Automatic Speech Recognition on VLSP-T1
11.2WERViP-VL
Evaluation Results
| Method | Links | |
|---|---|---|
| ViP-VLParams=78M, Pretrain=17,000h, Finetune=250h, Head=CTC2026.06 | 11.2 | |
| PhoWhisper-LargeParams=1.55B, Pretrain=680,000h, Finetune=800h, Head=AED2026.06 | 13.75 | |
| VietASRParams=68M, Pretrain=70,000h, Finetune=70,000h, Head=RNN-T2026.06 | 14.47 | |
| Wav2vec2-Large-ViParams=317M, Pretrain=13,000h, Finetune=250h, Head=CTC2026.06 | 15.18 | |
| Wav2vec2-Base-ViParams=95M, Pretrain=13,000h, Finetune=250h, Head=CTC2026.06 | 16.82 | |
| PhoWhisper-BaseParams=74M, Pretrain=680,000h, Finetune=800h, Head=AED2026.06 | 19.7 |