Text-to-Audio Generation on TTA-Bench Accuracy
3.539CE ScoreTangoFlux
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| TangoFluxParams=516M, NFE=502026.03 | 3.539 | 5.074 | 3.673 | 5.782 | 47.2 | 67.7 | 3.51 | 3.66 | |
| Tango-2-FullParams=866M, NFE=2002026.03 | 3.468 | 5.196 | 3.8 | 5.895 | 46.7 | 70.2 | — | — | |
| Resonate-GRPOParams=470M, NFE=252026.03 | 3.451 | 5.328 | 3.232 | 6.064 | 47.6 | 73.7 | 3.86 | 3.83 | |
| AudioLDM-2-LargeParams=718M, NFE=2002026.03 | 3.441 | 5.44 | 2.957 | 5.984 | 41.5 | 58.9 | — | — | |
| Resonate-PTParams=470M, NFE=252026.03 | 3.408 | 5.238 | 3.002 | 5.923 | 45.8 | 65.1 | 3.44 | 3.37 | |
| EzAudio-XLParams=857M, NFE=2002026.03 | 3.388 | 5.055 | 3.663 | 5.719 | 44.6 | 70.9 | 3.54 | 3.53 | |
| GenAU-L-FullParams=1253M, NFE=2002026.03 | 3.359 | 5.096 | 3.584 | 5.802 | 46.8 | 64.6 | 3.52 | 3.37 | |
| MeanAudio-L-FullParams=480M, NFE=252026.03 | 3.304 | 5.151 | 2.995 | 5.743 | 46.9 | 72.9 | — | — | |
| AudioLDM-L-FullParams=739M, NFE=2002026.03 | 3.275 | 5.137 | 3.219 | 5.854 | 44.1 | 57.4 | — | — | |
| Tango-FullParams=866M, NFE=2002026.03 | 3.264 | 5.151 | 3.36 | 5.954 | 44 | 59.9 | — | — |