Update app.py

app.py

@@ -35,7 +35,7 @@ def process_inputs(image, audio):
     with torch.inference_mode():
         outputs = model.generate(
             **input_ids,
-            max_new_tokens=max_tokens,
+            max_new_tokens=256,
             disable_compile=True
         )
     text = processor.batch_decode(