Generative Modeling Evaluation on ImageNet 256x256 (test)
40.55KGEL ET ScoreVQ-VAE2
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| VQ-VAE22023.06 | 40.55 | — | 2.933 | 2.73 | 5.851 | 5.723 | |
| VQGAN*VQGAN hyperparameters (k, t, p, a)=k = 600, t = 1.0, p = 0.922023.06 | 9.034 | 4.443 | 1.295 | 1.495 | 1.487 | 1.717 | |
| BIGGAN-DEEPTruncation parameter (tau)=0.62023.06 | 3.017 | 2.316 | 1.224 | 1.271 | 1.26 | 1.404 | |
| ADM2023.06 | 2.994 | 2.035 | 1.18 | 1.208 | 1.255 | 1.244 | |
| CDM2023.06 | 2.467 | 1.857 | 1.161 | 1.166 | 1.21 | 1.204 | |
| ADM-G (1.0)Classifier Guidance (G)=12023.06 | 2.289 | 1.786 | 1.155 | 1.151 | 1.185 | 1.188 | |
| VQGAN**VQGAN hyperparameters (k, t, p, a)=k = 600, a = 0.05, p = 1.02023.06 | 2.219 | 1.772 | 1.175 | 1.202 | 1.148 | 1.158 | |
| BIGGAN-DEEPTruncation parameter (tau)=12023.06 | 2.075 | 1.735 | 1.15 | 1.166 | 1.166 | 1.222 | |
| TRAINING SET2023.06 | 1.164 | 1.138 | 1.02 | 1.018 | 1.021 | 1.017 | |
| THEOR. OPT2023.06 | 1 | 1 | 1 | 1 | 1 | 1 |