ZixuanKe / PyContinual

PyContinual (An Easy and Extendible Framework for Continual Learning)
300 stars 62 forks source link

About Weight #29

Open Hephaestusxg opened 4 months ago

Hephaestusxg commented 4 months ago

Sry, excuse me.I would like to inquire about how to save model weights. In the newsgroup training, is it necessary to train 10 categories first and then proceed with incremental training? Or do you add ten categories to the existing BERT based on datasets_random?