funasr/bin/compute_audio_cmvn.py
@@ -73,8 +73,9 @@ iter_stop = int(kwargs.get("scale", -1.0) * len(dataloader_train)) log_step = iter_stop // 100 if batch_idx % log_step == 0: print(f"prcessed: {batch_idx}/{iter_stop}") logging.info(f"prcessed: {batch_idx}/{iter_stop}") if batch_idx >= iter_stop and iter_stop > 0.0: logging.info(f"prcessed: {iter_stop}/{iter_stop}") break fbank = batch["speech"].numpy()[0, :, :]
