Questions about vits fine tuning

[kbuwel]

I've been fine-tuning vits for a year.
I'm asking because the results aren't good.
I've received a lot of advice about vits.
The results were very disappointing.
I've heard that fine-tuning is possible with a 30-minute dataset.
What is the recommended amount of voice data time in coqui-ai?
I know it's the learning_rate learning speed.
I did it with 3060 and 12-e4.
What is the recommended value?