Image Generation On Celeba Hq 256X256
評価指標
FID
評価結果
このベンチマークにおける各モデルのパフォーマンス結果
比較表
| モデル名 | FID |
|---|---|
| flow-matching-in-latent-space | 5.26 |
| dual-contradistinctive-generative-autoencoder-1 | 15.81 |
| score-matching-model-for-unbounded-data-score-1 | 7.16 |
| vaebm-a-symbiosis-between-variational | 20.38 |
| robust-diffusion-gan-using-semi-unbalanced | 5.6 |
| tackling-the-generative-learning-trilemma-1 | 7.64 |
| wavelet-diffusion-models-are-fast-and | 5.94 |
| high-resolution-image-synthesis-with-latent | 5.11 |
| ddmi-domain-agnostic-latent-diffusion-models | 8.73 |
| taming-transformers-for-high-resolution-image | 10.2 |
| bellman-optimal-step-size-straightening-of | - |
| relay-diffusion-unifying-diffusion-process-1 | 3.15 |
| learning-energy-based-model-via-dual-mcmc-1 | 15.89 |
| styleswin-transformer-based-gan-for-high-1 | 3.25 |
| score-based-generative-modeling-in-latent | 7.22 |
| learning-joint-latent-space-ebm-prior-model-1 | 9.89 |
| how-to-train-your-neural-ode | - |
| learning-latent-space-hierarchical-ebm | 8.78 |
| glow-generative-flow-with-invertible-1x1 | 68.93 |