fairseq/models/speech_to_text/s2t_w2v2_transformer.py · 6e37ffbaccf440b5466d4ce19b7fa86281c5bb6a · xuchen / Fairseq-S2T · GitLab

Blame History Permalink

Cumulative updates. · 6e37ffba
```
implement mix-speech and optimize the ctc entropy loss
```
xuchen committed Oct 25, 2022
6e37ffba

s2t_w2v2_transformer.py 15.5 KB