train fastspeech2

[zcxgsq]

I used the LJSpeech dataset I generated myself to train with FastSpeech2, but there were errors

[eginhard]

Please always share code and error messages as text, not images for accessibility and readability.

The error comes from the soundfile package, not Coqui, so you should look for help there.