能否返回音素级的时间戳

Open toughhou opened this issue 8 months ago • 1 comments

希望模型能返回音素级的时间戳

May 24 '25 15:05 toughhou

不能，可以后续做asr试试

May 26 '25 03:05 aluminumbox

This issue is stale because it has been open for 30 days with no activity.

Jun 26 '25 02:06 github-actions[bot]

This issue was closed because it has been inactive for 14 days since being marked as stale.

Jul 10 '25 02:07 github-actions[bot]

不能，可以后续做asr试试

后续尝试使用Paraformer等ASR转后很多词汇或符合和原文本是不一样的，要这样去正确匹配时间戳几乎不太可能，期望CosyVoice后续能支持.. 现在看到Sambert是支持返回时间戳，但是音色效果又差很多..

Aug 27 '25 03:08 HuangJT