Retrieval-based-Voice-Conversion-WebUI
Retrieval-based-Voice-Conversion-WebUI copied to clipboard
Code for Training from Scratch (Creating Base Model)
Hello, first of all, thank you for all contributors which provide us amazing codes.
I would like to ask that if there is a code that trains the model from scratch without going through the pre-trained model. Currently, I am exploring the method of improved models of VITS to apply to the RVC model structure, and I am experiencing a problem that the parameters of the pre-trained model and the changed RVC model do not match.
As it is mentioned in issue 110 before, I was able to found traces of training-from-the-scratch with train_nsf_sim_cache_sid.py
, but not in the repo.
Could you please share the code for base model training? If you kindly share the method, I'm glad to share the methods by writing a guide for future developers.
Thank you.