Image Generation on CLEVR
0.81FIDHP GAN
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| HP GAN2026.02 | 0.81 | — | — | — | 0.42 | 69.3 | 71.8 | |
| Projected GANResolution=256x2562021.11 | 0.89 | — | — | — | — | — | — | |
| Projected GAN2026.02 | 0.89 | — | — | — | 0.51 | 64 | 73.5 | |
| FastGANResolution=256x2562021.11 | 3.24 | — | — | — | — | — | — | |
| FastGAN2026.02 | 3.24 | — | — | — | 2.64 | 60 | 65 | |
| Projected GANResolution=256x256, Training Point=First point surpassing SOTA2021.11 | 3.39 | — | — | — | — | — | — | |
| StyleGAN2Resolution=256x256, Training Point=Lowest FID in literature2021.11 | 5.05 | — | — | — | — | — | — | |
| GANsformersResolution=256x2562021.11 | 9.24 | — | — | — | — | — | — | |
| StyleGAN2-ADAResolution=256x2562021.11 | 10.17 | — | — | — | — | — | — | |
| ADA2026.02 | 10.17 | — | — | — | 8.15 | 37.3 | 56.9 | |
| Discriminator-Driven Diffusion ModelsGenerated samples=10K2026.01 | 24.16 | — | — | — | — | — | — | |
| Decomp DiffusionGenerated samples=10K2026.01 | 25.7 | — | — | — | — | — | — | |
| SAGANResolution=256x2562021.11 | 26.04 | — | — | — | — | — | — | |
| DATA2021.11 | — | 37 | 43 | — | — | — | — | |
| FastGAN2021.11 | — | 7 | 13 | — | — | — | — | |
| Projected GAN2021.11 | — | 42 | 31 | — | — | — | — | |
| StyleGAN2-ADA2021.11 | — | 14 | 13 | — | — | — | — |