Text-to-Image Generation on Text-to-image generation (test)
14.78FIDPISA
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| PISAModel=SD 3.5-M, Sparsity=70%2026.02 | 14.78 | 0.883 | 49.71 | 30.4 | 0.79 | 19.55 | 0.158 | 2.15 | |
| Dense-AttnModel=FLUX.1 schnell, Sparsity=0%2026.02 | 15.21 | 0.911 | — | 30.22 | — | — | — | 0.82 | |
| PISAModel=FLUX.1 schnell, Sparsity=85%2026.02 | 15.25 | 0.887 | 45.73 | 29.9 | 0.625 | 15.17 | 0.247 | 0.71 | |
| Dense-AttnModel=SD 3.5-M, Sparsity=0%2026.02 | 15.48 | 0.912 | — | 30.13 | — | — | — | 3.03 | |
| SpargeAttnModel=FLUX.1 schnell, Sparsity=80%2026.02 | 15.74 | 0.728 | 38.69 | 28.81 | 0.563 | 13.62 | 0.311 | 0.75 | |
| PISAModel=FLUX.1 dev, Sparsity=85%2026.02 | 15.91 | 1.001 | 51.98 | 31.69 | 0.682 | 17.07 | 0.241 | 6.87 | |
| Dense-AttnModel=FLUX.1 dev, Sparsity=0%2026.02 | 16.35 | 0.965 | — | 31.31 | — | — | — | 8.32 | |
| PISAModel=SD 3.5-L Turbo, Sparsity=85%2026.02 | 17.56 | 0.782 | 43.35 | 29.84 | 0.677 | 16.64 | 0.201 | 0.65 | |
| Dense-AttnModel=SD 3.5-L Turbo, Sparsity=0%2026.02 | 17.7 | 0.877 | — | 30.75 | — | — | — | 0.81 | |
| SpargeAttnModel=FLUX.1 dev, Sparsity=80%2026.02 | 19.2 | 0.906 | 48.13 | 31.31 | 0.539 | 12.88 | 0.296 | 7.47 | |
| SpargeAttnModel=SD 3.5-M, Sparsity=70%2026.02 | 19.52 | 0.465 | 32.32 | 26.3 | 0.627 | 14.6 | 0.269 | 2.69 | |
| SpargeAttnModel=SD 3.5-L Turbo, Sparsity=80%2026.02 | 22.27 | 0.235 | 26.03 | 27.15 | 0.558 | 13.88 | 0.281 | 0.71 |