List your company

Image Generation on ImageNet 512x512

334.7ISFREPix-XL
24.884105.317185.75266.183Dec 15, 2023May 17, 2024Oct 18, 2024Mar 21, 2025Aug 22, 2025Jan 23, 2026Jun 26, 2026

Evaluation results

MethodLinks
FREPix-XLParams=674M, Epochs=330, NFE=100×22026.05334.72.38—8059————
PixelU-H/32Params=1152M, Epochs=600, NFE=100×2, GFLOPs=138, Diffusion Space=Pixel-space2026.06322.11.925.988058————
DA-VAETraining Regime=Fine-tune, AutoEncoders=DA-VAE (f32c128p1), Token-nums=16×16, Training Epoches=802026.03314.3—————0.474.841.68
JiT-H/32Params=756M, Epochs=600, NFE=100×22026.05309.11.94———————
JiT-H/32Params=682M, Epochs=600, NFE=100×2, GFLOPs=183, Diffusion Space=Pixel-space2026.06309.11.94———————
JiT-G/32NFE=100x2, space=pixel, params=2B, Gflops=3842026.01306.81.78———————
DDT-XL/2params=675+49M, Gflops=525, Diffusion Space=Latent, Architecture=XL/22026.01305.11.28———————
DeCo-XL/16Params=682M, Epochs=340, NFE=100×22026.052902.22—8060————
DeCo-XL/16Params=682M, Epochs=340, NFE=100×2, Diffusion Space=Pixel-space2026.062902.224.678060————
pMF-H/32NFE=1, space=pixel, params=959M, Gflops=2722026.01284.92.48———————
VDM++NFE=256x2, space=pixel, params=2.5B, Gflops=5552026.01278.12.65———————
VDM++params=2B, Gflops=555, Diffusion Space=Pixel, Architecture=UViT/42026.01278.12.65———————
VDM++CFG=true2026.03278.12.65———————
VDM++Params=2B, Epochs=800, NFE=250×22026.05278.12.65———————
VDM++CFG scale=w = 1.352026.05278.12.65———————
VDM++Params=2B, Epochs=800, NFE=250×2, GFLOPs=555, Diffusion Space=Pixel-space2026.06278.12.65———————
DA-VAETraining Regime=Fine-tune, AutoEncoders=DA-VAE (f32c128p1), Token-nums=16×16, Training Epoches=252026.03277.6—————0.476.042.07
k-Diffparams=133M, Gflops=26, Diffusion Space=Pixel, Architecture=JiT-B/322026.01277.14.03———————
pMF-L/32NFE=1, space=pixel, params=413M, Gflops=1172026.01276.82.75———————
SiT-XL/2 + REPANFE=250x2, space=latent, params=675M (49M), Gflops=525 (1230)2026.01274.62.08———————
REPAparams=675+49M, Gflops=525, Diffusion Space=Latent, Architecture=SiT-XL/22026.01274.62.08———————
REPAEpochs=200, CFG=true2026.03274.62.084.198358————
REPATraining Regime=Scratch, Training Epoches=2002026.03274.6———————2.08
pMF-B/32NFE=1, space=pixel, params=120M, Gflops=342026.01271.93.7———————
PixelDiT-XLParams=797M, Epochs=360, NFE=100×22026.05271.12.21—7865————
JiT-B/32params=133M, Gflops=26, Diffusion Space=Pixel2026.012714.02———————
StyleGAN-XLNFE=1, space=pixel, params=168M, Gflops=20612026.01267.82.41———————
DiverseDiTEpochs=200, CFG=true2026.03267.121.994.198361————
SiT-XL/2 + sREPAEpochs=100, CFG scale=w = 1.352026.05260.32.294.188457————
RAE + DiTDH-XL/2NFE=50x2, space=latent, params=831M (415M), Gflops=642 (408)2026.01259.61.13———————
RAEparams=839+415M, Gflops=642, Diffusion Space=Latent, Architecture=DiTDH-XL/22026.01259.61.13———————
MakDiTEpochs=800, CFG=true2026.03256.32.55.18356————
