Speech To Text Translation
Speech-to-Text Translation is an important sub-task in the field of Natural Language Processing, aimed at converting speech signals from one language into text form in another language, which can be achieved through end-to-end or cascaded methods. The goal of this task is to enhance the efficiency and accuracy of cross-language communication, and it is widely applied in scenarios such as multilingual meeting transcription, international phone call transcription, online education, and telemedicine, making it highly valuable in practical applications.
CoVoST 2 eng-X
CoVoST 2 X-eng
FLEURS eng-X
FLEURS X-eng
libri-trans
Transformer + ASR Pretrain + SpecAug
MuST-C
Transformer with Adapters
MuST-C EN->DE
Task Modulation + Multitask Learning(ASR/MT) + Data Augmentation
MuST-C EN->ES
Transformer with Adapters
MuST-C EN->FR
Dual-decoder Transformer
MuST-C EN->NL
Speechformer