Comment by codingmoh
Hey, this is really cool! Curious how good the multi-language support is. Also - pretty wild that you trained the whole thing yourselves, especially without prior experience in speech models.
Might actually be helpful for others if you ever feel like documenting how you got started and what the process looked like. I’ve never worked with TTS models myself, and honestly wouldn’t know where to begin. Either way, awesome work. Big respect.
Thank you so much for the kind words :) We only support English at the moment, hopefully can do more languages in the future. We are planning to release a technical report on some of the details, so stay tuned for that!