microsoft/MASS

Issues

• who can share the model with me

  #183 opened a year ago by icegomic
  0

• This repo is missing important files

  #182 opened 2 years ago by microsoft-github-policy-service
  2

• per-training BlUE always 0.0000

  #158 opened 4 years ago by Nanamumuhan
  1

• supNMT pre-train problem with multi gpus

  #177 opened 3 years ago by Andrewlesson
  1

• Where is the file "fairseq-preprocess"

  #180 opened 2 years ago by xcssgzs
  0

• Does mass implement the translate method？

  #179 opened 3 years ago by mqy9787
  1

• how can you get the data for MASS supNMT?

  #178 opened 3 years ago by xiuzhilu
  0

• How does MASS supervised machine translation perform preprocessing?

  #176 opened 3 years ago by IdaBetsy
  0

• Mass_unsup has no problem on a single GPU, and errors are reported on multiple GPUs

  #175 opened 3 years ago by MayDomine
  0

• Fail to Reproduce the Result of UnsupMT-EnDe

  #141 opened 4 years ago by LibertFan
  3

• invalid task choice

  #174 opened 4 years ago by yaocheng95
  0

• Translation results on Zh-En pre-trained model

  #173 opened 4 years ago by riddlehk
  0

• How to create dictionary dict.lg.txt in MASS supNMT

  #172 opened 4 years ago by Ashmari
  0

• Question about data processing in Unsupervised NMT

  #171 opened 4 years ago by ElliottYan
  0

• Questions for SupNMT

  #170 opened 4 years ago by MSWon
  0

• Predictions on XSUM?

  #169 opened 4 years ago by danyaljj
  0

• Question towards the Pre-trained weight for the Neural Machine Translation under supNMT

  #168 opened 4 years ago by MichaelCaohn
  0

• Confusion about the amount of monolingual data used in the experiments

  #160 opened 4 years ago by cbaziotis
  1

• Incorrect dictionary format

  #166 opened 4 years ago by abdullahkhilji
  3

• How to create dictionary dict.lg.txt

  #167 opened 4 years ago by abdullahkhilji
  0

• Will BERT+transformer-decoder better than tensor2tensor for text-generation?

  #155 opened 4 years ago by guotong1988
  1

• Do two direction data for parallel data is necessary?

  #165 opened 4 years ago by SefaZeng
  0

• Experiment setting for Multilingual pretraining and Supervised NMT

  #150 opened 4 years ago by renziver
  1

• .

  #151 opened 4 years ago by masonreznov
  0

• Problem while running Supervised NMT

  #162 opened 4 years ago by RachitBansal
  2

• Quick question about "masked_block_start"

  #163 opened 4 years ago by Derekkk
  1

• Confusion regarding data

  #164 opened 4 years ago by kr-sundaram
  1

• Question of pretraining text-generation task, it seems that pretraining is not work for a small model?

  #161 opened 4 years ago by guotong1988
  3

• Questions on Table 3 of the MASS paper?

  #149 opened 4 years ago by Epsilon-Lee
  2

• What is the difference between pretrain-tensor2tensor and MASS?

  #156 opened 4 years ago by guotong1988
  5

• Unable to load Zh-En Pre-trained Model for fine-tuning

  #153 opened 4 years ago by riddlehk
  1

• How to reload checkpoint for UNMT?

  #154 opened 4 years ago by him-mah10
  8

• In text summarization

  #138 opened 4 years ago by KawhiZhao
  2

• is it typo in README (Fine-tuning (CNN / Daily Mail))?

  #144 opened 4 years ago by SoyChae
  1

• data/processed/en-fr/train.en.pth valid.en.pth. test.en.pth..........

  #157 opened 4 years ago by bozhenhhu
  0

• Fine Tuning with MBART pretrained model

  #148 opened 4 years ago by masonreznov
  0

• How to use MASS for Style Transfer?

  #143 opened 4 years ago by him-mah10
  2

• error in running training script for pre-training multiple monolingual data

  #146 opened 4 years ago by renziver
  0

• Outputs of summarization task

  #140 opened 4 years ago by shahbazsyed
  1

• how to preprocess data and use the finetuned model ?

  #147 opened 4 years ago by 15091444119
  1

• unable to set a proper batch_size in MASS-supNMT pretraining

  #142 opened 4 years ago by vikrant97
  2

• MASS-supNMT: Args say word_mask_keep_rand but code is word_mask_rand_keep

  #139 opened 5 years ago by JasonVann
  0

• In sup-NMT en2zh: the source sentence's unk data was predicted to other words, how can I make model keep the unk value so I can use post_process_prediction to replace_unk?

  #132 opened 5 years ago by rejae
  1

• How to prepare data pipeline and utilize the provided BPE codes?

  #135 opened 5 years ago by liuchongming74
  3

• How to implement fine-tuned model by myself?

  #137 opened 5 years ago by kaneyxx
  1

• Hyperparameter for low-resource experiment

  #136 opened 5 years ago by yukiyakiZ
  0

• How many Gpu needed for Text Summarization of CNNDM fine-tuning ?

  #134 opened 5 years ago by fseasy
  2

• Unable to fine-tune pre-trained model

  #133 opened 5 years ago by Valahaar
  0

• sup-NMT zh-en data processing.

  #131 opened 5 years ago by rejae
  1

• Can I extend zh-en dict to get rid of <UNK> problem in finetuning zh-en sup-NMT ?

  #130 opened 5 years ago by rejae
  2