Text-to-Image Generation on MSCOCO 2014 (FID-30k, Clip Score)
9.29FID (30k)ICC (CA-SVD)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ICC (CA-SVD)r=64, p=2, Sampler=DPM, #Steps=26, #MACs=3.94T2025.05 | 9.29 | 30.64 | |
| PixArt-αcfg=4.5, Sampler=DPM, #Steps=30, #MACs=8.33T2025.05 | 9.54 | 30.66 | |
| ICC (SVD)r=64, p=2, Sampler=DPM, #Steps=26, #MACs=3.94T2025.05 | 9.8 | 30.6 | |
| Naive Cachingp=2, Sampler=DPM, #Steps=30, #MACs=4.17T2025.05 | 10.39 | 30.54 | |
| ICC (CA-SVD)r=64, p=2, Sampler=DPM, #Steps=18, #MACs=2.73T2025.05 | 13.94 | 30.29 | |
| Dual-SolverNFE=6, Backbone=SANA2026.03 | 18.52 | 0.4837 | |
| ICC (SVD)r=64, p=2, Sampler=DPM, #Steps=18, #MACs=2.73T2025.05 | 18.69 | 30.04 | |
| Dual-SolverNFE=5, Backbone=SANA2026.03 | 18.81 | 0.4821 | |
| Dual-SolverNFE=4, Backbone=SANA2026.03 | 19.4 | 0.4801 | |
| Dual-SolverNFE=8, Backbone=SANA2026.03 | 19.43 | 0.4844 | |
| Dual-SolverNFE=7, Backbone=SANA2026.03 | 19.57 | 0.4843 | |
| Dual-SolverNFE=9, Backbone=SANA2026.03 | 19.77 | 0.4849 | |
| DS-SolverNFE=6, Backbone=SANA2026.03 | 20.65 | 0.4816 | |
| BNS-SolverNFE=6, Backbone=SANA2026.03 | 20.66 | 0.4807 | |
| BNS-SolverNFE=7, Backbone=SANA2026.03 | 20.79 | 0.4808 | |
| BNS-SolverNFE=5, Backbone=SANA2026.03 | 21.04 | 0.48 | |
| BNS-SolverNFE=8, Backbone=SANA2026.03 | 21.13 | 0.4809 | |
| DS-SolverNFE=7, Backbone=SANA2026.03 | 21.43 | 0.4818 | |
| BNS-SolverNFE=9, Backbone=SANA2026.03 | 21.64 | 0.4808 | |
| DS-SolverNFE=5, Backbone=SANA2026.03 | 21.66 | 0.4808 | |
| Dual-SolverNFE=3, Backbone=SANA2026.03 | 21.79 | 0.4795 | |
| DS-SolverNFE=8, Backbone=SANA2026.03 | 21.8 | 0.4818 | |
| DS-SolverNFE=9, Backbone=SANA2026.03 | 22.27 | 0.4816 | |
| DPM-Solver++NFE=6, Backbone=SANA2026.03 | 22.48 | 0.4821 | |
| DPM-Solver++NFE=5, Backbone=SANA2026.03 | 22.56 | 0.4813 | |
| DPM-Solver++NFE=7, Backbone=SANA2026.03 | 22.79 | 0.4822 | |
| DPM-Solver++NFE=8, Backbone=SANA2026.03 | 22.9 | 0.482 | |
| DDIMNFE=8, Backbone=SANA2026.03 | 22.96 | 0.4824 | |
| DDIMNFE=9, Backbone=SANA2026.03 | 22.97 | 0.4824 | |
| DDIMNFE=7, Backbone=SANA2026.03 | 22.99 | 0.4824 | |
| DDIMNFE=6, Backbone=SANA2026.03 | 23.06 | 0.4821 | |
| DPM-Solver++NFE=9, Backbone=SANA2026.03 | 23.11 | 0.482 | |
| DDIMNFE=5, Backbone=SANA2026.03 | 23.93 | 0.4813 | |
| Naive Cachingp=2, Sampler=DPM, #Steps=20, #MACs=2.78T2025.05 | 25.47 | 29.68 | |
| DPM-Solver++NFE=4, Backbone=SANA2026.03 | 26.12 | 0.4779 | |
| BNS-SolverNFE=4, Backbone=SANA2026.03 | 26.37 | 0.4772 | |
| DDIMNFE=4, Backbone=SANA2026.03 | 27.72 | 0.4782 | |
| DS-SolverNFE=4, Backbone=SANA2026.03 | 29.15 | 0.4765 | |
| PixArt-αcfg=4.5, Sampler=DPM, #Steps=15, #MACs=4.17T2025.05 | 39.57 | 29.11 | |
| DDIMNFE=3, Backbone=SANA2026.03 | 45.05 | 0.4656 | |
| DPM-Solver++NFE=3, Backbone=SANA2026.03 | 45.33 | 0.4652 | |
| BNS-SolverNFE=3, Backbone=SANA2026.03 | 48.16 | 0.4651 | |
| DS-SolverNFE=3, Backbone=SANA2026.03 | 48.65 | 0.4635 | |
| PixArt-αcfg=4.5, Sampler=DPM, #Steps=10, #MACs=2.78T2025.05 | 86.54 | 27.21 |