Scene-based Audio Classification on HEAR Speech tasks
0.656CREMA-D Scorewav2vec2.0
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| wav2vec2.0Backbone=wav2vec2.02022.09 | 0.656 | 0.692 | 0.838 | 0.879 | 0.08 | 0.493 | 0.606 | — | |
| BYOL-A*Backbone=AudioNTT2022.09 | 0.623 | 0.788 | 0.896 | 0.924 | 0.137 | 0.39 | 0.626 | — | |
| ABTBackbone=AudioNTT2022.09 | 0.594 | 0.745 | 0.882 | 0.91 | 0.111 | 0.324 | 0.594 | 17 | |
| ABTBackbone=ViTC-B, Patch Size=16x82022.09 | 0.581 | 0.812 | 0.724 | 0.771 | 0.087 | 0.312 | 0.548 | 140 | |
| CREPEBackbone=CREPE2022.09 | 0.383 | 0.499 | 0.18 | 0.211 | 0.051 | 0.142 | 0.244 | — |