Full-audio music detection on AIME Full Audio
98.67AccuracyFusion Segment Transformer
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Fusion Segment TransformerArchitecture=Fusion Segment Transformer, Feature Extractor=MERT2026.01 | 98.67 | 99 | 98.35 | 98.68 | 99.95 | 98.99 | |
| Segment TransformerArchitecture=Segment Transformer, Feature Extractor=MERT2026.01 | 98.58 | 99.17 | 98.02 | 98.59 | 99.92 | 99.16 | |
| Fusion Segment TransformerArchitecture=Fusion Segment Transformer, Feature Extractor=FXencoder2026.01 | 98.5 | 98.43 | 98.21 | 98.32 | 99.71 | 98.73 | |
| Segment TransformerArchitecture=Segment Transformer, Feature Extractor=FXencoder2026.01 | 98.3 | 98.43 | 97.77 | 98.1 | 99.37 | 98.73 | |
| MERT (Standalone)Architecture=Standalone with linear classifier, Feature Extractor=MERT2026.01 | 97.08 | 97.5 | 98.02 | 98.59 | 99.39 | 97.47 | |
| Segment TransformerArchitecture=Segment Transformer, Feature Extractor=Music2vec2026.01 | 96.35 | 95.6 | 97.1 | 96.35 | 98.68 | 95.61 | |
| Fusion Segment TransformerArchitecture=Fusion Segment Transformer, Feature Extractor=Music2vec2026.01 | 96.13 | 95.79 | 96.43 | 96.11 | 99.12 | 95.83 | |
| Segment TransformerArchitecture=Segment Transformer, Feature Extractor=Wav2vec 2.02026.01 | 90.82 | 88.42 | 93.75 | 91.01 | 96.42 | 87.94 | |
| Fusion Segment TransformerArchitecture=Fusion Segment Transformer, Feature Extractor=Wav2vec 2.02026.01 | 90.71 | 87.6 | 94.64 | 90.99 | 96.59 | 86.84 | |
| Segment TransformerArchitecture=Segment Transformer, Feature Extractor=Muffin2026.01 | 87.39 | 83.81 | 92.41 | 87.9 | 92.13 | 82.46 | |
| Fusion Segment TransformerArchitecture=Fusion Segment Transformer, Feature Extractor=Muffin2026.01 | 87.28 | 83.77 | 92.19 | 87.78 | 92.46 | 82.46 |