Update README.md

README.md

@@ -61,7 +61,7 @@ with torch.no_grad():
   logits = asr_model(features).logits
 
 predicted_ids = torch.argmax(logits, dim=-1)
-predictions = processor.batch_decode(predicted_ids)
+predictions = processor.batch_decode(predicted_ids, skip_special_tokens=True)
 ```
 
 ## Training Hyperparameters