Closed suzhenghang closed 2 months ago
Sorry, it doesn't support stereo because in my experience, the model that supports stereo may take twice or more gpu resources to train. However, if your gpu resource is sufficient, the training may work with dealing audio tokens with stereo pattern like MusicGen did.
is stereo supported?