Using SoundStorm

[camille-vanhoffelen]

Any plans on using SoundStorm? I gather from the README that bark it is built upon an custom-trained AudioLM. SoundStorm effectively can replace AudioLM's coarse and fine acoustic models, with much faster inference and higher quality audio. lucidrains is also working on an open-source pytorch implementation...