diff options
| author | Marcin Junczys-Dowmunt <junczys@amu.edu.pl> | 2018-03-14 03:29:46 +0300 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2018-03-14 03:29:46 +0300 |
| commit | 2fdb213e0e964e88e6ce70ae1941e73494f29e1d (patch) | |
| tree | 2dedcf382ea34881f88f396adec2791186f81eec | |
| parent | e468216f454ec14857b5b248eefeb3a39fef8dfe (diff) | |
Update README.mdwmt2017
| -rw-r--r-- | wmt2017-transformer/README.md | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/wmt2017-transformer/README.md b/wmt2017-transformer/README.md index 88122d2..2e9b873 100644 --- a/wmt2017-transformer/README.md +++ b/wmt2017-transformer/README.md @@ -67,6 +67,6 @@ Running the complete script from start to end shoud results in numbers similar t System | test2014 | test2015 | test2016(valid) | test2017 | |------|----------|----------|-----------------|----------| |Edinburgh WMT17| -- | -- | 36.20 |28.30| -|Example | 29.08 | 31.04 | 36.80 | 29.50| +|This example | 29.08 | 31.04 | 36.80 | 29.50| Improving on Edinburgh's system submission by 1.2 BLEU. Training all components for more than 8 epochs is likely to improve results. So could increasing model dimensions, but the could require careful hyperparamter tuning, especially dropout regularization. |