FLC777/GLAT

Issues

• when import 'torch_imputer' . fail with the error : ninja: build stopped: subcommand failed.

  #21 opened 8 months ago by hhh07
  0

• What is “linear annealing (from 3e − 4 to 1e − 5) ” schedule？

  #16 opened 2 years ago by thunder123321
  3

• train command gives error related to 'torch_imputer'

  #20 opened a year ago by Shailashree-KS
  3

• cannot reproduce results on wmt14 ende distill

  #13 opened 3 years ago by fernando9torres
  1

• About CTC

  #19 opened a year ago by thunder123321
  0

• About Fairseq 1.0.0a0

  #2 opened a year ago by pengwei-iie
  1

• About the CTC implementation

  #1 opened a year ago by ybai-nlp
  2

• About Data Set Selection

  #17 opened a year ago by Yonghao-Li
  1

• about mask_id in "glat_ctc.py" 41 line

  #18 opened a year ago by thunder123321
  1

• Inference with NPD. Issue: AttributeError: 'dict' object has no attribute '_get_node_flag'

  #14 opened 3 years ago by chuanmingliu
  1

• KeyError: 'bleu'

  #15 opened 2 years ago by Andrewlesson
  0

• Which part of your code copies the encoder hidden when glancing?

  #4 opened 3 years ago by zkx06111
  1

• Problem of inference speed

  #12 opened 3 years ago by zhajiahe
  0

• How to reproduce "rerank" result in paper

  #11 opened 3 years ago by zhajiahe
  0

• Training Speed

  #10 opened 3 years ago by hemingkx
  1

• Why was the src_embedding_copy flag code changed?

  #9 opened 3 years ago by SirRob1997
  0

• About the CTC/NPD implementation

  #5 opened 3 years ago by baoalita
  1

• backpropagation

  #8 opened 3 years ago by guoliangq
  0

• size mismatch for encoder.embed_tokens.weight

  #7 opened 3 years ago by Rexbalaeniceps
  0

• The argument "--cpu" is not supported

  #6 opened 3 years ago by Beau-xu
  0

• training fail with the following error

  #3 opened 3 years ago by gyin94
  4