fudan-generative-vision/hallo

Any problem in improving the model's performance on Mandarin Chinese?

Opened this issue · 0 comments

Why it is late to improve the model's performance on Mandarin Chinese?
Is that because wav2vec can not embed Chinese speech well?