Upload folder using huggingface_hub

script.py

@@ -5,8 +5,7 @@ import os
 from tqdm import tqdm
 import timm
 import torchvision.transforms as T
-import timm,wandb,albumentations as A
-from albumentations.pytorch import ToTensorV2
+# from albumentations.pytorch import ToTensorV2
 
 from PIL import Image
 import torch
@@ -25,40 +24,14 @@ def is_gpu_available():
     """Check if the python package `onnxruntime-gpu` is installed."""
     return torch.cuda.is_available()
 
-VALID_AUG = A.Compose([
-    A.SmallestMaxSize(max_size=SZ + 16, p=1.0),
-    A.CenterCrop(height=SZ, width=SZ, p=1.0),
-    A.Normalize(),
-    ToTensorV2(),
-])
+# VALID_AUG = A.Compose([
+#     A.SmallestMaxSize(max_size=SZ + 16, p=1.0),
+#     A.CenterCrop(height=SZ, width=SZ, p=1.0),
+#     A.Normalize(),
+#     ToTensorV2(),
+# ])
 
 
-class LoadImagesAndLabels(torch.utils.data.Dataset):
-    
-    def __init__(self, df, transforms, mode='train'):
-        self.df = df
-        self.transforms = transforms
-        self.mode = mode
-        
-    def __len__(self): return len(self.df)
-    
-    def __getitem__(self,index):
-        row = self.df.iloc[index]
-        image_path = str(row.filename)
-        images_root_path="/tmp/data/private_testset"
-        image_path = os.path.join(images_root_path, str(row.filename))
-        img = Image.open(image_path).convert("RGB")
-        img = np.array(img) 
-        
-        if self.transforms is not None:
-            img = self.transforms(image=img)['image']
-        
-        if self.mode == 'test':
-            return img
-            
-        label = torch.tensor(labels_class_map[row.binomial_name]).long()
-        return img, label
-
 def get_corn_model(model_name, pretrained=True, **kwargs):
     model = timm.create_model(model_name, pretrained=pretrained, **kwargs)
     model = nn.Sequential(
@@ -82,7 +55,10 @@ class PytorchWorker:
             model_ckpt = torch.load("./NB_EXP_V2_008/vit_base_patch16_224_224_bs32_ep16_lr6e05_wd0.05_mixup_cutmix_CV_0.pth", map_location=self.device)
             model.load_state_dict(model_ckpt)
             return model.to(self.device)
-
+        
+        self.transforms = T.Compose([T.Resize((SZ, SZ)),
+                                            T.ToTensor(),
+                                            T.Normalize([0.5, 0.5, 0.5], [0.5, 0.5, 0.5])])
         self.model = _load_model()
 
     def predict_image(self, image: np.ndarray) -> list():
@@ -90,8 +66,8 @@ class PytorchWorker:
         :param image: Input image as numpy array.
         :return: A list with logits and confidences.
         """
-        image = image.to(self.device)
-        outputs = self.model(image.unsqueeze(dim = 0))
+        image_data = self.transforms(image).unsqueeze(0).to(self.device)
+        outputs = self.model(image_data)
         logits = outputs
         return logits.tolist()
 
@@ -100,10 +76,15 @@ def make_submission(test_metadata, model_path, model_name, output_csv_path="./su
     """Make submission with given """
 
     model = PytorchWorker()
-    data = LoadImagesAndLabels(test_metadata, VALID_AUG, mode='test')
+    data = LoadImagesAndLabels(test_metadata, None, mode='test')
     predictions = []
 
-    for image in data:
+    for _, row in tqdm(test_metadata.iterrows(), total=len(test_metadata)):
+        image_path = os.path.join(images_root_path, row.filename)
+        # image_path = row.filename
+
+        image = Image.open(image_path).convert("RGB")
+
         output = model.predict_image(image)
         string_label_dup = LABELS.get(str(np.argmax(output)), 'Acanthophis antarcticus')
         prediction_class = ORIGINAL_LABELS.get(string_label_dup, 1)

submission.csv

@@ -0,0 +1,4 @@
+observation_id,class_id
+1,419
+2,419
+3,419