MaskDiTEpochs=800, CFG scale=w = 1.352026.05256.32.55.18356————
SiT-XL/2 + REPAEpochs=100, CFG scale=w = 1.352026.05255.72.324.168456————
DiT-XLTraining Regime=Scratch, AutoEncoders=SD-VAE (f8c4p2), Token-nums=32×32, Training Epoches=24002026.03255.3—————0.4812.043.04
LightningDiT-XLTraining Regime=Fine-tune, AutoEncoders=VA-VAE (f16c32p2), Token-nums=16×16, Training Epoches=802026.03254.5—————0.511.313.12
SiT-XL/2 + sREPAEpochs=80, CFG scale=w = 1.352026.05254.32.44.248454————
SiT-XL/2NFE=250x2, space=latent, params=675M (49M), Gflops=525 (1230)2026.01252.22.62———————
SiT-XL/2params=675+49M, Gflops=525, Diffusion Space=Latent, Architecture=XL/22026.01252.22.62———————
SiT-XL/2Epochs=600, CFG=true2026.03252.22.624.188457————
SiT-XL/2Params=675M + 86M, Epochs=600, NFE=250×22026.05252.22.62—8457————
SiT-XL/2Epochs=600, CFG scale=w = 1.352026.05252.22.624.188457————
SiT-XL/2Params=675M + 86M, Epochs=600, NFE=250×2, GFLOPs=525, Diffusion Space=Latent-space2026.06252.22.624.188457————
SiD, UViTNFE=1000x2, space=pixel, params=2.5B, Gflops=5552026.01248.73.02———————
SiDparams=2B, Gflops=555, Diffusion Space=Pixel, Architecture=UViT/42026.01248.73.02———————
SiT-XL/2 + REPAEpochs=80, CFG scale=w = 1.352026.05247.32.444.218456————
PixNerdparams=700 M, Gflops=583, Diffusion Space=Pixel, Architecture=XL/162026.01245.62.84———————
PixNerd-XL/16Params=700M, Epochs=340, NFE=100×22026.05245.62.84—8059————
PixNerd-XL/16Params=700M, Epochs=340, NFE=100×2, GFLOPs=583, Diffusion Space=Pixel-space2026.06245.62.845.958059————
DiT w/ Faster DiffusionSampling Steps (T)=250, Image Resolution=512x512, Latency (s/image)=17.352023.12245.133.255.058351————
DiverseDiTEpochs=80, CFG=true2026.03241.082.214.318261————
DiTSampling Steps (T)=250, Image Resolution=512x512, Latency (s/image)=26.252023.12240.823.045.028454————
DiT-XL/2NFE=250x2, space=latent, params=675M (49M), Gflops=525 (1230)2026.01240.83.04———————
DiT-XL/2params=675+49M, Gflops=525, Diffusion Space=Latent, Architecture=XL/22026.01240.83.04———————
DiT-XL/2Epochs=600, CFG=true2026.03240.83.045.028454————
DiT-XL/2Params=675M + 86M, Epochs=600, NFE=250×22026.05240.83.04—8454————
DiT-XL/2Epochs=600, CFG scale=w = 1.352026.05240.83.045.028454————
DiT-XL/2Params=675M + 86M, Epochs=600, NFE=250×2, GFLOPs=525, Diffusion Space=Latent-space2026.06240.83.045.028454————
LightningDiT-XL*Training Regime=Scratch, AutoEncoders=VA-VAE (f16c32p2), Token-nums=16×16, Training Epoches=802026.03229.7—————0.521.793.98
ADM-GEpochs=400, CFG=true2026.03221.72.855.868453————
ADM-G, ADM-UEpochs=400, CFG scale=w = 1.352026.05221.72.855.868453————
RINparams=320M, Gflops=415, Diffusion Space=Pixel2026.012163.95———————
RINParams=320M, NFE=250, GFLOPs=415, Diffusion Space=Pixel-space2026.062163.95———————
RINParams=320M, NFE=2502026.052103.95———————
Simple diffusion (U-ViT, L)Epochs=800, CFG scale=w = 1.352026.05205.34.53———————
