Class-conditional image generation on ImageNet 1K 512x512 (test)
1.28FIDDDT-XL/2
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| DDT-XL/2Training Steps=500K, CFG=3, Interval Guidance=[0.3, 1]2025.04 | 1.28 | — | — | 305.1 | 80 | 63 | 4.22 | |
| REPA-XL/22025.04 | 2.08 | — | — | 274.6 | 83 | 58 | 4.19 | |
| DiT-XL/2 with CADSGuidance scale (w_CFG)=2.52023.10 | 2.31 | — | — | 239.56 | 80 | 61 | — | |
| DiT-XL/2 with CADSGuidance scale (w_CFG)=2.252023.10 | 2.32 | — | — | 233.03 | 80 | 60 | — | |
| StyleGAN-XL2023.10 | 2.41 | — | — | 267.75 | 77 | 52 | — | |
| StyleGAN-XL2025.04 | 2.41 | — | — | 267.75 | 77 | 52 | 4.06 | |
| FlowDCN-XL/22025.04 | 2.44 | — | — | 252.8 | 84 | 54 | 4.53 | |
| DiT-XL/2 with CADSGuidance scale (w_CFG)=22023.10 | 2.53 | — | — | 219.08 | 80 | 61 | — | |
| SiT-XL/22025.04 | 2.62 | — | — | 252.21 | 84 | 57 | 4.18 | |
| VAR-d36Param.=2.0 B, Steps=102025.03 | 2.63 | — | — | 303.2 | — | — | — | |
| VDM++2023.10 | 2.65 | — | — | 278.1 | — | — | — | |
| ARPG-XLParam.=719 M, Steps=642025.03 | 2.82 | 35.53 | 13.98 | 277.5 | 82 | 56 | — | |
| GIVT-Causal-L+At=0.9, DB-CFG=0.9, Steps=512, Activation caching=true2023.12 | 2.92 | — | — | — | 84 | 55 | — | |
| SimpleDiffusion2023.10 | 3.02 | — | — | 248.7 | — | — | — | |
| DiT-XL/2Param.=675 M, Steps=2502025.03 | 3.04 | 0.18 | 4.7 | 240.8 | 84 | 54 | — | |
| DiT-XL/2-GGuidance scale (w_CFG)=1.52023.10 | 3.04 | — | — | 240.82 | 84 | 54 | — | |
| DiT-XL/2-GClassifier-Free Guidance (CFG)=1.5, Steps=2502023.12 | 3.04 | — | — | — | 84 | 54 | — | |
| DiT-XL/22025.04 | 3.04 | — | — | 240.82 | 84 | 54 | 5.02 | |
| ADM-G, ADM-U2023.10 | 3.85 | — | — | 221.72 | 84 | 53 | — | |
| ADM-G, ADM-U2025.04 | 3.85 | — | — | 221.72 | 84 | 53 | 5.86 | |
| RIN+NoiseSchedule2023.10 | 3.95 | — | — | 216 | — | — | — | |
| GIVT-MaskGITtc=140, Steps=162023.12 | 4.86 | — | — | — | 88 | 48 | — | |
| MaskGITParam.=227 M, Steps=122025.03 | 7.32 | 4.48 | 7.63 | 156 | 78 | 50 | — | |
| ADM-GClassifier Guidance (CG)=1, Steps=2502023.12 | 7.72 | — | — | — | 87 | 42 | — | |
| ADM-G2025.04 | 7.72 | — | — | 172.71 | 87 | 42 | 6.57 | |
| MaskGITtc=4.5, Steps=162023.12 | 7.8 | — | — | — | 86 | 46 | — | |
| GIVT-Causal-Lt=0.9, Steps=512, Activation caching=true2023.12 | 8.35 | — | — | — | 79 | 61 | — | |
| BigGAN-deep2023.10 | 8.43 | — | — | 177.9 | 88 | 29 | — | |
| BigGAN-deep2025.04 | 8.43 | — | — | 177.9 | 88 | 29 | 8.13 | |
| DiT-XL/2Steps=2502023.12 | 12.03 | — | — | — | 75 | 64 | — | |
| ADMSteps=2502023.12 | 23.2 | — | — | — | 73 | 60 | — | |
| VQGANParam.=1.4 B, Steps=10242025.03 | 26.52 | 0.63 | 44.12 | 66.8 | 73 | 31 | — |