elephantmipt/bert-distillation

Issues

• Is it possible to distill an ensemble network based on Roberta?

  #8 opened 3 years ago by boscoj2008
  0

• Key Error :masked_lm_labels

  #7 opened 4 years ago by hbatra21
  2

• for NER downstream task, the data format of train.csv/test.csv should be different , right?

  #6 opened 5 years ago by marcusau
  6