Research map / Image generation / GANs
Slimmable Generative Adversarial Networks
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 |
|---|---|---|---|
| Replace independently trained width-specific generators with a slimmable generator | DCGAN · CIFAR-10 | FID at 0.25× width: 46.9 -> 37.3 | Helped |
| Replace independently trained width-specific generators with a slimmable generator | ResNet · STL-10 | FID at 0.25× width: 66.6 -> 69.1 | Hurt |
| Use sliceable conditional batch normalization instead of naive conditional batch normalization | Slimmable cGAN-pd · CIFAR-10 | FID at 0.25× width: 21.7 -> 19.5 | Helped |
| Use sliceable conditional batch normalization instead of naive conditional batch normalization | Slimmable cGAN-pd · CIFAR-100 | FID at 0.25× width: 26.8 -> 23.8 | Helped |
| Add stepwise inplace distillation (ablation) | SlimDCGAN · CIFAR-10 | mIC: 282.7 -> 231.3 | Helped |
| Add stepwise inplace distillation (ablation) | SlimResGAN · STL-10 | mIC: 342.4 -> 248.7 | Helped |
| Share partial parameters across discriminators (ablation) | Slimmable G · CIFAR-10 | FID AVG: 35.8 -> 30.8 | Helped |
| Add stepwise inplace distillation to shared-discriminator SlimGAN (ablation) | DCGAN · CIFAR-10 | FID AVG: 30.8 -> 29.2 | Helped |
| Use one discriminator for all generator widths (ablation) | Slimmable generator · CIFAR-10 | FID AVG: 30.8 -> 154.3 | Hurt |
| Make the discriminator slimmable rather than sharing partial discriminator parameters (ablation) | Slimmable generator · CIFAR-10 | FID AVG: 30.8 -> 35.9 | Hurt |
| Remove adversarial loss for narrow generators (ablation) | SlimGAN · CIFAR-10 | FID AVG: 29.2 -> 52.7 | Hurt |
| Distill all narrow generators directly from the widest generator rather than stepwise (ablation) | SlimGAN · CIFAR-10 | FID AVG: 29.2 -> 29.6 | No clear effect |
Closest papers by meaning
- Training Generative Adversarial Networks with Limited Data (NeurIPS 2020 · similarity 0.88)
- RareGAN: Generating Samples for Rare Classes (AAAI 2022 · similarity 0.88)
- Training Generative Adversarial Networks in One Stage (CVPR 2021 · similarity 0.88)
- Large Scale GAN Training for High Fidelity Natural Image Synthesis (ICLR 2019 · similarity 0.88)
- Diversity-Sensitive Conditional Generative Adversarial Networks (ICLR 2019 · similarity 0.88)
- SAN: Inducing Metrizability of GAN with Discriminative Normalized Linear Layer (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.