teslacool/SCA

Issues

• setups for reproducing IWSLT14 De-en

  #16 opened 5 years ago by BaohaoLiao
  4

• can you help me?

  #21 opened 4 years ago by genbei
  2

• Error training language model

  #22 opened 4 years ago by genbei
  1

• A little problem about IWSLT

  #20 opened 4 years ago by genbei
  1

• Install fairseq error

  #19 opened 4 years ago by genbei
  4

• Preprocessing script for Es-En and He-En

  #17 opened 4 years ago by marshmallow111
  1

• Why do you want to train two language models?

  #18 opened 4 years ago by genbei
  1

• unexpected key when generate

  #15 opened 5 years ago by BaohaoLiao
  3

• Can anybody reproduce the results of paper ?

  #13 opened 5 years ago by KelleyYin
  1

• no improvement with huge data

  #10 opened 5 years ago by nicolabertoldi
  2

• not able to train a LM with SCA

  #11 opened 5 years ago by nicolabertoldi
  1

• use of the language model layers during inference

  #12 opened 5 years ago by nicolabertoldi
  3

• Pull Request into fairseq

  #8 opened 5 years ago by nicolabertoldi
  3

• multilingual engine with SCA

  #9 opened 5 years ago by nicolabertoldi
  3

• Problem of training efficiency

  #7 opened 5 years ago by pemywei
  2

• Why not train the lm-nmt with a baseline for warmup?

  #6 opened 5 years ago by pemywei
  1

• problem with interactive.py

  #5 opened 5 years ago by nicolabertoldi
  12

• which is the meaning of "I shift a sentence twice in decoder input"

  #2 opened 5 years ago by nicolabertoldi
  3

• *inf* value for ppl

  #4 opened 5 years ago by nicolabertoldi
  3

• optimization of SCA

  #3 opened 5 years ago by nicolabertoldi
  2

• Which ‘arch’ should I choose when training the language model

  #1 opened 5 years ago by socaty
  2