Update app.py

app.py

@@ -23,6 +23,13 @@ outputs = pipeline(
     messages[1]["content"],  # Use only the user message here
     max_new_tokens=256,  # Limit the number of tokens generated
 )
+pipeline = transformers.pipeline(
+    "text-generation",
+    model=model_id,
+    model_kwargs={"torch_dtype": torch.bfloat16},
+    device=-1,  # Use CPU (avoid device_map)
+)
+
 
 # Print the generated text from the output
 print(outputs[0]["generated_text"])