Audio Tagging on MTT
40.8MTT APMATPAC++
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MATPAC++Role=Teacher, Size=86M2026.04 | 40.8 | — | |
| M2DRole=Teacher, Size=86M2026.04 | 40.6 | — | |
| S-SONDO (MobileNetV3)Teacher=MATPAC++, Size=2.9M2026.04 | 40.2 | — | |
| S-SONDO (DyMN)Teacher=MATPAC++, Size=8.7M2026.04 | 39.9 | — | |
| S-SONDO (MobileNetV3)Teacher=M2D, Size=2.9M2026.04 | 39.5 | — | |
| S-SONDO (ERes2Net)Teacher=MATPAC++, Size=1.4M2026.04 | 39.4 | — | |
| S-SONDO (DyMN)Teacher=M2D, Size=8.7M2026.04 | 39.1 | — | |
| DyMNTeacher=None (Supervised), Size=8.7M2026.04 | 38.8 | — | |
| S-SONDO (ERes2Net)Teacher=M2D, Size=1.4M2026.04 | 38.8 | — | |
| MobileNetV3Teacher=None (Supervised), Size=2.9M2026.04 | 38.7 | — | |
| DuoTokRate=25 Hz, Size (S)=2×32768, Kbps=0.752025.11 | 35 | 87 | |
| ERes2NetTeacher=None (Supervised), Size=1.4M2026.04 | 34.3 | — | |
| SemantiCodecRate=50 Hz, Size (S)=2×8192, Kbps=1.302025.11 | 32 | 88 | |
| X-CodecRate=50 Hz, Size (S)=8×1024, Kbps=4.002025.11 | 32 | 87 | |
| X-Codec (YuE)Rate=50 Hz, Size (S)=16×1024, Kbps=8.002025.11 | 32 | 87 | |
| MuCodec (LeVo)Rate=25 Hz, Size (S)=2×16384, Kbps=0.702025.11 | 26 | 84 | |
| DACRate=75 Hz, Size (S)=8×1024, Kbps=6.002025.11 | 20 | 79 | |
| EnCodecRate=75 Hz, Size (S)=8×1024, Kbps=6.002025.11 | 18 | 76 | |
| WavTokenizerRate=40 Hz, Size (S)=1×4096, Kbps=0.482025.11 | 17 | 74 |