Text-to-Audio on AudioCaps multi-event prompts
75.2FDopenl3TANGOFLUX
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| TANGOFLUX#Params.=515M, Duration=30 sec2024.12 | 75.2 | 1.2 | 48.8 | 11.1 | |
| TANGOFLUX-base#Params.=515M, Duration=30 sec2024.12 | 79.7 | 1.23 | 43.8 | 10.7 | |
| Stable Audio Open#Params.=1056M, Duration=47 sec2024.12 | 88.5 | 2.67 | 28.6 | 9.3 | |
| AudioLDM 2-large#Params.=712M, Duration=10 sec2024.12 | 107.9 | 1.83 | 41.5 | 7.3 | |
| Tango 2#Params.=866M, Duration=10 sec2024.12 | 108.3 | 1.14 | 45.2 | 8.4 |