Image Generation on Fashion-MNIST (test)
10.3FIDGLF
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| GLFperceptual loss=true2019.05 | 10.3 | — | — | — | — | |
| GLANNperceptual loss=true2019.05 | 13 | — | — | — | — | |
| GLF2019.05 | 21.3 | — | — | — | — | |
| RAE + GMM2019.05 | 25.1 | — | — | — | — | |
| Two-Stage VAE2019.05 | 26.1 | — | — | — | — | |
| WAE-GAN2019.05 | 31.5 | — | — | — | — | |
| VAEprior=flow2019.05 | 51.8 | — | — | — | — | |
| VAEposterior=flow2019.05 | 55.1 | — | — | — | — | |
| VAE2019.05 | 57.5 | — | — | — | — | |
| Lagging VAE2023.01 | — | 0.4 | 0.6 | 1.6 | -248.5 | |
| LIDGMVAE2023.01 | — | 1 | 1.6 | 2.6 | -242.3 | |
| SA-VAE2023.01 | — | 0.2 | 0.3 | 1.3 | -252.2 | |
| VAE2023.01 | — | 0.1 | 0.2 | 0.9 | -258.8 | |
| β-VAEbeta=0.22023.01 | — | 0.6 | 1.2 | 2.4 | -245.3 |