About training and finetune.

### Checks - [x] This template is only for research question, not usage problems, feature requests or bug reports. - [x] I have thoroughly reviewed the project documentation and read the related paper(s). - [x] I have searched for existing issues, including closed ones, no similar questions. - [x] I am using English to submit this issue to facilitate community communication. ### Question details Hi, i'm trying to train it from the scratch with my train script, i have some question: 1, How many step did it take to have hearable wav? 2, I have librispeech-100h, is it enough for training process? How many hour needed for training. 3, Does reduce transformer parameters below small config make it impossible to train? Sorry for my bad English. I'm trying to understand the whole project.