Offline Speech-to-Speech Translation on CVSS-C v1 (test)
25.5ASR-BLEU (DE)Fairseq
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| FairseqMODEL TYPE=Discrete Multi-Decoder (UnitY)2023.04 | 25.5 | 32.3 | 30.9 | 29.6 | |
| ESPnet-ST-v2MODEL TYPE=Discrete Multi-Decoder (UnitY)2023.04 | 23.7 | 32 | 33.1 | 29.6 |