Questions about the pipeline

[xwkya]

SO: https://arxiv.org/abs/2111.02392
VITS: https://arxiv.org/abs/2106.06103

Can anyone explain the training and inference pipeline?
What are the components that are actually being trained when running the train command, what are the components that are already pretrained?

Thanks

[34j]

Sorry, but I cannot explain in detail now.
It may be helpful to refer to the Wiki (where I described the model structure) or the code.