Audio Editing (Replace) on AudioCaps Subset
1.8833LSDHybrid Diffusion Transformer
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Hybrid Diffusion TransformerMethod=Ours2026.06 | 1.8833 | 0.4138 | 28.77 | 3.11 | 1.52 | 4.16 | |
| RFM-EditingMethod=RFM-Editing2026.06 | 1.9004 | 0.3983 | 29.35 | 3.43 | 1.67 | 4.29 | |
| AudioEditorMethod=AudioEditor2026.06 | 1.9446 | 0.4752 | 31.72 | 2.73 | 2.03 | 7.66 | |
| Zero-ShotMethod=Zero-Shot2026.06 | 1.9611 | 0.4209 | 61.22 | 4.88 | 2.14 | 3.91 | |
| AUDITMethod=AUDIT2026.06 | 2.6302 | 0.1826 | 45.7 | 6.86 | 3.11 | 5.87 |