Add `length_penalty`

app.py

@@ -11,6 +11,7 @@ def predict(text):
         input_ids=tokenized_inputs["input_ids"],
         attention_mask=tokenized_inputs["attention_mask"],
         max_new_tokens=256,
+        length_penalty=0.8,
         num_beams=2,
         do_sample=True
     )