Speech Recognition on Vietnamese (Vi) (test)
14.05Average WERvinai/PhoWhisper-base
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| vinai/PhoWhisper-base2025.09 | 14.05 | 9.4 | 14.33 | 16.08 | 16.42 | |
| TSPCTraining Strategy=SSL P2T enc + joint FT2025.09 | 15.87 | 12.42 | 9.94 | 18.01 | 22.39 | |
| TSPCTraining Strategy=Joint FT2025.09 | 16.47 | 13.64 | 10.63 | 18.3 | 23.33 | |
| TSPCTraining Strategy=baseline2025.09 | 18.13 | 14.9 | 11.59 | 20.93 | 25.13 | |
| Wav2vec2-vn-base2025.09 | 21.7 | 13.71 | 16.4 | 32.42 | 24.31 | |
| Qwen3-ASR-0.6B2025.09 | 22.47 | 12.65 | 10.49 | 40.97 | 25.76 | |
| openai/whisper-large-v3-turbo2025.09 | 45.23 | 27.77 | 70.31 | 41.19 | 41.65 | |
| openai/whisper-base2025.09 | 74.83 | 52.01 | 92.94 | 92.9 | 61.5 | |
| facebook/mms-1b-all2025.09 | 93.14 | 92.49 | 92.86 | 91.88 | 95.33 |