Automatic Speech Recognition on Europarl-ST Es-En v1.0 (test)
0.62ASR-chrFS2TT-TTS
Evaluation results
| Method | Links | |
|---|---|---|
| S2TT-TTSPre-training=wav2vec2.0 (encoder)2022.12 | 0.62 | |
| ASR-MT-TTSPre-training=wav2vec2.0 (encoder)2022.12 | 0.618 | |
| UnitY + t-mBARTDecoder layers=12L->2L, Pre-training=wav2vec2.0 (encoder), t-mBART2022.12 | 0.606 | |
| UnitYDecoder layers=6L->6L, Pre-training=wav2vec2.0 (encoder)2022.12 | 0.591 | |
| S2UT + u-mBARTPre-training=wav2vec2.0 (encoder), u-mBART2022.12 | 0.587 | |
| S2SpecT2 + t-mBARTDecoder layers=12L->6L, Pre-training=wav2vec2.0 (encoder), t-mBART2022.12 | 0.485 | |
| S2SpecT2Decoder layers=6L->6L, Pre-training=wav2vec2.0 (encoder)2022.12 | 0.484 |