Another one is Soprano-1.1.

It seems like it is being trained by one person, and it is surprisingly natural for such a small model.

I remember when TTS always meant the most robotic, barely comprehensible voices.

https://www.reddit.com/r/LocalLLaMA/comments/1qcusnt/soprano...

https://huggingface.co/ekwek/Soprano-1.1-80M