Image Generation on MNIST latent dimension 16 (test)
8.53FIDVAE - OURS
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| VAE - OURSSampling=Geometry-aware2022.09 | 8.53 | 98 | 97 | |
| RAE (GP)Latent Prior=10-component GMM, Regularization=GP2022.09 | 9.44 | 97 | 98 | |
| AE - GMMLatent Prior=10-component GMM2022.09 | 9.6 | 95 | 90 | |
| RAE (L2)Latent Prior=10-component GMM, Regularization=L22022.09 | 9.89 | 97 | 98 | |
| RAE (SN)Latent Prior=10-component GMM, Regularization=Spectral2022.09 | 11.22 | 97 | 98 | |
| RAELatent Prior=10-component GMM2022.09 | 11.23 | 98 | 98 | |
| VAE - GMMLatent Prior=10-component GMM2022.09 | 13.13 | 95 | 92 | |
| HVAE2022.09 | 15.54 | 97 | 95 | |
| WAE2022.09 | 20.71 | 93 | 88 | |
| VAMP2022.09 | 34.02 | 83 | 88 | |
| RHVAE2022.09 | 36.51 | 73 | 28 | |
| VAE - N(0, 1)Latent Prior=N(0, 1)2022.09 | 40.7 | 83 | 75 | |
| AE - N(0, 1)Latent Prior=N(0, 1)2022.09 | 46.41 | 86 | 77 |