Libri2Mix
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Speech Separation | Libri2Mix (test) | SI-SNRi (dB)22 | 45 | |
| Target Speaker Extraction | Libri2Mix clean (test) | DNSMOS SIG3.656 | 9 | |
| Target Speech Extraction | Libri2Mix 100 2-speaker+noise 16kHz (test) | PESQ2.33 | 8 | |
| Multi-speaker Automatic Speech Recognition | Libri2Mix | CP-WER14.34 | 8 | |
| Speech Recognition | Libri2Mix max mode (test) | WER (%)15.6 | 8 | |
| Target Speech Extraction | Libri2Mix-100 (1-speaker+noise) 16kHz (test) | SI-SDR14.51 | 7 | |
| Target Speaker Extraction | Libri2Mix Noisy | PESQ2.21 | 7 | |
| Source Separation | Libri2Mix Noisy (test) | SI-SNRi15.9 | 7 | |
| Source Separation | Libri2Mix Clean (test) | SI-SNRi20.6 | 7 | |
| Speech Separation | Libri2Mix max mode (test) | Delta SI-SNR (dB)18.3 | 6 | |
| Target Speech Extraction | Libri2Mix 2-Speaker+Noise 16 kHz (test) | SI-SDR11.12 | 5 | |
| Speech Separation | Libri2mix noisy 8 kHz (test) | ΔSI-SDR15.2 | 5 | |
| Speech Separation | Libri2mix clean 8 kHz (test) | Delta SI-SDR20.5 | 5 | |
| Target Speaker Extraction | Libri2Mix Noisy (test) | SI-SDR13.3 | 5 | |
| Target Speaker Extraction | Libri2Mix | SI-SNRi (dB)16.45 | 4 | |
| Target-speaker Automatic Speech Recognition | Libri2Mix Clean | tcpWER0.028 | 3 | |
| Target-speaker Automatic Speech Recognition | Libri2Mix Both | tcpWER (%)7.7 | 3 | |
| Target Speech Extraction | Libri2Mix 360 2-speaker+noise 16kHz (test) | SI-SDR9.7 | 1 |
Showing 18 of 18 rows