Lwasinam

/

voicera

like 22

Hello, sorry i never responded, this project gave me a job and I've been busy with that
Many more strong ar-tts models are available for fine-tune, I suggest you try those,
They'll be much better than this

Oh, my! Glad to hear that! By the way, I'm currently looking for an audio tokenizer that could with speech (and maybe music, but anyway should be exactly 48kHz!). Do you know any good that don't require much data, and I have fine-tune (or maybe even train from scratch)?