Audio Scene Classification on HEAR Music 2021
0.966BeijingABT AudioNTT
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| ABT AudioNTTEncoder=AudioNTT2022.09 | 0.966 | 0.818 | 0.962 | 0.97 | 0.932 | 0.476 | 0.74 | 0.838 | -1 | |
| HEAR CREPEModel=CREPE2022.09 | 0.928 | 0.645 | 0.929 | 0.898 | 0.824 | 0.87 | 0.9 | 0.856 | — | |
| HEAR BYOL-A*Model=BYOL-A*2022.09 | 0.919 | 0.835 | 0.969 | 0.97 | 0.9 | 0.29 | 0.642 | 0.789 | — | |
| HEAR wav2vec2.0Model=wav2vec2.02022.09 | 0.907 | 0.78 | 0.946 | 0.943 | 0.828 | 0.402 | 0.653 | 0.78 | — | |
| ABT ViTC-B (16 x 8)Encoder=ViTC-B, Patch Size=16x82022.09 | 0.869 | 0.765 | 0.992 | 0.952 | 0.897 | 0.28 | 0.632 | 0.769 | 15 |