Update app.py

app.py

@@ -100,9 +100,7 @@ def reply(message: str, history: list[str]) -> str:
     rag_prompt = preprocess(message)
 
     # Generate a response from the language model
-    response = llm(
-        rag_prompt, max_new_tokens=512, return_full_text=False, device="cuda"
-    )
+    response = llm(rag_prompt, max_new_tokens=512, return_full_text=False)
 
     # Return the generated response
     return response[0]["generated_text"].strip("= ")