Update app.py

app.py

@@ -79,7 +79,7 @@ def respond(message, max_tokens, temperature, top_p):
             return violation_message
 
     assistant_prompt = f"{system_message}\nUser: {message}\nAssistant:"
-    inputs = tokenizer(assistant_prompt, return_tensors="pt")
+    inputs = tokenizer(assistant_prompt, return_tensors="pt").to(model.device)
     outputs = model.generate(
         **inputs,
         max_new_tokens=max_tokens,