diff options
| author | Marcin Junczys-Dowmunt <marcinjd@microsoft.com> | 2018-11-26 06:30:40 +0300 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2018-11-26 06:30:40 +0300 |
| commit | a7f7ac5e2f91f3ecd65d7d8aae70c2926e6c82b1 (patch) | |
| tree | 43d6ed019215b757e3ec663b517816dd20536ccd | |
| parent | 067a5b3ba12601f5a4984294bf30153b33505702 (diff) | |
Update README.md
| -rw-r--r-- | training-basics-spm/README.md | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/training-basics-spm/README.md b/training-basics-spm/README.md index 1da24ac..8bd2b84 100644 --- a/training-basics-spm/README.md +++ b/training-basics-spm/README.md @@ -235,7 +235,7 @@ after which BLEU scores for the dev and test set are reported. sacreBLEU/sacrebleu.py -t wmt16/dev -l ro-en < data/newsdev2016.ro.output sacreBLEU/sacrebleu.py -t wmt16 -l ro-en < data/newstest2016.ro.output ``` -You should see results somewhere in the area of: +You should see results somewhere in the area of 36.5 BLEU for the dev set and 35.1 BLEU for the test set. This is actually a bit better than for the BPE version from `marian/examples/training-basics` with the complex preprocessing. ``` BLEU+case.mixed+lang.ro-en+numrefs.1+smooth.exp+test.wmt16/dev+tok.13a+version.1.2.12 = 36.5 67.9/42.7/29.4/20.9 (BP = 1.000 ratio = 1.006 hyp_len = 49816 ref_len = 49526) BLEU+case.mixed+lang.ro-en+numrefs.1+smooth.exp+test.wmt16+tok.13a+version.1.2.12 = 35.1 66.6/41.3/28.0/19.6 (BP = 1.000 ratio = 1.005 hyp_len = 47804 ref_len = 47562) |