Image Generation On Imagenet 256X256
Métriques
FID
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | FID |
|---|---|
| simple-diffusion-end-to-end-diffusion-for | 3.71 |
| relay-diffusion-unifying-diffusion-process-1 | 1.99 |
| Modèle 3 | 2.74 |
| randomized-autoregressive-visual-generation | 1.70 |
| randomized-autoregressive-visual-generation | 1.95 |
| randomized-autoregressive-visual-generation | 1.50 |
| cascaded-diffusion-models-for-high-fidelity | 4.88 |
| masked-diffusion-transformer-is-a-strong | 1.79 |
| an-image-is-worth-32-tokens-for | 1.97 |
| bigroc-boosting-image-generation-via-a-robust | 3.69 |
| refining-generative-process-with | 4.45 |
| scaling-up-gans-for-text-to-image-synthesis | 3.45 |
| maskgit-masked-generative-image-transformer | 6.18 |
| representation-alignment-for-generation | 1.42 |
| refining-generative-process-with | 1.83 |
| diffusion-models-without-classifier-free-1 | 1.34 |
| beyond-next-token-next-x-prediction-for | 1.28 |
| entropy-driven-sampling-and-training-scheme | 3.96 |
| stylegan-xl-scaling-stylegan-to-large-diverse | 2.30 |
| autoregressive-image-generation-using | 3.83 |
| diffusion-models-beat-gans-on-image-synthesis | 4.59 |
| flowar-scale-wise-autoregressive-image | 1.65 |
| beyond-next-token-next-x-prediction-for | 1.24 |
| scalable-adaptive-computation-for-iterative | 4.51 |
| acdit-interpolating-autoregressive | 2.37 |
| scalable-diffusion-models-with-transformers | 2.27 |
| entropy-driven-sampling-and-training-scheme | 4.09 |
| maskgit-masked-generative-image-transformer | 4.02 |
| stabilize-the-latent-space-for-image | 3.39 |
| robust-latent-matters-boosting-image | 1.60 |
| diffit-diffusion-vision-transformers-for | 1.73 |
| elucidating-the-design-space-of-language | 1.54 |
| fasterdit-towards-faster-diffusion | 2.03 |
| diffusion-models-beat-gans-on-image-synthesis | 3.94 |
| autoregressive-image-generation-with | 2.1 |
| bigroc-boosting-image-generation-via-a-robust | 3.63 |
| efficient-diffusion-training-via-min-snr | 1.57 |
| generative-modeling-with-explicit-memory | 1.53 |
| autoregressive-image-generation-without | 2.31 |
| an-image-is-worth-32-tokens-for | 2.48 |
| reconstruction-vs-generation-taming-1 | 1.35 |
| improved-denoising-diffusion-probabilistic-1 | 12.3 |
| taming-transformers-for-high-resolution-image | 5.2 |
| adversarially-slicing-generative-networks | 2.14 |
| pagoda-progressive-growing-of-a-one-step | 1.56 |
| Modèle 46 | 4.29 |
| polarity-sampling-quality-and-diversity | 6.82 |
| simple-diffusion-end-to-end-diffusion-for | 3.75 |
| autoregressive-image-generation-without | 1.78 |
| alleviating-distortion-in-image-generation | 1.63 |
| visual-autoregressive-modeling-scalable-image | 1.73 |
| robust-latent-matters-boosting-image | 1.83 |
| autoregressive-image-generation-with | 2.44 |
| Modèle 54 | 11.84 |
| diffusion-models-need-visual-priors-for-image | 2.79 |
| large-scale-gan-training-for-high-fidelity | 8.1 |
| language-model-beats-diffusion-tokenizer-is | 3.65 |
| refining-generative-process-with | 3.18 |
| open-magvit2-an-open-source-project-toward | 2.33 |
| language-model-beats-diffusion-tokenizer-is | 1.78 |
| diffusion-models-need-visual-priors-for-image | 1.83 |
| alleviating-distortion-in-image-generation | 1.70 |
| cads-unleashing-the-diversity-of-diffusion | 1.70 |
| givt-generative-infinite-vocabulary | 2.59 |
| autoregressive-image-generation-without | 1.55 |
| learning-stackable-and-skippable-lego-bricks | 2.05 |
| autoregressive-image-generation-with | 1.94 |
| simpler-diffusion-sid2-1-5-fid-on-imagenet512 | 1.38 |
| generative-modeling-with-explicit-memory | 1.32 |
| masked-diffusion-transformer-is-a-strong | 1.58 |
| autoregressive-model-beats-diffusion-llama | 2.18 |
| draft-and-revise-effective-image-generation | 3.41 |
| self-conditioned-image-generation-via | 3.49 |
| flow-matching-in-latent-space | 4.46 |
| taming-transformers-for-high-resolution-image | 6.59 |
| polynomial-implicit-neural-representations | 2.86 |
| maskbit-embedding-free-image-generation-via | 1.52 |
| an-image-is-worth-32-tokens-for | 2.77 |
| randomized-autoregressive-visual-generation | 1.48 |