List your company

Conditional Image Generation on ImageNet 256x256 1.0 (train val)

1.42FIDSiT-XL/2 + REPA
1.05443.52225.998.4578Dec 14, 2024

Evaluation results

MethodLinks
SiT-XL/2 + REPACFG=true2024.121.42305.70.80.65———
MAR-HTok. Model=KL, # Tokens=256, CFG=true2024.121.55303.70.810.62145.080.121.22
MDTv2-XL/2CFG=true2024.121.58314.70.790.65125.430.59—
SiT-XL (SoftVQ-L)Tok. Model=SoftVQ-L, # Tokens=64, CFG=true2024.121.86293.60.810.6329.235.390.61
MAR-H (SoftVQ-BL)Tok. Model=SoftVQ-BL, # Tokens=64, CFG=true2024.121.93289.40.80.6186.550.890.65
TiTok-S-128Tok. Model=VQ, # Tokens=128, CFG=true2024.121.97281.8——33.036.51.61
SiT-XL/2CFG=true2024.122.06270.30.820.5981.920.54—
DiT-XL/2CFG=true2024.122.27278.20.830.5380.730.51—
U-ViT-H/2Tok. Model=KL, # Tokens=1024, Sampler=DPM-Solver, Sampling Steps=50, CFG=true2024.122.29263.90.820.57128.890.980.62
MAR-HTok. Model=KL, # Tokens=256, CFG=false2024.122.35227.80.790.62145.080.121.22
SiT-XL (SoftVQ-L)Tok. Model=SoftVQ-L, # Tokens=32, CFG=true2024.122.44310.60.770.6314.7310.110.74
MAR-H (SoftVQ-L)Tok. Model=SoftVQ-L, # Tokens=32, CFG=true2024.122.54273.60.780.6167.932.190.61
MAR-H (SoftVQ-BL)Tok. Model=SoftVQ-BL, # Tokens=64, CFG=false2024.122.81218.30.780.6286.550.890.65
DiT-XL (SoftVQ-L)Tok. Model=SoftVQ-L, # Tokens=64, CFG=true2024.122.93268.50.840.5528.814.510.61
LlamaGen-3BTok. Model=VQ, # Tokens=256, CFG=true2024.123.06279.70.80.58781.562.92.08
LDM-4Tok. Model=KL, # Tokens=4096, CFG=true2024.123.6247.70.870.48157.920.370.27
DiT-XL (SoftVQ-L)Tok. Model=SoftVQ-L, # Tokens=32, CFG=true2024.123.69270.40.830.5314.528.740.74
RQ-Trans.Tok. Model=RQ, # Tokens=256, CFG=false2024.123.8323.7——908.917.853.2
MAR-H (SoftVQ-L)Tok. Model=SoftVQ-L, # Tokens=32, CFG=false2024.123.83211.20.770.6167.932.190.61
Taming-Trans.Tok. Model=VQ, # Tokens=256, CFG=false2024.125.2290.3————7.94
MaskGITTok. Model=VQ, # Tokens=256, CFG=false2024.126.18182.10.80.51——2.28
MAGETok. Model=VQ, # Tokens=256, CFG=false2024.126.93195.8—————
LDM-4Tok. Model=KL, # Tokens=4096, CFG=false2024.1210.56103.50.710.62157.920.370.27