Update handler.py

handler.py

@@ -28,7 +28,7 @@ class EndpointHandler:
         logging.warn("---end---")
 
         # ignoring parameters! Default to configs in generation_config.json.
-        messages = [{"role": "user", "content": data}]
+        messages = [{"role": "user", "content": data.pop("inputs", data)}]
         response = self.model.chat(self.tokenizer, messages)
         logging.warn("---start chat response---")
         logging.warn(response)