndimage.interpolation import zoom\nfrom sklearn.metrics import log_loss\n\ntrain = pd.read_csv('../input/cassava-leaf-disease-classification/train.csv')\nsubmission = pd.read_csv('../input/cassava-leaf-disease-classification/sample_submission.csv')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"CFG = {\n    'normalize_mean':[0.42984136, 0.49624753, 0.3129598],\n    'normalize_std':[0.21417203, 0.21910103, 0.19542212],\n    'device': 'cuda:0',\n    'fold_num': 5,\n    'seed': 42,\n    'valid_bs': 32,\n    'num_workers': 4,\n\n    'model_arch': ['tf_efficientnet_b4_ns',\n                   'tf_efficientnet_b4_ns',\n                   'tf_efficientnet_b4_ns',\n                   'tf_efficientnet_b4_ns',\n                   'tf_efficientnet_b4_ns',\n                  ], #### [\"tf_efficientnet_b4_ns\", \"vit_large_patch16_384\", \"deit_base_patch16_384\", 'resnext50_32x4d', 'vit_base_patch16_384']\n    'img_size': 512,\n    'tta': 3,  ####\n    'fold_list':[0],\n    'used_epochs': [\"../input/train3012/tf_efficientnet_b4_ns_3012_fold_0_14\",\n                    \"../input/train3012/tf_efficientnet_b4_ns_3012_fold_1_14\",\n                    \"../input/train3012/tf_efficientnet_b4_ns_3012_fold_2_14\",\n                    \"../input/train3012/tf_efficientnet_b4_ns_3012_fold_3_9\",\n                    \"../input/train3012/tf_efficientnet_b4_ns_3012_fold_4_12\"], ####\n    'weights': [1,1,1,1,1], ####\n} ","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Helper Functions"},{"metadata":{"trusted":true},"cell_type":"code","source":"def 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    torch.backends.cudnn.benchmark = False\n    torch.cuda.manual_seed_all(seed)\n    \ndef get_img(path):\n    im_bgr = cv2.imread(path)\n    im_rgb = im_bgr[:, :, ::-1]\n    return im_rgb","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Dataset"},{"metadata":{"trusted":true},"cell_type":"code","source":"class CassavaDataset(Dataset):\n    def __init__(\n        self, df, data_root, transforms=None, output_label=True\n    ):\n        \n        super().__init__()\n        self.df = df.reset_index(drop=True).copy()\n        self.transforms = transforms\n        self.data_root = data_root\n        self.output_label = output_label\n    \n    def __len__(self):\n        return self.df.shape[0]\n    \n    def __getitem__(self, index: int):\n        # get labels\n        if self.output_label:\n            target = self.df.iloc[index]['label']\n          \n        path = \"{}/{}\".format(self.data_root, self.df.iloc[index]['image_id'])\n        \n        img  = get_img(path)\n        \n        if self.transforms:\n            img = self.transforms(image=img)['image']\n            \n        if self.output_label == True:\n            return img, target\n        else:\n            return img","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"from albumentations import (\n    HorizontalFlip, VerticalFlip, IAAPerspective, ShiftScaleRotate, CLAHE, RandomRotate90,\n    Transpose, ShiftScaleRotate, Blur, OpticalDistortion, GridDistortion, HueSaturationValue,\n    IAAAdditiveGaussianNoise, GaussNoise, MotionBlur, MedianBlur, IAAPiecewiseAffine, RandomResizedCrop,\n    IAASharpen, IAAEmboss, RandomBrightnessContrast, Flip, OneOf, Compose, Normalize, Cutout, CoarseDropout, ShiftScaleRotate, CenterCrop, Resize,\n    RandomCrop\n)\n\nfrom albumentations.pytorch import ToTensorV2\n\ndef get_inference_transforms():\n    return Compose([\n        RandomCrop(512, 512),\n        Transpose(p=0.5),\n        HorizontalFlip(p=0.5),\n        VerticalFlip(p=0.5),\n        HueSaturationValue(hue_shift_limit=10, sat_shift_limit=10, val_shift_limit=10, p=0.5),\n        RandomBrightnessContrast(brightness_limit=(-0.1,0.1), contrast_limit=(-0.1, 0.1), p=0.5),\n        Normalize(mean=CFG[\"normalize_mean\"], std=CFG[\"normalize_std\"], max_pixel_value=255.0, p=1.0),\n        ToTensorV2(p=1.0),\n    ], p=1.)