opus-2020-07-19.yml

release: eng-iir/opus-2020-07-19.zip
release-date: 2020-07-19
dataset-name: opus
modeltype: transformer
pre-processing: normalization + SentencePiece (spm32k,spm32k)
subwords:
   source: spm32k
   target: spm32k
subword-models:
   source: source.spm
   target: target.spm
source-languages:
   - eng
target-languages:
   - asm
   - awa
   - ben
   - bho
   - gom
   - guj
   - hif
   - hin
   - jdt
   - kur
   - mai
   - mar
   - npi
   - ori
   - oss
   - pan
   - pes
   - pnb
   - pus
   - rom
   - san
   - sin
   - snd
   - tgk
   - tly
   - urd
   - zza
use-target-labels:
   - ">>asm<<"
   - ">>awa<<"
   - ">>ben<<"
   - ">>bho<<"
   - ">>gom<<"
   - ">>guj<<"
   - ">>hif_Latn<<"
   - ">>hin<<"
   - ">>jdt_Cyrl<<"
   - ">>kur_Arab<<"
   - ">>kur_Latn<<"
   - ">>mai<<"
   - ">>mar<<"
   - ">>npi<<"
   - ">>ori<<"
   - ">>oss<<"
   - ">>pan_Guru<<"
   - ">>pes<<"
   - ">>pes_Latn<<"
   - ">>pes_Thaa<<"
   - ">>pnb<<"
   - ">>pus<<"
   - ">>rom<<"
   - ">>san_Deva<<"
   - ">>sin<<"
   - ">>snd_Arab<<"
   - ">>tgk_Cyrl<<"
   - ">>tly_Latn<<"
   - ">>urd<<"
   - ">>zza<<"
test-data:
   Tatoeba-test.eng-asm: 117/530
   Tatoeba-test.eng-awa: 279/1087
   Tatoeba-test.eng-ben: 2500/10942
   Tatoeba-test.eng-bho: 42/224
   Tatoeba-test.eng-fas: 3768/27610
   Tatoeba-test.eng-guj: 154/680
   Tatoeba-test.eng-hif: 36/177
   Tatoeba-test.eng-hin: 5000/31361
   Tatoeba-test.eng-jdt: 2/4
   Tatoeba-test.eng-kok: 1/5
   Tatoeba-test.eng-kur: 296/1410
   Tatoeba-test.eng-lah: 32/180
   Tatoeba-test.eng-mai: 8/16
   Tatoeba-test.eng-mar: 10000/47665
   Tatoeba-test.eng-multi: 10000/55975
   Tatoeba-test.eng-nep: 115/398
   Tatoeba-test.eng-ori: 33/190
   Tatoeba-test.eng-oss: 57/260
   Tatoeba-test.eng-pan: 87/570
   Tatoeba-test.eng-pus: 31/273
   Tatoeba-test.eng-rom: 671/4283
   Tatoeba-test.eng-san: 144/366
   Tatoeba-test.eng-sin: 45/189
   Tatoeba-test.eng-snd: 4/17
   Tatoeba-test.eng-tgk: 29/147
   Tatoeba-test.eng-tly: 46/109
   Tatoeba-test.eng-urd: 1663/12106
   Tatoeba-test.eng-zza: 523/2388
BLEU-scores:
   Tatoeba-test.eng-asm: 1.7
   Tatoeba-test.eng-awa: 0.2
   Tatoeba-test.eng-ben: 14.9
   Tatoeba-test.eng-bho: 0.4
   Tatoeba-test.eng-fas: 3.3
   Tatoeba-test.eng-guj: 17.3
   Tatoeba-test.eng-hif: 1.0
   Tatoeba-test.eng-hin: 16.5
   Tatoeba-test.eng-jdt: 9.5
   Tatoeba-test.eng-kok: 8.1
   Tatoeba-test.eng-kur: 3.1
   Tatoeba-test.eng-lah: 0.9
   Tatoeba-test.eng-mai: 9.8
   Tatoeba-test.eng-mar: 19.3
   Tatoeba-test.eng-multi: 13.1
   Tatoeba-test.eng-nep: 0.9
   Tatoeba-test.eng-ori: 1.3
   Tatoeba-test.eng-oss: 2.2
   Tatoeba-test.eng-pan: 8.5
   Tatoeba-test.eng-pus: 1.0
   Tatoeba-test.eng-rom: 1.3
   Tatoeba-test.eng-san: 1.0
   Tatoeba-test.eng-sin: 10.8
   Tatoeba-test.eng-snd: 2.8
   Tatoeba-test.eng-tgk: 6.9
   Tatoeba-test.eng-tly: 0.6
   Tatoeba-test.eng-urd: 11.8
   Tatoeba-test.eng-zza: 0.5
chr-F-scores:
   Tatoeba-test.eng-asm: 0.194
   Tatoeba-test.eng-awa: 0.031
   Tatoeba-test.eng-ben: 0.447
   Tatoeba-test.eng-bho: 0.110
   Tatoeba-test.eng-fas: 0.219
   Tatoeba-test.eng-guj: 0.366
   Tatoeba-test.eng-hif: 0.079
   Tatoeba-test.eng-hin: 0.451
   Tatoeba-test.eng-jdt: 0.136
   Tatoeba-test.eng-kok: 0.040
   Tatoeba-test.eng-kur: 0.123
   Tatoeba-test.eng-lah: 0.036
   Tatoeba-test.eng-mai: 0.374
   Tatoeba-test.eng-mar: 0.456
   Tatoeba-test.eng-multi: 0.380
   Tatoeba-test.eng-nep: 0.037
   Tatoeba-test.eng-ori: 0.190
   Tatoeba-test.eng-oss: 0.194
   Tatoeba-test.eng-pan: 0.337
   Tatoeba-test.eng-pus: 0.123
   Tatoeba-test.eng-rom: 0.221
   Tatoeba-test.eng-san: 0.106
   Tatoeba-test.eng-sin: 0.382
   Tatoeba-test.eng-snd: 0.205
   Tatoeba-test.eng-tgk: 0.324
   Tatoeba-test.eng-tly: 0.024
   Tatoeba-test.eng-urd: 0.396
   Tatoeba-test.eng-zza: 0.033