Automatic Speech Recognition on SPGISpeech (val)
3.43WER (%)READ (Segment-level combination)
Evaluation results
| Method | Links | |
|---|---|---|
| READ (Segment-level combination)SNR=20dB2026.06 | 3.43 | |
| ROVERSNR=20dB2026.06 | 3.59 | |
| Best single-systemSNR=20dB2026.06 | 3.65 | |
| READ (Segment-level combination)SNR=10dB2026.06 | 4.06 | |
| Best single-systemSNR=10dB2026.06 | 4.11 | |
| ROVERSNR=10dB2026.06 | 4.7 | |
| READ (Segment-level combination)SNR=0dB2026.06 | 11.26 | |
| Best single-systemSNR=0dB2026.06 | 11.69 | |
| ROVERSNR=0dB2026.06 | 18.11 |