\n\ndef get_inference_transforms_vit384():\n    return Compose([\n        RandomCrop(384, 384),\n        Transpose(p=0.5),\n        HorizontalFlip(p=0.5),\n        VerticalFlip(p=0.5),\n        HueSaturationValue(hue_shift_limit=10, sat_shift_limit=10, val_shift_limit=10, p=0.5),\n        RandomBrightnessContrast(brightness_limit=(-0.1,0.1), contrast_limit=(-0.1, 0.1), p=0.5),\n        Normalize(mean=CFG[\"normalize_mean\"], std=CFG[\"normalize_std\"], max_pixel_value=255.0, p=1.0),\n        ToTensorV2(p=1.0),\n    ], p=1.)\n\n","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Model"},{"metadata":{"trusted":true},"cell_type":"code","source":"class CassvaImgClassifier(nn.Module):\n    def __init__(self, model_arch, n_class, pretrained=False):\n        super().__init__()\n        self.model = timm.create_model(model_arch, pretrained=pretrained, num_classes=5)\n    def forward(self, x):\n        x = self.model(x)\n        return x","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Main Loop"},{"metadata":{"trusted":true},"cell_type":"code","source":"def inference_one_epoch(model, data_loader, device):\n    model.eval()\n    image_preds_all = []\n    pbar = tqdm(enumerate(data_loader), total=len(data_loader))\n    for step, (imgs) in pbar:\n        imgs = imgs.to(device).float()\n        image_preds = model(imgs)   #output = model(input)\n        image_preds_all += [torch.softmax(image_preds, 1).detach().cpu().numpy()]\n    image_preds_all = np.concatenate(image_preds_all, axis=0)\n    return image_preds_all","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"seed_everything(CFG['seed'])\ntst_preds = []\ndevice = torch.device(CFG['device'])\n\ntest = pd.DataFrame()\ntest['image_id'] = list(os.listdir('../input/cassava-leaf-disease-classification/test_images/'))\n\n\nfor i, sub_model in enumerate(CFG['used_epochs']): \n    if \"vit\" not in sub_model:\n        test_ds = CassavaDataset(test, '../input/cassava-leaf-disease-classification/test_images/', transforms=get_inference_transforms(), output_label=False)\n    else:\n        test_ds = CassavaDataset(test, '../input/cassava-leaf-disease-classification/test_images/', transforms=get_inference_transforms_vit384(), output_label=False)\n        \n    tst_loader = torch.utils.data.DataLoader(\n        test_ds, \n        batch_size=CFG['valid_bs'],\n        num_workers=CFG['num_workers'],\n        shuffle=False,\n        pin_memory=True,\n    )\n\n    model = CassvaImgClassifier(CFG['model_arch'][i], train.label.nunique()).to(device)\n    model.load_state_dict(torch.load(sub_model, map_location=CFG['device']))\n    print(sub_model)\n\n    with torch.no_grad():\n        for tta_ in range(CFG['tta']):\n            tst_preds += [CFG['weights'][i]/sum(CFG['weights'])/CFG['tta']*inference_one_epoch(model, tst_loader, device)]\n\ntst_preds = np.sum(tst_preds, axis=0)\n\n\ndel model\ntorch.cuda.empty_cache()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"test['label'] = np.argmax(tst_preds, axis=1)\ntest.to_csv('submission.csv', index=False)","execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat":4,"nbformat_minor":4}