Speechformer

Reported on 3 benchmarks across 1 task · 1 paper · 1 SOTA

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

Natural Language Processing3 results

Speech-to-Text TranslationonMuST-C EN->NL
Case-sensitive sacreBLEU· 2021-09-09
27.7
SOTA
Speechformer: Reducing Information Loss in Direct Speech Translation arXiv:2109.04574
Speech-to-Text TranslationonMuST-C EN->DE
Case-sensitive sacreBLEU· 2021-09-09
23.6
best: 28.88 (Task Modulation + Multitask Learning(ASR/MT) + Data Augmentation)
Speechformer: Reducing Information Loss in Direct Speech Translation arXiv:2109.04574
Speech-to-Text TranslationonMuST-C EN->ES
Case-sensitive sacreBLEU· 2021-09-09
28.5
best: 28.73 (Transformer with Adapters)
Speechformer: Reducing Information Loss in Direct Speech Translation arXiv:2109.04574