RIFF Rampage
For some reason, a lot of the audio files I use for voice cloning provide an error talking about something related to RIFF or whatever and I don't know why
For some reason, a lot of the audio files I use for voice cloning provide an error talking about something related to RIFF or whatever and I don't know why
Thank you for your report.
I have reviewed the issue.
The problem occurs because the model only accepts audio files in the ".wav" format.
Most of the documentation and voice lists provided by Kyutai use the ".wav" format, so this was the initial assumption. However, this assumption may not be entirely accurate, as there is currently no official documentation specifying which file extensions are supported for voice cloning.
To address this, I have added validator and converter to handle unsupported formats.