Image Generation on CIFAR100
3.86FIDDLSM
Evaluation Results
| Method | Links | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DLSM2022.03 | 3.86 | 11.62 | — | 61 | 63 | 82 | 71 | 0.28 | 42 | 68 | 43 | 52 | — | |
| Base method2022.03 | 4.52 | 11.53 | — | 62 | 62 | 84 | 71 | 0.15 | 33 | 68 | 30 | 38 | — | |
| Omni-INR-GANweight decay=True, differentiable data augmentation=True2020.11 | 6.7 | 14.15 | — | — | — | — | — | — | — | — | — | — | — | |
| Omni-INR-GANweight decay=True2020.11 | 7.63 | 13.9 | — | — | — | — | — | — | — | — | — | — | — | |
| ImAC-GANweight decay=True2020.11 | 8.11 | 13.38 | — | — | — | — | — | — | — | — | — | — | — | |
| FQ-GAN2020.11 | 8.23 | 10.62 | — | — | — | — | — | — | — | — | — | — | — | |
| Omni-GANweight decay=True2020.11 | 8.41 | 13.3 | — | — | — | — | — | — | — | — | — | — | — | |
| ImAC-GANweight decay=True, differentiable data augmentation=True2020.11 | 8.45 | 14.54 | — | — | — | — | — | — | — | — | — | — | — | |
| BigGANweight decay=False, differentiable data augmentation=True2020.11 | 8.63 | 10.69 | — | — | — | — | — | — | — | — | — | — | — | |
| QC-NCSN++Sampler=RK45 ODE solver, NFE=1312022.09 | 8.9 | 10.12 | — | 59 | 64 | — | — | — | — | — | — | — | — | |
| U-NCSN++Sampler=RK45 ODE solver, NFE=1682022.09 | 8.95 | 10.09 | — | 59 | 63 | — | — | — | — | — | — | — | — | |
| BigGANweight decay=False2020.11 | 10.12 | 10.96 | — | — | — | — | — | — | — | — | — | — | — | |
| Omni-GANweight decay=True, differentiable data augmentation=True2020.11 | 11.39 | 15.37 | — | — | — | — | — | — | — | — | — | — | — | |
| BigGAN + DiffAug + DigGANArchitecture=BigGAN, Regularization=DigGAN, Data Augmentation=DiffAug2022.11 | 11.63 | 11.45 | — | — | — | — | — | — | — | — | — | — | — | |
| BigGAN + DiffAug + RLCArchitecture=BigGAN, Regularization=RLC, Data Augmentation=DiffAug2022.11 | 11.84 | 10.77 | — | — | — | — | — | — | — | — | — | — | — | |
| BigGAN + RLC2021.04 | 11.88 | — | — | — | — | — | — | — | — | — | — | — | — | |
| BigGAN + RLCArchitecture=BigGAN, Regularization=RLC2022.11 | 11.88 | 11.18 | — | — | — | — | — | — | — | — | — | — | — | |
| BigGAN + DiffAugArchitecture=BigGAN, Data Augmentation=DiffAug2022.11 | 12 | 11.87 | — | — | — | — | — | — | — | — | — | — | — | |
| HMCParameterization=Energy2023.07 | 12.4 | — | — | — | — | — | — | — | — | — | — | — | 84.9 | |
| LS GAN2021.04 | 12.43 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Scaling method2022.03 | 12.58 | 11.59 | — | 65 | 52 | 93 | 61 | 0.24 | 59 | 41 | 78 | 51 | — | |
| LAParameterization=Energy2023.07 | 12.7 | — | — | — | — | — | — | — | — | — | — | — | 80 | |
| BigGAN + DigGANArchitecture=BigGAN, Regularization=DigGAN2022.11 | 12.93 | 10.92 | — | — | — | — | — | — | — | — | — | — | — | |
| HMC-3CParameterization=Score2023.07 | 13.3 | — | — | — | — | — | — | — | — | — | — | — | 75.8 | |
| BigGANArchitecture=BigGAN2022.11 | 13.54 | 10.58 | — | — | — | — | — | — | — | — | — | — | — | |
| BigGAN2021.04 | 13.6 | — | — | — | — | — | — | — | — | — | — | — | — | |
| BigGANweight decay=True2020.11 | 13.73 | 9.98 | — | — | — | — | — | — | — | — | — | — | — | |
| Non saturated GAN2021.04 | 13.87 | — | — | — | — | — | — | — | — | — | — | — | — | |
| BigGANweight decay=True, differentiable data augmentation=True2020.11 | 13.97 | 9.79 | — | — | — | — | — | — | — | — | — | — | — | |
| RaHinge GAN2021.04 | 14.61 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Multi-hinge2020.11 | 14.62 | 13.35 | — | — | — | — | — | — | — | — | — | — | — | |
