Image Generation On Imagenet 512X512
Metrics
FID
Inception score
Results
Performance results of various models on this benchmark
Comparison Table
| Model Name | FID | Inception score |
|---|---|---|
| scalable-diffusion-models-with-transformers | 3.04 | 240.82 |
| autoregressive-image-generation-without | 1.73 | - |
| self-improving-diffusion-models-with | 1.73 | - |
| guiding-a-diffusion-model-with-a-bad-version | 1.34 | - |
| adversarial-score-identity-distillation | 2.06 | - |
| language-model-beats-diffusion-tokenizer-is | 3.07 | 213.1 |
| polynomial-implicit-neural-representations | 3.81 | - |
| adversarial-score-identity-distillation | 1.488 | - |
| pagoda-progressive-growing-of-a-one-step | 1.80 | - |
| adversarial-score-identity-distillation | 1.413 | - |
| adversarial-score-identity-distillation | 3.353 | - |
| maskgit-masked-generative-image-transformer | 4.46 | 342.0 |
| language-model-beats-diffusion-tokenizer-is | 1.91 | 324.3 |
| diffusion-models-beat-gans-on-image-synthesis | 7.72 | 172.71 |
| adversarial-score-identity-distillation | 1.379 | - |
| adversarial-score-identity-distillation | 2.707 | - |
| an-image-is-worth-32-tokens-for | 2.49 | - |
| discrete-predictor-corrector-diffusion-models | 3.54 | 350.2 |
| generative-modeling-with-explicit-memory | 1.71 | - |
| high-resolution-image-synthesis-with-latent | 3.60 | 247.67 |
| simple-diffusion-end-to-end-diffusion-for | 4.28 | 171 |
| analyzing-and-improving-the-training-dynamics | 1.85 | - |
| an-image-is-worth-32-tokens-for | 2.13 | - |
| adversarial-score-identity-distillation | 2.156 | - |
| applying-guidance-in-a-limited-interval | 1.68 | - |
| analyzing-and-improving-the-training-dynamics | 2.23 | - |
| simpler-diffusion-sid2-1-5-fid-on-imagenet512 | 1.48 | - |
| adversarial-score-identity-distillation | 1.969 | - |
| analyzing-and-improving-the-training-dynamics | 2.01 | - |
| adversarial-score-identity-distillation | 1.366 | - |
| cads-unleashing-the-diversity-of-diffusion | 2.31 | - |
| sa-solver-stochastic-adams-solver-for-fast | 2.80 | - |
| alleviating-distortion-in-image-generation | 2.89 | - |
| deep-compression-autoencoder-for-efficient | 1.72 | - |
| simple-diffusion-end-to-end-diffusion-for | 4.53 | 205.3 |
| analyzing-and-improving-the-training-dynamics | 2.91 | - |
| applying-guidance-in-a-limited-interval | 1.40 | - |
| givt-generative-infinite-vocabulary | 2.92 | - |
| adversarial-score-identity-distillation | 1.669 | - |
| analyzing-and-improving-the-training-dynamics | 1.81 | - |
| diffit-diffusion-vision-transformers-for | 2.67 | 252.12 |
| diffusion-models-beat-gans-on-image-synthesis | 3.85 | 221.72 |
| stylegan-xl-scaling-stylegan-to-large-diverse | 2.40 | - |
| adversarial-score-identity-distillation | 1.907 | - |
| adversarial-score-identity-distillation | 1.888 | - |
| maskgit-masked-generative-image-transformer | 7.32 | 156.0 |
| analyzing-and-improving-the-training-dynamics | 1.88 | - |
| guiding-a-diffusion-model-with-a-bad-version | 1.25 | - |