What is the best state of the art text to speech model suitable for converting an English written book to an audiobook (mp3) usable from Hugging Face transformers library? 

I don't want to use reader apps, because their performance is not so good, I can leave it running overnight and prefer better intonation and more natural sounding voice. 

 I haven't tried it yet, but I heard speechnote is good on Linux Unplugged.

I'm not sure id it uses hugging face either. 

 I observed relatively good results with XTTS2 https://huggingface.co/spaces/coqui/xtts
then there are VITS2 and StyleTTS, but I don't have a personal experience with them.