how to train a better contentvec model applied on svc task

[li-henan]

dear author, thanks for your excellent code.Could I ask the detailed steps of training a better contentvec model applied on svc task? Is the script similar to the photo below（load pretrained checkpoint_best_legacy.pt and my dataset, train as the photo below）？Thanks very much