Audio Editing (Add) on AudioSetCaps Subset (test)
1.7979LSDHybrid Diffusion Transformer
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Hybrid Diffusion Transformerguidance scale=2.0, sampling steps=2002026.06 | 1.7979 | 0.5027 | 20.65 | 1.47 | 0.81 | 3.02 | |
| RFM-Editingtraining-based=true2026.06 | 1.8036 | 0.4993 | 21.51 | 1.65 | 0.83 | 2.88 | |
| Zero-Shottraining-free=true2026.06 | 2.0904 | 0.5412 | 57.49 | 1.75 | 1.1 | 3.15 | |
| AudioEditortraining-free=true2026.06 | 2.5196 | 0.5028 | 34.31 | 3.57 | 1.46 | 3.55 | |
| AUDITtraining-based=true2026.06 | 2.9971 | 0.2141 | 49.86 | 10.51 | 3 | 3.46 |