Image Generation on CelebA 256x256
Metric: FID (lower is better)
Metric: FID (lower is better)
| # | Model↕ | FID▲ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | StyleSwin | 3.25 | No | StyleSwin: Transformer-based GAN for High-resolu... | 2021-12-20 | Code |
| 2 | HiT-B | 3.39 | No | Improved Transformer for High-Resolution GANs | 2021-06-14 | Code |
| 3 | LSGM | 7.22 | No | Score-based Generative Modeling in Latent Space | 2021-06-10 | Code |
| 4 | VQGAN | 10.2 | No | Taming Transformers for High-Resolution Image Sy... | 2020-12-17 | Code |
| 5 | StyleALAE | 19.21 | No | Adversarial Latent Autoencoders | 2020-04-09 | Code |
| 6 | NCP-VAE | 24.79 | No | A Contrastive Learning Approach for Training Var... | 2020-10-06 | - |
| 7 | MSP | 35 | No | Latent Space Factorisation and Manipulation via ... | 2019-07-26 | Code |
| 8 | GLF+perceptual loss (ours) | 41.8 | No | Generative Latent Flow | 2019-05-24 | Code |