Text-to-image generation on MSCOCO
4.57FIDBiFM
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| BiFMModel=MMDiT, NFE=502026.03 | 4.57 | — | — | — | — | — | — | — | |
| REPAModel=MMDiT, NFE=502026.03 | 4.73 | — | — | — | — | — | — | — | |
| MeanFlowModel=MMDiT, NFE=502026.03 | 5.02 | — | — | — | — | — | — | — | |
| FMModel=MMDiT, NFE=502026.03 | 6.05 | — | — | — | — | — | — | — | |
| SesaHand2026.02 | 18.63 | 0.0055 | 17.77 | 0.0072 | 96.6 | 1.015 | 1.537 | 4.55 | |
| AttentionHand2026.02 | 20.71 | 0.003 | 27.09 | 0.0129 | 96.5 | 1.026 | 1.986 | 4.05 | |
| ControlNet2026.02 | 21.67 | 0.0066 | 40.32 | 0.021 | 81 | 1.252 | 2.182 | 3.21 | |
| T2I-Adapter2026.02 | 22 | 0.0076 | 32.08 | 0.0157 | 91.4 | 1.546 | 2.451 | 2.91 | |
| Uni-ControlNet2026.02 | 30.34 | 0.0074 | 37.77 | 0.02 | 85.5 | 2.105 | 3.039 | 2.56 | |
| Stable Diffusion2026.02 | 40.52 | 0.0068 | 50.78 | 0.0255 | 65.1 | 2.932 | 4.591 | 1.5 |