Hi,
Does anyone know what is the correct sequence to train a new language from scratch with new phoneset? is it tcsinger ->sad->sdlm?
and if I want to finetune English chinese pretrained model to another language, do I need to train also every submodel?