cleaning mem

app.py

@@ -224,7 +224,7 @@ def save_model(
             repo_id=repo_name,
             repo_type="model",
         )
-        progress(1.00, desc="Model pushed to Hub")
+        progress(0.95, desc="Model pushed to Hub")
 
     # Get model architecture as string
     import io
@@ -330,7 +330,16 @@ def quantize_and_save(
             public,
             progress,
         )
-
+        # Clean up the model to free memory
+        del quantized_model
+        # Force garbage collection to release memory
+        import gc
+        gc.collect()
+        
+        if torch.cuda.is_available():
+            torch.cuda.empty_cache()
+            
+        progress(1.0, desc="Memory cleaned")
         return final_message
 
     except Exception as e: