|
smarter way to avoid applying encoder key mask
|
alexeib
|
|
6 years ago |
|
caching v3 (cache keys, values, process only last time step) (#241)
|
Alexei Baevski
|
|
6 years ago |
|
Fix buffers in sinusoidal positional embeddings
|
Myle Ott
|
|
6 years ago |
|
Fix flake8
|
Myle Ott
|
|
6 years ago |
|
Bug fixes
|
Myle Ott
|
|
6 years ago |
|
Pass args around to cleanup parameter lists
|
Myle Ott
|
|
6 years ago |
|
Remove Google batching stategy (it's not needed)
|
Myle Ott
|
|
6 years ago |
|
Add Transformer model
|
Myle Ott
|
|
6 years ago |
|
address comments
|
alexeib
|
|
6 years ago |
|
fix optim history
|
alexeib
|
|
6 years ago |
|
Fix LSTM
|
Myle Ott
|
|
6 years ago |
|
Faster fconv generation
|
Myle Ott
|
|
6 years ago |
|
Remove sweep_log prefix from json progress bar
|
Myle Ott
|
|
6 years ago |
|
0.4.0 -> 0.5.0
|
Myle Ott
|
|
6 years ago |
|
Merge internal changes (#163)
|
Myle Ott
|
|
6 years ago |
|
Update dataset code for use by https://github.com/pytorch/translate/pull/62 (#161)
|
theweiho
|
|
6 years ago |
|
Fix old model checkpoints after #151 (fixes #156) (#157)
|
Myle Ott
|
|
6 years ago |
|
Flake8
|
Myle Ott
|
|
6 years ago |
|
Add pretrained embedding support (#151)
|
Sai
|
|
6 years ago |
|
use implicit padding when possible (#152)
|
ngimel
|
|
6 years ago |
|
Update README.md
|
Myle Ott
|
|
6 years ago |
|
Disallow --batch-size in interactive.py
|
Myle Ott
|
|
6 years ago |
|
make interactive mode print out alignment nicely
|
alexeib
|
|
6 years ago |
|
Merge internal changes (#136)
|
Myle Ott
|
|
6 years ago |
|
Merge pull request #134 from hitvoice/master
|
Sergey Edunov
|
|
6 years ago |
|
Update training command for IWSLT14
|
Runqi Yang
|
|
6 years ago |
|
Update training commands
|
Runqi Yang
|
|
6 years ago |
|
Update training commands
|
Runqi Yang
|
|
6 years ago |
|
Merge remote-tracking branch 'upstream/master'
|
杨润琦
|
|
6 years ago |
|
fix typo in data/README (#131)
|
Runqi Yang
|
|
6 years ago |