zhjgao/difformer

Issues

• Question about data processing

  #9 opened 3 months ago by hhhhhyyyy
  0

• Thank you for your excellent work，I have some questions about this experimental analysis.

  #5 opened 4 months ago by CCCCCCCCCdut
  3

• embedding

  #7 opened 4 months ago by X-fxx
  1

• BLEU=26.08 on IWSLT14 DE->EN

  #8 opened 4 months ago by chiral-carbon
  2

• ize mismatch for encoder.embed_tokens.weight: copying a param with shape torch.Size([10096, 128]) from checkpoint, the shape in current model is torch.Size([151000, 128]).

  #4 opened 7 months ago by Markkk111
  4

• cannot reproduce results on `iwslt14_de_en`

  #3 opened a year ago by chenxwh
  5

• Question regarding QQP / Wiki-Auto results

  #6 opened 9 months ago by zzbuzzard
  1

• Does the preprocessing step of the fairseq model have an impact on the original fairness of the dataset?

  #1 opened a year ago by xyz321123
  1

• how do you obtain `transformer.pt`?

  #2 opened a year ago by chenxwh
  1