SimpleDiffusionParams=2B, Epochs=800, NFE=250×2, GFLOPs=555, Diffusion Space=Pixel-space2026.062053.54———————
CubiDParams (B)=0.95, Tokenizer=DC-AE-f32c32, Token Dimension=322026.03188.7————1.58———
USiT-2BParams (B)=1.58, Tokenizer=DC-AE-f32c32, Token Dimension=322026.03187.68————1.72———
USiT-HParams (B)=0.50, Tokenizer=DC-AE-f32c32, Token Dimension=322026.03174.58————1.89———
ADM-GNFE=250x2, space=pixel, params=559M, Gflops=19832026.01172.77.72———————
ADM-Gparams=559M, Gflops=1983, Diffusion Space=Pixel2026.01172.77.72———————
ADM-GParams=554M, Epochs=400, NFE=2502026.05172.77.72—8753————
ADM-GParams=554M, Epochs=400, NFE=250, GFLOPs=1983, Diffusion Space=Pixel-space2026.06172.77.726.578753————
Simple diffusion (U-Net)Epochs=800, CFG scale=w = 1.352026.051714.28———————
BigGAN-deepNFE=1, space=pixel, params=56M, Gflops=762026.01152.87.5———————
DiverseDiTBase Model=REPA-XL, Iter.=400k, CFG=1.352026.03137.097.185.387864————
SiT-XLParams (B)=0.67, Tokenizer=DC-AE-f32c32, Token Dimension=322026.03131.37————2.41———
REPA-XLIter.=400k, CFG=1.352026.03127.837.915.417965————
DC-Gen-DiT-XLTraining Regime=Fine-tune, AutoEncoders=DC-AE (f32c32p1), Token-nums=16×16, Training Epoches=802026.03122.5—————0.668.212.22
DiT-XLTraining Regime=Scratch, AutoEncoders=DC-AE (f32c32p1), Token-nums=16×16, Training Epoches=24002026.03117.5—————0.669.562.84
DiverseDiTBase Model=REPA-L, Iter.=400k, CFG=1.352026.031149.835.497864————
REPA-LIter.=400k, CFG=1.352026.03106.4310.825.527863————
DiC-HG FLOPs (Wino.)=817.2 (388.4), TP=53.3, Training Steps=400K2024.12101.7812.89———————
DiC-XLG FLOPs (Wino.)=464.3 (228.7), TP=84.2, Training Steps=400K2024.1293.5515.32———————
Subtraction-based Boundary RF ModelBackbone=DiT-B, Sampler=SDE2025.0680.9214.58———————
Mask-based Boundary RF ModelBackbone=DiT-B, Sampler=SDE2025.0677.2515.69———————
DiverseDiTBase Model=SiT-XL, Iter.=400k, CFG=1.352026.0376.4517.685.487760————
Subtraction-based Boundary RF ModelBackbone=DiT-B, Sampler=Euler2025.0675.9615.08———————
Vanilla RF ModelBackbone=DiT-B, Sampler=SDE2025.0672.9316.7———————
DiverseDiTBase Model=SiT-L, Iter.=400k, CFG=1.352026.0371.7819.195.747661————
SiT-XLIter.=400k, CFG=1.352026.0371.5719.655.557560————
Mask-based Boundary RF ModelBackbone=DiT-B, Sampler=Euler2025.0670.3515.21———————
Vanilla RF ModelBackbone=DiT-B, Sampler=Euler2025.0669.9216.35———————
DiT-XL/2G FLOPs (Wino.)=524.7, TP=16.2, Training Steps=400K2024.1266.320.94———————
DiverseDiTBase Model=REPA-B, Iter.=400k, CFG=1.352026.0364.6223.827.767063————
SiT-LIter.=400k, CFG=1.352026.0364.0522.755.787363————
REPA-BIter.=400k, CFG=1.352026.0353.9230.137.796864————
DiverseDiTBase Model=SiT-B, Iter.=400k, CFG=1.352026.0345.0933.186.926765————
SiT-BIter.=400k, CFG=1.352026.0336.843.467.536064————
HSOParadigm=Hierarchical Opt., Prep. Cost=~8s, NFE=42025.11—15.71———————