Update app.py

app.py

@@ -16,8 +16,8 @@ else:
 @spaces.GPU
 def generate_response(passage: str, question: str) -> str:
     # Prepare the input text by combining the passage and question
-    user_message = f"Passage: {passage}\nQuestion: {question}"
-    inputs = tokenizer(user_message, return_tensors="pt").to(model.device)
+    message = [{"role": "user", "content": f"Passage: {passage}\nQuestion: {question}"}]
+    inputs = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt")
     
     # Generate text, focusing only on the new tokens added by the model
     outputs = model.generate(inputs.input_ids, max_new_tokens=150)