Update app.py

app.py

@@ -59,7 +59,7 @@ def generate_answer_vilt(processor, model, image, question):
     encoding = processor(images=image, text=question, return_tensors="pt")
 
     with torch.no_grad():
-        outputs = model(**encoding)
+        outputs = model(**encoding, max_length=256)
 
     predicted_class_idx = outputs.logits.argmax(-1).item()