| AC-GANweight decay=True, differentiable data augmentation=True2020.11 | 14.91 | 11.01 | — | — | — | — | — | — | — | — | — | — | — | |
| LA-10LParameterization=Score2023.07 | 15.5 | — | — | — | — | — | — | — | — | — | — | — | 75.2 | |
| AC-GANweight decay=True2020.11 | 16.41 | 9.74 | — | — | — | — | — | — | — | — | — | — | — | |
| Adaptive log-SNR reweightingmode=adaptive2026.03 | 20.89 | — | — | — | — | — | — | — | — | — | — | — | — | |
| C-NCSN++Sampler=RK45 ODE solver, NFE=2972022.09 | 21.89 | 7.84 | — | 55 | 56 | — | — | — | — | — | — | — | — | |
| Log-normal2026.03 | 23.31 | — | — | — | — | — | — | — | — | — | — | — | — | |
| U-LAParameterization=Energy2023.07 | 24.6 | — | — | — | — | — | — | — | — | — | — | — | 87.3 | |
| U-HMCParameterization=Energy2023.07 | 25.4 | — | — | — | — | — | — | — | — | — | — | — | 87.2 | |
| ImAC-GANweight decay=False2020.11 | 25.64 | 11.34 | — | — | — | — | — | — | — | — | — | — | — | |
| U-LAParameterization=Score2023.07 | 25.9 | — | — | — | — | — | — | — | — | — | — | — | 82.9 | |
| Omni-GANweight decay=False2020.11 | 26.51 | 11.45 | — | — | — | — | — | — | — | — | — | — | — | |
| U-HMCParameterization=Score2023.07 | 28.6 | — | — | — | — | — | — | — | — | — | — | — | 79 | |
| ReverseParameterization=Score2023.07 | 31.8 | — | — | — | — | — | — | — | — | — | — | — | 74.2 | |
| RiskyDiffTarget Model=ResNet-502025.12 | 32.9 | — | — | — | — | — | — | — | — | — | — | — | — | |
| RiskyDiffTarget Model=DenseNet-1212025.12 | 33.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| ImAC-GANweight decay=False, differentiable data augmentation=True2020.11 | 33.3 | 11.78 | — | — | — | — | — | — | — | — | — | — | — | |
| RiskyDiffTarget Model=GoogleNet2025.12 | 33.4 | — | — | — | — | — | — | — | — | — | — | — | — | |
| ReverseParameterization=Energy2023.07 | 33.4 | — | — | — | — | — | — | — | — | — | — | — | 72.6 | |
| AC-GANweight decay=False2020.11 | 34.19 | 9.31 | — | — | — | — | — | — | — | — | — | — | — | |
| RiskyDiffTarget Model=WideResNet2025.12 | 34.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| AC-GANweight decay=False, differentiable data augmentation=True2020.11 | 38.01 | 9.19 | — | — | — | — | — | — | — | — | — | — | — | |
| Omni-GANweight decay=False, differentiable data augmentation=True2020.11 | 39.9 | 10.66 | — | — | — | — | — | — | — | — | — | — | — | |
| AdvDiffTarget Model=WideResNet2025.12 | 44.1 | — | — | — | — | — | — | — | — | — | — | — | — | |
| AdvDiffuserTarget Model=ResNet-502025.12 | 45 | — | — | — | — | — | — | — | — | — | — | — | — | |
| AdvDiffTarget Model=ResNet-502025.12 | 45 | — | — | — | — | — | — | — | — | — | — | — | — | |
| AdvDiffuserTarget Model=GoogleNet2025.12 | 45.1 | — | — | — | — | — | — | — | — | — | — | — | — | |
| AdvDiffuserTarget Model=DenseNet-1212025.12 | 45.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| AdvDiffuserTarget Model=WideResNet2025.12 | 45.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| AdvDiffTarget Model=GoogleNet2025.12 | 45.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| AdvDiffTarget Model=DenseNet-1212025.12 | 45.3 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Omni-INR-GANweight decay=False2020.11 | 53.29 | 9.05 | — | — | — | — | — | — | — | — | — | — | — | |
| Omni-INR-GANweight decay=False, differentiable data augmentation=True2020.11 | 53.88 | 11.06 | — | — | — | — | — | — | — | — | — | — | — |