Update new_approach/spa_ensemble.py

new_approach/spa_ensemble.py

@@ -12,10 +12,11 @@ from sklearn.preprocessing import StandardScaler
 import torchvision.transforms as transforms
 import open_clip
 import joblib
-from huggingface_hub import hf_hub_download  # <--- REQUIRED for downloading weights
+from huggingface_hub import hf_hub_download
 
 # --- CONFIGURATION ---
 CONFIDENCE_THRESHOLD = 0.99
+# The list directory remains in the root of the Space
 LIST_DIR = Path("list") 
 
 # ==============================================================================
@@ -43,6 +44,8 @@ class FeatureExtractor:
     @staticmethod
     def extract_texture_features(img):
         img_np = np.array(img); gray = cv2.cvtColor(img_np, cv2.COLOR_RGB2GRAY); features = {}
+        # Optimization: Canny/Sobel can be slow on huge images.
+        # We assume image is resized in extract_all_features
         edges = cv2.Canny(gray, 50, 150)
         gx, gy = cv2.Sobel(gray, cv2.CV_64F, 1, 0, ksize=3), cv2.Sobel(gray, cv2.CV_64F, 0, 1, ksize=3)
         features.update({
@@ -74,6 +77,7 @@ class FeatureExtractor:
     @staticmethod
     def extract_all_features(img):
         img = img.convert('RGB')
+        # OPTIMIZATION: Resize for Handcrafted Features to speed up Canny/Sobel
         max_size = 1024
         if max(img.size) > max_size:
              img.thumbnail((max_size, max_size), Image.Resampling.LANCZOS)
@@ -170,7 +174,7 @@ class ModelManager:
         print(f"Initializing SPA Ensemble on {self.device}...")
         
         # --- CONFIG: YOUR MODEL REPO ID ---
-        # ⚠️ REPLACE 'YourUsername' with your actual HuggingFace username!
+        # Using the correct repo ID provided
         self.REPO_ID = "FrAnKu34t23/ensemble_models_plant" 
         
         self.class_to_idx, self.idx_to_class, self.id_to_name = self.load_class_info()
@@ -192,11 +196,13 @@ class ModelManager:
         # 3. Download & Load Scaler
         print("SPA Ensemble: Downloading Scaler...")
         try:
+            # Now fetching scaler.joblib from the Model Repo
             scaler_path = hf_hub_download(repo_id=self.REPO_ID, filename="scaler.joblib")
             self.scaler = joblib.load(scaler_path)
-            print("✓ Scaler loaded.")
+            print("✓ Scaler downloaded and loaded.")
         except Exception as e:
-            print(f"Warning: Could not download scaler: {e}. Using dummy.")
+            print(f"Warning: Could not download scaler from {self.REPO_ID}: {e}.")
+            print("Using dummy scaler (predictions may be inaccurate).")
             self.scaler = StandardScaler()
             self.scaler.fit(np.zeros((1, 49)))