Update app.py

app.py

@@ -12,7 +12,7 @@ def predict(text):
     input_ids = tokenizer(text, max_length=156, return_tensors="pt", padding="max_length", truncation=True).input_ids.to(device)
     outputs = model.generate(input_ids=input_ids, max_new_tokens=528, do_sample=True)
     pred = tokenizer.batch_decode(outputs.detach().cpu().numpy(), skip_special_tokens=True)[0]
-    return pred
+    return pred[len('<extra_id_0>'):]
 
 title = 'VTA-QA Demo'
 article = "Loaded model from https://huggingface.co/kietnt0603/randeng-t5-vta-qa-lora"