Update handler.py

handler.py

@@ -95,7 +95,7 @@ class EndpointHandler:
 
             with torch.no_grad(), torch.autocast("cuda"):
                 generated_ids = self.model.generate(
-                    **processed_inputs, **generation_config
+                    **processed_inputs, generation_config=generation_config
                 )
                 generated_texts = self.processor.batch_decode(
                     generated_ids,