Text-to-Image Generation on GenEval 2 (Component Scores)
83.59Object AccuracyHyperNoise
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| HyperNoiseBase Model=SANA-Sprint, Time(s)=0.5322026.05 | 83.59 | 54.18 | 34.84 | 36.32 | 20.66 | 45.92 | |
| ReNoBase Model=SANA-Sprint, Time(s)=14.9642026.05 | 83.16 | 56.57 | 34.73 | 38.36 | 13.44 | 45.25 | |
| LENSBase Model=SANA-Sprint, Time(s)=0.3522026.05 | 82.98 | 55.27 | 36.12 | 36.38 | 18.21 | 45.79 | |
| ReNoBase Model=SD-Turbo, Time(s)=11.3712026.05 | 81.02 | 50.42 | 31.91 | 32.58 | 6.9 | 40.57 | |
| SANA-SprintBase Model=SANA-Sprint, Time(s)=0.3362026.05 | 80.23 | 54.73 | 33.01 | 33.31 | 15.53 | 43.36 | |
| ReNoBase Model=Hyper-SDXL, Time(s)=45.132026.05 | 80.06 | 56.79 | 32.61 | 27.53 | 3.59 | 40.12 | |
| LENSBase Model=SD-Turbo, Time(s)=0.1122026.05 | 75.07 | 43.08 | 26.08 | 28.46 | 5.78 | 35.69 | |
| HyperNoiseBase Model=SD-Turbo, Time(s)=0.1942026.05 | 74.94 | 43 | 23.95 | 27.67 | 5.23 | 34.96 | |
| LENSBase Model=Hyper-SDXL, Time(s)=0.4242026.05 | 72.96 | 53.47 | 27.84 | 21.45 | 4.39 | 36.02 | |
| SD-TurboBase Model=SD-Turbo, Time(s)=0.1072026.05 | 72.27 | 42.65 | 21.15 | 28.1 | 5.6 | 33.95 | |
| HyperNoiseBase Model=Hyper-SDXL, Time(s)=0.6942026.05 | 71.97 | 52.02 | 27.06 | 21.58 | 4.76 | 35.48 | |
| Hyper-SDXLBase Model=Hyper-SDXL, Time(s)=0.4142026.05 | 71.94 | 51.97 | 26.88 | 21.14 | 3.28 | 35.04 |