4 ms·
Is the approach similar to what they did for Riffusion? (On front page a while ago) https://www.riffusion.com/ https://www.riffusion.com/
by lachlan_gray 3y ago
Is the approach similar to what they did for Riffusion? (On front page a while ago)
https://www.riffusion.com/ https://www.riffusion.com/
- woodson 3y agoNo, it’s not similar (other than superficially, as both use diffusion methods in some way). It uses diffusion to generate latent vectors that are entered in a neural audio codec model to produce speech.