Class-conditional image generation on ImageNet-1k 256x256
1.48FIDRAR-XXL
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| RAR-XXLType=AR, Params=1.4B, Arbitrary Resolutions=✗, Tokens=2562026.04 | 1.48 | 326 | — | — | — | |
| MaskBitType=Mask, Params=305B, Arbitrary Resolutions=✗, Tokens=2562026.04 | 1.52 | 328.6 | — | — | — | |
| VAR-d30-reType=VAR, Params=2B, Arbitrary Resolutions=✗, Tokens=6802026.04 | 1.73 | 350.2 | — | — | — | |
| MAGVIT-v2Type=Mask, Params=307M, Arbitrary Resolutions=✗, Tokens=2562026.04 | 1.78 | 319.4 | — | — | — | |
| VAR-d30Type=VAR, Params=2B, Arbitrary Resolutions=✗, Tokens=6802026.04 | 1.92 | 323.1 | — | — | — | |
| Open-MAGVIT2-XLType=AR, Params=1.5B, Arbitrary Resolutions=✗, Tokens=2562026.04 | 2.33 | 271.77 | — | — | — | |
| LlamaGen-XXLType=AR, Params=1.4B, Arbitrary Resolutions=✗, Tokens=5762026.04 | 2.34 | 253.91 | — | — | — | |
| TiTok-S/B-128Type=Mask, Params=287M, Arbitrary Resolutions=✗, Tokens=1282026.04 | 2.48 | — | — | — | — | |
| UniTok-XXLType=AR, Params=1.4B, Arbitrary Resolutions=✗, Tokens=2562026.04 | 2.51 | 216.7 | — | — | — | |
| VIM-L-reType=AR, Params=1.7B, Arbitrary Resolutions=✗, Tokens=10242026.04 | 3.04 | 227.4 | — | — | — | |
| VibeToken-Gen-XXLType=AR, Params=1.5B, Arbitrary Resolutions=✓, Tokens=642026.04 | 3.62 | 226.93 | — | — | — | |
| MaskGITType=Mask, Params=177M, Arbitrary Resolutions=✗, Tokens=2562026.04 | 4.02 | 355.6 | — | — | — | |
| GPT2-reType=AR, Params=1.4B, Arbitrary Resolutions=✗, Tokens=2562026.04 | 5.2 | 280.3 | — | — | — | |
| VibeToken-Gen-BType=AR, Params=87M, Arbitrary Resolutions=✓, Tokens=642026.04 | 7.62 | 185.92 | — | — | — | |
| SSCModel=SiT-XL/2, Implementation=Inference2026.05 | 14.26 | — | 5.59 | 23.7 | 30.9 | |
| MAFMModel=SiT-XL/2, Implementation=Train2026.05 | 16.68 | — | 4.67 | 10.8 | 42.3 | |
| SSCModel=SiT-L/2, Implementation=Inference2026.05 | 18 | — | 5.77 | 22.3 | 34.3 | |
| Flow MatchingModel=SiT-XL/22026.05 | 18.7 | — | 8.09 | — | — | |
| MAFMModel=SiT-L/2, Implementation=Train2026.05 | 20.55 | — | 4.99 | 11.4 | 43.2 | |
| Flow MatchingModel=SiT-L/22026.05 | 23.18 | — | 8.79 | — | — | |
| SSCModel=SiT-B/2, Implementation=Inference2026.05 | 32.72 | — | 6.46 | 19.5 | 42 | |
| MAFMModel=SiT-B/2, Implementation=Train2026.05 | 35.83 | — | 6.12 | 11.8 | 45.1 | |
| Flow MatchingModel=SiT-B/22026.05 | 40.64 | — | 11.14 | — | — | |
| SSCModel=SiT-S/2, Implementation=Inference2026.05 | 52.85 | — | 7.83 | 18.5 | 51.1 | |
| MAFMModel=SiT-S/2, Implementation=Train2026.05 | 57.25 | — | 8.53 | 11.7 | 46.7 | |
| Flow MatchingModel=SiT-S/22026.05 | 64.83 | — | 16 | — | — |