Transformer + ASR Pretrain + SpecAug

Reported on 6 benchmarks across 1 task · 1 paper · 4 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing6 results

Speech-to-Text Translationonlibri-trans
Case-insensitive sacreBLEU· 2020-12-18
17.2
SOTA
NeurST: Neural Speech Translation Toolkit arXiv:2012.10018
Speech-to-Text Translationonlibri-trans
Case-insensitive tokenized BLEU· 2020-12-18
18.7
SOTA
NeurST: Neural Speech Translation Toolkit arXiv:2012.10018
Speech-to-Text Translationonlibri-trans
Case-sensitive sacreBLEU· 2020-12-18
16.3
SOTA
NeurST: Neural Speech Translation Toolkit arXiv:2012.10018
Speech-to-Text Translationonlibri-trans
Case-sensitive tokenized BLEU· 2020-12-18
17.8
SOTA
NeurST: Neural Speech Translation Toolkit arXiv:2012.10018
Speech-to-Text TranslationonMuST-C EN->FR
Case-sensitive sacreBLEU· 2020-12-18
33.3
best: 33.45 (Dual-decoder Transformer)
NeurST: Neural Speech Translation Toolkit arXiv:2012.10018
Speech-to-Text TranslationonMuST-C EN->ES
Case-sensitive sacreBLEU· 2020-12-18
27.4
best: 28.73 (Transformer with Adapters)
NeurST: Neural Speech Translation Toolkit arXiv:2012.10018