Image Generation On Lsun Horse 256 X 256
评估指标
Clean-FID (trainfull)
评测结果
各个模型在此基准测试上的表现结果
模型名称 | Clean-FID (trainfull) | Paper Title | Repository |
|---|---|---|---|
| PGGAN | 14.09 ± 0.06 | Progressive Growing of GANs for Improved Quality, Stability, and Variation | |
| Vision-aided GAN | 2.11 | Ensembling Off-the-shelf Models for GAN Training | |
| StyleGAN2 | 4.06 ± 0.03 | Analyzing and Improving the Image Quality of StyleGAN | |
| Projected GAN | - | Projected GANs Converge Faster | |
| StyleGAN | 4.78 ± 0.03 | A Style-Based Generator Architecture for Generative Adversarial Networks | |
| ADM (dropout) | - | Diffusion Models Beat GANs on Image Synthesis |
0 of 6 row(s) selected.