Conditional Image Generation On Imagenet 2
평가 지표
FID
Inception score
평가 결과
이 벤치마크에서 각 모델의 성능 결과
모델 이름 | FID | Inception score | Paper Title | Repository |
|---|---|---|---|---|
| IC-GAN (chx96) + DA | 8.2±0.1 | 173.8±0.9 | Instance-Conditioned GAN | |
| DiGIT | 3.39 | 205.96 | Stabilize the Latent Space for Image Autoregressive Modeling: A Unified Perspective | |
| ADM-G | 4.59 | 186.7 | Diffusion Models Beat GANs on Image Synthesis | |
| BigGAN+ [Brock et al.] (chx96) | 8.1 | 144.2 | Instance-Conditioned GAN | |
| ADM-G + EDS + ECT (ED-DPM, classifier_scale=1.0) | 4.09 | 221.57 | Entropy-driven Sampling and Training Scheme for Conditional Diffusion Generation |
0 of 5 row(s) selected.