Update handler.py

handler.py

@@ -11,8 +11,27 @@ class EndpointHandler:
         # load the model
         self.tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True)
         self.model = AutoModelForCausalLM.from_pretrained(path, device_map="auto", torch_dtype=dtype, trust_remote_code=True)
+        # create inference pipeline
+        self.pipeline = pipeline("text-generation", model=model, tokenizer=tokenizer)
 
     def __call__(self, data: Any) -> List[List[Dict[str, float]]]:
+        inputs = data.pop("inputs", data)
+        parameters = data.pop("parameters", None)
+
+        # pass inputs with all kwargs in data
+        if parameters is not None:
+            prediction = self.pipeline(inputs, **parameters)
+        else:
+            prediction = self.pipeline(inputs)
+        print("---start---")
+        print(prediction)
+        print("---end---")
+
         # ignoring parameters! Default to configs in generation_config.json.
         messages = [{"role": "user", "content": data}]
-        return self.model.chat(self.tokenizer, messages)
+        response = self.model.chat(self.tokenizer, messages)}
+        print("---start chat response---")
+        print(response)
+        print("---end chat response---")
+
+        return [{'generated_text': response}]