Hi author, I have a problem about train model?

[funny000]

when i use data of the code hub to pretrained the model, and get 300~1500 files with model_config.json，
so how to use the model? use it replace which model, replace Automodel.from_pretrained or AutoTokenizer.from_pretrained there model path?