Text-to-audio infilling on AudioCaps (test)
13.28IS (Inception Score)Ground-truth
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Ground-truth2023.12 | 13.28 | — | — | — | 0.49 | — | 3.13 | 4.21 | |
| AUDIOBOX SOUND2023.12 | 12.05 | 1.29 | 7.19 | 0.65 | 0.63 | 0.77 | 2.95 | 4.2 | |
| TANGOFine-tuning=Full2023.12 | 8.95 | 1.86 | 15 | 0.71 | 0.56 | 0.78 | 2.79 | 4.07 | |
| TANGO2023.12 | 8.53 | 1.25 | 18.02 | 0.78 | 0.53 | 0.78 | 2.75 | 3.94 | |
| AudioLDM-L-Full2023.12 | 8.27 | 2.65 | 21.27 | 0.84 | 0.51 | 0.76 | 2.58 | 3.58 |