Commits · 43b2a8709251b93b1c24d9a7ed1f68e0d7641ab2 · Emmay / WMT19-1.0.14 · GitLab

07 Mar, 2019 1 commit
- reviese transformer_dla hparameter set: relu_dropout=0.1 · 43b2a870
  libei committed Mar 07, 2019
  
  43b2a870 Browse Files
06 Mar, 2019 1 commit
- fix bugs · fb957510
  libei committed Mar 06, 2019
  
  fb957510 Browse Files
01 Mar, 2019 1 commit
- suport share decoder input and softmax embedding · 2e5f98d6
  libei committed Mar 01, 2019
  
  2e5f98d6 Browse Files
21 Feb, 2019 1 commit
- add fast decoding · 8ab393f2
  libei committed Feb 21, 2019
  
  8ab393f2 Browse Files
18 Feb, 2019 3 commits
- remove no use file and import settings · 81667eab
  libei committed Feb 18, 2019
  
  81667eab Browse Files
- fix bugs in transformer_dla · 42676b59
  libei committed Feb 18, 2019
  
  42676b59 Browse Files
- add transformer_dla.py to support Dynamic Layer Aggeration training · a440c641
  libei committed Feb 18, 2019
  
  a440c641 Browse Files
16 Feb, 2019 1 commit
- reviese bugs · 90010cd3
  libei committed Feb 16, 2019
  
  90010cd3 Browse Files
15 Feb, 2019 3 commits
- reset num_hidden_layers into encoder_layers and decoder_layers · fb9ee9e7
```
set tf.random_seed to initial parameters
```
  libei committed Feb 15, 2019
  fb9ee9e7 Browse Files
- remove file · 6097530a
  libei committed Feb 15, 2019
  
  6097530a Browse Files
- WMT19 training system based Tensor2Tensor-1.0.14 · 037d45dc
  libei committed Feb 15, 2019
  
  037d45dc Browse Files