Update app.py

app.py

@@ -17,7 +17,7 @@ def respond_to_input(user_input):
     inputs = tokenizer(prompt, return_tensors="pt")
     max_length = 256
     outputs = model.generate(
-        inputs.input_ids.cuda(),
+        inputs.input_ids,
         max_length=max_length,
         eos_token_id=tokenizer.eos_token_id,
         do_sample=True,