Update

backend/model_histo.py

@@ -401,35 +401,39 @@ class BreastCancerClassifier:
             batch = images[i:i + batch_size]
             processed_batch = self.preprocess_image_batch(batch)
             
-        # Calculate batch number before try block to ensure it's available in except block
-        batch_num = i // batch_size + 1
-        
-        try:
-            # Handle different model interface types
-            if hasattr(self.path_foundation, 'signatures') and "serving_default" in self.path_foundation.signatures:
-                # TensorFlow SavedModel format
-                infer = self.path_foundation.signatures["serving_default"]
-                batch_embeddings = infer(tf.constant(processed_batch))
-            elif hasattr(self.path_foundation, 'predict'):
-                # Standard Keras model
-                batch_embeddings = self.path_foundation.predict(processed_batch, verbose=0)
-            else:
-                # Direct callable
-                batch_embeddings = self.path_foundation(processed_batch)
+            # Calculate batch number before try block to ensure it's available in except block
+            batch_num = i // batch_size + 1
             
-            # Handle different output formats
-            if isinstance(batch_embeddings, dict):
-                key = list(batch_embeddings.keys())[0]
-                if hasattr(batch_embeddings[key], 'numpy'):
-                    batch_embeddings = batch_embeddings[key].numpy()
+            try:
+                # Handle different model interface types
+                if hasattr(self.path_foundation, 'signatures') and "serving_default" in self.path_foundation.signatures:
+                    # TensorFlow SavedModel format
+                    infer = self.path_foundation.signatures["serving_default"]
+                    batch_embeddings = infer(tf.constant(processed_batch))
+                elif hasattr(self.path_foundation, 'predict'):
+                    # Standard Keras model
+                    batch_embeddings = self.path_foundation.predict(processed_batch, verbose=0)
                 else:
-                    batch_embeddings = batch_embeddings[key]
-            elif hasattr(batch_embeddings, 'numpy'):
-                batch_embeddings = batch_embeddings.numpy()
-            
-            embeddings.append(batch_embeddings)
-            
-            # Progress reporting
+                    # Direct callable
+                    batch_embeddings = self.path_foundation(processed_batch)
+                
+                # Handle different output formats
+                if isinstance(batch_embeddings, dict):
+                    key = list(batch_embeddings.keys())[0]
+                    if hasattr(batch_embeddings[key], 'numpy'):
+                        batch_embeddings = batch_embeddings[key].numpy()
+                    else:
+                        batch_embeddings = batch_embeddings[key]
+                elif hasattr(batch_embeddings, 'numpy'):
+                    batch_embeddings = batch_embeddings.numpy()
+                
+                embeddings.append(batch_embeddings)
+                
+                # Progress reporting
+                if batch_num % 5 == 0:
+                    print(f"Processed batch {batch_num}/{num_batches}")
+                    
+            except Exception as e:
                 print(f"Error processing batch {batch_num}: {e}")
                 continue