Some of the audio file may have issue

#2
by neil0306 - opened

Hi,
Thank you for your effort and open source this amazing dataset, but I have some issues and hope you can give me some suggestions.

I ran the dataset downloading script suggested in this blog and did all following steps, but I stuck in python webui_preprocess.py, after the clicked "预处理音频文件", I saw the following errors:
image.png

This error will fail the training process, do you have any idea to solve it?

Thank you for your time, and have a good day!
Ning.

update:
know why my training crashed; it was because I failed to generate BERT representations. The CPU used on my server is an AMD Threadripper, not an Intel CPU, which led to some .so files being incompatible, just like the MLK. My solution was to set an environment variable.
image.png

neil0306 changed discussion status to closed

呢個問題我都遇到過,唔係數據集嘅問題而係佢哋個模型嘅問題。嗰個blog用嘅係fish audio,我以前都試過同你得到相同嘅bug,我後尾改用SoVITS嚟練就成功咗,你可以跟我呢個教程 https://gist.github.com/laubonghaudoi/da84f1c47ff6ca25e618febf4249ff46

最尾得到個效果demo係噉嘅 https://huggingface.co/spaces/laubonghaudoi/zoengjyutgaai_tts

Sign up or log in to comment