Libri-Light
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Spoken Language Modeling | Libri-Light 6k | sWUGGY (all)71.89 | 9 | |
| Acoustic unit discovery | Libri-Light other (test) | ABX Error (Within Speaker)8.14 | 9 | |
| Acoustic unit discovery | Libri-Light clean (test) | ABX Error (Within Speaker)5.83 | 9 | |
| Phoneme Recognition | Libri-Light other (test) | PER48.93 | 8 | |
| Phoneme Recognition | Libri-Light clean (test) | PER32.67 | 8 | |
| Phoneme Recognition | Libri-Light other (dev) | PER46.67 | 8 | |
| Phoneme Recognition | Libri-Light clean (dev) | PER31.16 | 8 | |
| Self-supervised pretraining | Libri-Light 6k | Pretraining Time (hr)23 | 7 | |
| Speech Meaningfulness Evaluation | Libri-Light LL-OTHER | MMOS4.05 | 5 |
Showing 9 of 9 rows