Update handler.py

handler.py

@@ -26,9 +26,10 @@ class EndpointHandler():
         """
         
         prompts = [f"<human>: {prompt}\n<bot>:" for prompt in data["inputs"]]
-        
-        
-        inputs = self.tokenizer(prompts, padding=True, return_tensors='pt').to(model.device)
+
+        print("prompts")
+        inputs = self.tokenizer(prompts, padding=True, return_tensors='pt').to(self.model.device)
+        raise ValueError(inputs)
         input_length = inputs.input_ids.shape[1]
         outputs = self.model.generate(
             **inputs, max_new_tokens=512, do_sample=True, temperature=0.7, top_p=0.7, top_k=50