Research map / Label-efficient and robust learning / Recognition and long-tail
Feature Quantization Improves GAN Training
Read the paperCheck 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 |
|---|---|---|---|
| Add discriminator feature quantization | BigGAN · CIFAR-10 | FID: 6.04 -> 5.34 | Helped |
| Add discriminator feature quantization | TAC-GAN · CIFAR-100 | FID: ~7.22 -> 7.15 | Helped |
| Add discriminator feature quantization | BigGAN · CIFAR-100 | FID: 8.64 -> 7.36 | Helped |
| Add discriminator feature quantization | BigGAN · ImageNet-1000 64×64 | FID: 12.75 -> 12.62 | Helped |
| Add discriminator feature quantization | BigGAN · ImageNet-1000 128×128 | FID: ~22.77 -> ~19.11 | Helped |
| Add discriminator feature quantization | StyleGAN · FFHQ 32² | FID-50k: 3.28 -> 3.01 | Helped |
| Add discriminator feature quantization | StyleGAN · FFHQ 64² | FID-50k: 4.82 -> 4.36 | Helped |
| Add discriminator feature quantization | StyleGAN · FFHQ 128² | FID-50k: 6.33 -> 5.98 | Helped |
| Add discriminator feature quantization | StyleGAN · FFHQ 1024² | FID-50k: 5.24 -> 4.89 | Helped |
| Add discriminator feature quantization | StyleGAN2 · FFHQ | FID: 3.31 -> 3.19 | Helped |
| Add feature quantization in the generator (ablation) | GAN · CIFAR-100 | FID: 9.01{ ± .44} -> 8.96{ ± .26} | No clear effect |
| Use a smaller dictionary size (ablation) | FQ-GAN · CIFAR-100 | FID | Helped |
Closest papers by meaning
- Improving GAN with Neighbors Embedding and Gradient Matching (AAAI 2019 · similarity 0.88)
- Style Quantization for Data-Efficient GAN Training (CVPR 2025 · similarity 0.88)
- Ensembling Off-the-Shelf Models for GAN Training (CVPR 2022 · similarity 0.88)
- NICE: NoIse-modulated Consistency rEgularization for Data-Efficient GANs (NeurIPS 2023 · similarity 0.87)
- MCGAN: Enhancing GAN Training with Regression-Based Generator Loss (AAAI 2025 · similarity 0.87)
- SMaRt: Improving GANs with Score Matching Regularity (ICML 2024 · similarity 0.86)
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.