Top-k Training of GANs: Improving GAN Performance by Throwing Away Bad Samples

Aug-15-2025, 15:54:15 GMT–Neural Information Processing Systems

We also apply our method to recent GAN variants and improve state-of-the-art FID for conditional generation from 9.21 to 8.57 on CIFAR-10.

arxiv preprint arxiv, generative adversarial network, generator, (12 more...)

Neural Information Processing Systems

Aug-15-2025, 15:54:15 GMT

Conferences PDF

Country:
- North America
  - Canada
    - Ontario > Toronto (0.14)
    - Quebec > Montreal (0.04)
  - United States > California
    - Orange County > Irvine (0.04)

Genre:
- Research Report > New Finding (0.46)

Technology:
- Information Technology > Artificial Intelligence
  - Machine Learning
    - Neural Networks (0.71)
    - Statistical Learning (0.95)
  - Vision (0.70)

Duplicate Docs Excel Report

Title
Top-k Trainingof GANs: Improving GAN Performanceby Throwing Away Bad Samples

Similar Docs Excel Report more

Title	Similarity	Source
None found