Commits · v0.6.1 · Simon Will / fairseq · GitLab

Feb 09, 2019
- Add fairseq to PyPI (#495) · fbd4cef9
  Myle Ott authored Feb 08, 2019
  View commits for tag v0.6.1 v0.6.1
  fbd4cef9
Feb 07, 2019
- stitch preprocessing pipeline · cea0e4b9
  Ruty Rinott authored Feb 06, 2019
  cea0e4b9
Feb 06, 2019
- Add CheckpointManager to keep avg checkpoint weights in memory to reduce disk... · c49c292c
  Wei Ho authored Feb 06, 2019
  c49c292c
Feb 05, 2019
- Add standalone binaries · 829bd8ce
  Myle Ott authored Feb 05, 2019
  829bd8ce
Feb 01, 2019
- Support custom Dictionary implementations in 'preprocess.py' (#448) · bbb4120b
  Davide Caroselli authored Feb 01, 2019
  bbb4120b
Jan 30, 2019
- Do distributed init after data loading · ec6f8ef9
  Myle Ott authored Jan 30, 2019
  ec6f8ef9
- Add --input option to interactive.py to support reading from file · 3dce7c9f
  Myle Ott authored Jan 30, 2019
  3dce7c9f
- Merge internal changes (#483) · 42be3ebd
  Myle Ott authored Jan 30, 2019
  42be3ebd
Jan 29, 2019
- make dictionary class as input for fairseq preprocess functions (#482) · 66ce2175
  Jingfei Du authored Jan 29, 2019
  66ce2175
Jan 25, 2019
- Add code for "Pay Less Attention with Lightweight and Dynamic Convolutions" (#473) · b41c74dc
  Myle Ott authored Jan 25, 2019
  b41c74dc
- refactor AdversarialTrainer factor out helper functions · bc8ae449
  Xian Li authored Jan 25, 2019
  bc8ae449
- Adafactor Optimizer (#472) · 3e67386b
  Lucio Dery authored Jan 25, 2019
  3e67386b
- Only use c10d distributed primitives · 7e0d222c
  Myle Ott authored Jan 25, 2019
  7e0d222c
Jan 24, 2019
- LSTM improvements (fixes #414) · 9196c0b6
  Myle Ott authored Jan 24, 2019
  9196c0b6
- Print model and number of trained params · d0ebcec4
  Myle Ott authored Jan 24, 2019
  d0ebcec4
- Enforce UTF-8 when open() text files (#460) · 38f1dee9
  Davide Caroselli authored Jan 24, 2019
  38f1dee9
- Better error message for improperly formatted dictionaries · ef3e6ab5
  Myle Ott authored Jan 24, 2019
  ef3e6ab5
- change f"{args}" to "{}".format(args) (#467) · 8eb49c84
  vufg authored Jan 24, 2019
  8eb49c84
- Fix iteration bug in GroupedIterator. Correct sent size filter. (#455) · 37b9c235
  frankang authored Jan 24, 2019
  37b9c235
Jan 17, 2019
- Fix stories generation · d259ffa9
  Myle Ott authored Jan 16, 2019
  d259ffa9
- Fix initial learning rate (#453) · 2210fa71
  Myle Ott authored Jan 16, 2019
  2210fa71
Jan 16, 2019
- FIX: '--user-dir' on multi-gpu (#449) · 7853818c
  Davide Caroselli authored Jan 16, 2019
  7853818c
- Add --checkpoint-upper-bound to average_checkpoints.py (#452) · bdec179b
  Myle Ott authored Jan 16, 2019
  bdec179b
- optimizations for token_block_dataset · d1dc66d9
  Ruty Rinott authored Jan 15, 2019
  d1dc66d9
Jan 15, 2019
- Fixed wrong help message shown on '--help' (#446) · cefe3f8a
  Davide Caroselli authored Jan 15, 2019
  cefe3f8a
- '--user-dir' documentation (correct) (#447) · ebaf8c50
  Davide Caroselli authored Jan 15, 2019
  ebaf8c50
Jan 14, 2019
- New command line option '--user-dir' (#440) · b15f5f53
  Davide Caroselli authored Jan 14, 2019
  b15f5f53
- Fixes (#442) · d9284ee7
  Huihui Fan authored Jan 14, 2019
  d9284ee7
Jan 10, 2019
- Make error message for trying to train after make_generation_fast work correctly · 315fa5cb
  Wei Ho authored Jan 09, 2019
  315fa5cb
Jan 09, 2019
- Misc fixes · 4b1f4788
  Myle Ott authored Jan 09, 2019
  4b1f4788
- Fix broken link in README.md (#436) · 73876ce3
  Art Matsak authored Jan 09, 2019
  73876ce3
Jan 07, 2019
- Update docs for --lazy-load and torch.distributed.launch · 14bd9c62
  Myle Ott authored Jan 07, 2019
  14bd9c62
Jan 05, 2019
- Cleanup more files · 7d66726b
  Myle Ott authored Jan 05, 2019
  
  7d66726b
- rm fb_train.py (#432) · df4d566d
  Myle Ott authored Jan 05, 2019
  
  df4d566d
- Merge internal changes (#283) · 7633129b
  Myle Ott authored Jan 04, 2019
  7633129b
Dec 28, 2018
- Make multiprocessing_train.py work with multi-node setups · 0cb87130
  Myle Ott authored Dec 28, 2018
  0cb87130
- Fix resuming from FP16 checkpoints (#424) · 58dd1862
  Myle Ott authored Dec 27, 2018
  58dd1862
- Fix backtranslation dataset on IndexedCachedDataset (#410) · 31a43973
  Paul Michel authored Dec 27, 2018
  31a43973
Dec 26, 2018
- Merge internal changes (#422) · 8ce6499d
  Myle Ott authored Dec 26, 2018
  8ce6499d
- Add option to disable positional embeddings in TransformerModel (#421) · 19c17b74
  Emanuele Bugliarello authored Dec 26, 2018
  19c17b74