Update app.py

app.py

@@ -17,7 +17,7 @@ else:
 def generate_response(passage: str, question: str) -> str:
     # Prepare the input text by combining the passage and question
     message = [f"Passage: {passage}\nQuestion: {question}"]
-    inputs = tokenizer(message, return_tensors='pt', return_token_type_ids=False)
+    inputs = tokenizer(message, return_tensors='pt', return_token_type_ids=False).to('cuda')
 
     response = model.generate(**inputs, max_new_tokens=100)