Image Reconstruction On Imagenet
평가 지표
FID
LPIPS
PSNR
SSIM
평가 결과
이 벤치마크에서 각 모델의 성능 결과
비교 표
| 모델 이름 | FID | LPIPS | PSNR | SSIM |
|---|---|---|---|---|
| taming-transformers-for-high-resolution-image | 3.64 | 0.177 | 19.93 | 0.542 |
| open-magvit2-an-open-source-project-toward | 1.17 | - | 21.90 | - |
| an-image-is-worth-32-tokens-for | 1.71 | - | - | - |
| scaling-the-codebook-size-of-vqgan-to-100000 | 2.62 | 0.120 | 23.80 | 0.589 |
| preventing-local-pitfalls-in-vector | 0.91 | 0.066 | 27.57 | 0.729 |
| maskbit-embedding-free-image-generation-via | 1.66 | - | - | - |
| taming-scalable-visual-tokenizer-for | 1.00 | 0.2030 | - | - |
| vector-quantized-image-modeling-with-improved-1 | 1.28 | - | - | - |
| maskgit-masked-generative-image-transformer | 2.28 | - | - | - |
| autoregressive-image-generation-using | 1.83 | - | - | - |
| preventing-local-pitfalls-in-vector | 1.00 | 0.076 | 26.59 | 0.717 |
| movq-modulating-quantized-vectors-for-high | 1.12 | 0.113 | 22.42 | 0.673 |