Update handler.py

handler.py

@@ -12,10 +12,10 @@ class EndpointHandler():
             path,
             torch_dtype=torch.bfloat16,
             device_map="cuda:0",  # or "mps" if on Apple Silicon
-            attn_implementation="flash_attention_2" if is_flash_attn_2_available() else None,
+            attn_implementation="flash_attention_2" # if is_flash_attn_2_available() else None,
         ).eval()
         self.processor = ColQwen2Processor.from_pretrained(path)
-        self.model = torch.compile(self.model)
+        # self.model = torch.compile(self.model)
         print(f"Model and processor loaded {'with' if is_flash_attn_2_available() else 'without'} FA2")
 
     def __call__(self, data: Dict[str, Any]) -> Dict[str, Any]: