- dataset: opus
- model: transformer-align
- pre-processing: normalization + SentencePiece
- download: opus-2020-01-08.zip
- test set translations: opus-2020-01-08.test.txt
- test set scores: opus-2020-01-08.eval.txt
| testset | BLEU | chr-F |
|---|---|---|
| JW300.en.fj | 34.0 | 0.561 |
| Tatoeba.en.fj | 62.5 | 0.781 |
- dataset: opus+bt
- model: transformer-align
- source language(s): en
- target language(s): fj
- model: transformer-align
- pre-processing: normalization + SentencePiece (spm4k,spm4k)
- download: opus+bt-2020-05-23.zip
- test set translations: opus+bt-2020-05-23.test.txt
- test set scores: opus+bt-2020-05-23.eval.txt
- en-fj: QED (23) Tatoeba (32) wikimedia (2)
- en-fj: total size = 57
- unused dev/test data is added to training data
- total size (opus+bt): 370746
-
en-fj: JW300
-
devset = top 2500 lines of JW300.src.shuffled!
-
testset = next 2500 lines of JW300.src.shuffled!
-
remaining lines are added to traindata
| testset | BLEU | chr-F |
|---|---|---|
| JW300.en.fj | 34.6 | 0.570 |