Update app.py

app.py

@@ -17,7 +17,7 @@ def generate_response(messages):
     # Apply chat template and encode messages
     encodeds = tokenizer.apply_chat_template(messages, return_tensors="pt")
     # Move inputs to device
-    model_inputs = encodeds.to("cuda")  # Assuming CUDA device is available
+    model_inputs = encodeds 
     # Generate response
     generated_ids = model.generate(model_inputs, max_new_tokens=1000, do_sample=True)
     # Decode the generated response