lemmonation/jm-nat

Issues

• why 'Index file doesn\'t match expected format. Make sure that --dataset-impl is configured properly.'? I preproessed the data using fairseq-preprocess

  #5 opened 3 years ago by CandisXue
  1

• 为什么genertate出来的是包含【mask】的不完整句子？？

  #6 opened 3 years ago by CandisXue
  0

• Could you please provide the download link for the jointly masked pre-trained model in wmt？

  #3 opened 4 years ago by yangkexin
  2

• Question about the criterion in the training script.

  #4 opened 4 years ago by chynphh
  0

• 报一个bug：source中的mask_index和为词表中第一个词的index

  #2 opened 5 years ago by hanyaqian
  3

• Generate distilled target samples by the autoregressive model, i.e., set --gen-subset train while decoding.

  #1 opened 5 years ago by hanyaqian
  2