Commits · master · Simon Will / fairseq · GitLab

May 06, 2019
- Add script for preparing WMT18-Multi30k with BPE · a9bf5d5f
  Simon Will authored May 06, 2019
  
  a9bf5d5f
May 04, 2019
- Bugfix in size of multi-corpus dataset · 657a8836
  Kritika Singh authored May 03, 2019
  657a8836
May 03, 2019
- an option to raise exception if oom happens during fairseq.trainer.train_step (#2) · a2901f98
  Yongqiang Wang authored May 03, 2019
  a2901f98
- added bert large architecture (#698) · f5fbcaaf
  Naman Goyal authored May 03, 2019
  f5fbcaaf
May 02, 2019
- Make learned positional embedding optional · 39264559
  Peng-Jen Chen authored May 02, 2019
  39264559
- Move distributed_init into DistributedFairseqModel (#687) · 34726d56
  Myle Ott authored May 02, 2019
  34726d56
- Validate on all sets based on --save-interval-updates · fb18be00
  Myle Ott authored May 02, 2019
  fb18be00
- Fix inconsistent gradient check · 4a30a5f6
  Myle Ott authored May 02, 2019
  4a30a5f6
- Make CTC work with more encoder-only models · ffc9c8cc
  Kritika Singh authored May 01, 2019
  ffc9c8cc
May 01, 2019
- Make MultiCorpusSampledDataset and IndexedCachedDataset Picklable · e112d501
  Myle Ott authored May 01, 2019
  e112d501
- add ConcatDataset support for XLM · 91c78477
  taineleau authored May 01, 2019
  91c78477
- Support dataset upsampling / relative ratio in PytorchTranslateTask (#494) · ff74ca94
  Ning Dong authored May 01, 2019
  ff74ca94
- Better OOM recovery · da9e493e
  Myle Ott authored May 01, 2019
  da9e493e
- Add default noising argument in WordNoiser initialization (#664) · 37420855
  Ning Dong authored Apr 30, 2019
  37420855
Apr 30, 2019
- addding polynomial lr scheduler (#683) · 9421e978
  Naman Goyal authored Apr 30, 2019
  9421e978
- Merge internal changes · 6b8cb7db
  Myle Ott authored Apr 30, 2019
  6b8cb7db
- Add rm_pt.py helper script for removing checkpoint files · f5e52c19
  Myle Ott authored Apr 30, 2019
  f5e52c19
- Merge internal changes (#654) · d45db804
  Myle Ott authored Apr 29, 2019
  d45db804
- Add more details in error message when sentence length > max tokens (#672) · 89a69616
  Liezl Puzon authored Apr 29, 2019
  89a69616
- Fix upgrade_state_dict for XLM Transformer sentence encoder (#680) · 121877f5
  Liezl Puzon authored Apr 29, 2019
  121877f5
Apr 29, 2019
- Update README.md (#679) · ace8f724
  Myle Ott authored Apr 29, 2019
  ace8f724
- Update comments and citations · 849605a0
  Myle Ott authored Apr 29, 2019
  849605a0
Apr 27, 2019
- Add args and sys.argv to tensorboard (#673) · 257a3b89
  Noe Casas authored Apr 27, 2019
  257a3b89
- Add small comments for MonolingualDataset and TokenBlockDataset · 8bf8399d
  Myle Ott authored Apr 27, 2019
  8bf8399d
Apr 26, 2019
- Passing kwargs in setup_task in fairseq_task (#670) · f701aa8c
  Mohammad Sadegh Rasooli authored Apr 26, 2019
  f701aa8c
Apr 25, 2019
- Fix fairseq unittest timeouts (#667) · 57b6a6db
  Liezl Puzon authored Apr 25, 2019
  57b6a6db
- XLM for NMT: option to only load encoder or decoder (#666) · 5008fd4e
  Liezl Puzon authored Apr 25, 2019
  5008fd4e
- Load a XLM model into transformer encoder / decoder for MT training (#629) · 8da9b1c5
  Liezl Puzon authored Apr 25, 2019
  8da9b1c5
- Add gelu and gelu_fast as possible activation functions (#653) · 8500bdd0
  Liezl Puzon authored Apr 25, 2019
  8500bdd0
- Added link to blog post (#662) · d8d03745
  ankur6ue authored Apr 24, 2019
  d8d03745
- added link to sample stories · 5ecedd69
  Angela Fan authored Apr 24, 2019
  5ecedd69
Apr 24, 2019
- Don't reload best validation loss when using --reset-optimizer · 0020477a
  Myle Ott authored Apr 24, 2019
  0020477a
Apr 22, 2019
- Fix generation with --no-early-stop (#627) · fa52d202
  Max Ryabinin authored Apr 22, 2019
  fa52d202
- reduce memory footprint for average_checkpoints (#647) · d63477e1
  Yongqiang Wang authored Apr 21, 2019
  d63477e1
Apr 17, 2019
- Open BlockPairDataset for MaskedLMData to work (#641) · d2f3007c
  Kartikay Khandelwal authored Apr 17, 2019
  d2f3007c
- Enable custom sampling strategy in MultiCorpusSampledDataset (#639) · 90d6eac2
  Ning Dong authored Apr 16, 2019
  90d6eac2
- Black formatting for multi_corpus_sampled_dataset.py (#638) · 17cef3f6
  Ning Dong authored Apr 16, 2019
  17cef3f6
Apr 16, 2019
- Open Source MLM Implementation in Fairseq (#635) · 8776928c
  Kartikay Khandelwal authored Apr 16, 2019
  8776928c
Apr 15, 2019
- Better distributed init · 303b95ce
  Myle Ott authored Apr 15, 2019
  303b95ce
- fix checkpoint timer (#634) · de8aeab5
  freewym authored Apr 15, 2019
  de8aeab5