Research map / Image generation / GANs
Sliced Wasserstein Generative Models
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Replace the WD objective with the proposed dual SWD blocks | SWGAN · CIFAR-10 | FID: 17.6 ± 0.7 -> 17.0 ± 1.0 | Helped |
| Replace the WD objective with the proposed dual SWD blocks | SWGAN · CelebA | FID: 15.8 ± 0.6 -> 13.2 ± 0.7 | Helped |
| Replace the WD objective with the proposed dual SWD blocks | SWGAN · LSUN | FID: 19.5 ± 1.2 -> 14.9 ± 1.0 | Helped |
| Use the primal SWD blocks rather than adversarial regularization | SWAE · CelebA | FID: 49.3±5.8 -> 48.9±4.3 | No clear effect |
| Use the progressive SWGAN objective instead of the WGAN objective | Progressive GAN · CelebA-HQ | FID: 7.5 -> 5.5 | Helped |
| Use the progressive SWGAN objective instead of the WGAN objective | Progressive GAN · LSUN | FID: 8.4 -> 8.0 | Helped |
| Use the progressive SWGAN objective instead of the WGAN objective | Progressive 3D GAN · TrailerFaces | video FID: 462.6 -> 404.1 | Helped |
| Use ConvNet without normalization rather than ResNet with normalization (ablation) | SWAE · CelebA | FID: 63.2 -> 48.6 | Helped |
| Use ConvNet with normalization rather than ResNet with normalization (ablation) | SWGAN · CelebA | FID: 13.0 -> 19.2 | Hurt |
| Use ConvNet without normalization rather than ResNet with normalization (ablation) | SWGAN · CelebA | FID: 13.0 -> 18.8 | Hurt |
Closest papers by meaning
- Generative Modeling Using the Sliced Wasserstein Distance (CVPR 2018 · similarity 0.92)
- Asymptotic Guarantees for Learning Generative Models with the Sliced-Wasserstein Distance (NeurIPS 2019 · similarity 0.91)
- Max-Sliced Wasserstein Distance and Its Use for GANs (CVPR 2019 · similarity 0.91)
- Distributional Sliced-Wasserstein and Applications to Generative Modeling (ICLR 2021 · similarity 0.91)
- Banach Wasserstein GAN (NeurIPS 2018 · similarity 0.89)
- Sliced Wasserstein Estimation with Control Variates (ICLR 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.