path

pages/Upload_and_Predict.py

@@ -58,17 +58,18 @@ class_names = ['No DR', 'Mild', 'Moderate', 'Severe', 'Proliferative DR']
 
 @st.cache_data
 def load_sample_images_from_csv():
-    dataset = load_dataset("Ci-Dave/test_labels.csv", split="test") 
+    csv_url = "https://huggingface.co/datasets/Ci-Dave/DDR_dataset_train_test/resolve/main/splits/test_labels.csv"
+    df = pd.read_csv(csv_url)
     samples = defaultdict(list)
 
     for i in range(5):
         class_name = class_names[i]
-        class_samples = dataset.filter(lambda example: example['label'] == i)[:5]
-        for example in class_samples:
-            img_path = example['new_path']  # Ensure this column exists or adjust accordingly
-            if os.path.exists(img_path):  # Optional: skip if not local
+        class_samples = df[df['label'] == i].head(5)
+        for _, row in class_samples.iterrows():
+            img_path = row['new_path']
+            if os.path.exists(img_path):  # works only if images are local
                 samples[class_name].append(img_path)
-
+    
     return samples
 
 # Load pretrained model