Research map / Image generation / GANs
Wasserstein GAN With Quadratic Transport Cost
Huidong Liu, Xianfeng Gu, Dimitris Samaras
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| use WGAN-QC instead of WGAN-div | WGAN · CelebA | FID: 15.2 -> 12.9 | Improved |
| use WGAN-QC instead of WGAN-div | WGAN · CelebA-HQ | FID: 13.5 -> 7.7 | Improved |
| use WGAN-QC instead of WGAN-div | WGAN · LSUN bedroom | FID: 15.9 -> 13.9 | Improved |
| increase gamma from 1e-4 to 1 (ablation) | WGAN-QC · CelebA | FID: 23.4 -> 11.5 | Improved |
| increase gamma from 1 to 10 (ablation) | WGAN-QC · CelebA | FID: 11.5 -> 15.7 | Hurt |
| use WGAN-QC instead of BigGAN (supervised) | unsupervised GAN · ImageNet dog | Inception Score: 10.39 -> 10.48 | Improved |
| use WGAN-QC instead of WGAN-div for speed | WGAN · CelebA | training time: 2.1 days -> 0.6 days | Improved |
| use WGAN-QC instead of WGAN-div for speed | WGAN · LSUN bedroom | training time: 2.4 days -> 1.3 days | Improved |
| use quadratic transport cost with OTR instead of WGAN-div regularization | WGAN · Dirac distribution | generator convergence | Improved |
Closest papers by meaning
- Wasserstein Divergence for GANs (ECCV 2018 · similarity 0.91)
- Kantorovich Strikes Back! Wasserstein GANs are not Optimal Transport? (NeurIPS 2022 · similarity 0.90)
- Improving the Improved Training of Wasserstein GANs: A Consistency Term and Its Dual Effect (ICLR 2018 · similarity 0.90)
- Do WGANs succeed because they minimize the Wasserstein Distance? Lessons from Discrete Generators (ICLR 2025 · similarity 0.89)
- A Two-Step Computation of the Exact GAN Wasserstein Distance (ICML 2018 · similarity 0.89)
- Some Theoretical Insights into Wasserstein GANs (JMLR 2021 · similarity 0.89)
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.
