Research map / Image generation / GANs
Online Multi-Granularity Distillation for GAN Compression
Yuxi Ren, Jie Wu, Xuefeng Xiao, Jianchao Yang
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 online distillation instead of offline two-stage (ablation) | OMGD (U-Net) · Cityscapes | mIoU: 45.76 -> 48.91 | Improved |
| use online distillation instead of offline two-stage (ablation) | OMGD (U-Net) · edges->shoes | FID: 26.19 -> 25.00 | Improved |
| use online distillation instead of offline two-stage (ablation) | OMGD (Res-Net) · horse->zebra | FID: 77.09 -> 51.92 | Improved |
| add deeper teacher to the wider teacher (ablation) | OMGD (U-Net) · edges->shoes | FID: 33.88 -> 25.00 | Improved |
| add deeper teacher to the wider teacher (ablation) | OMGD (U-Net) · Cityscapes | mIoU: 44.04 -> 48.91 | Improved |
| add channel distillation of intermediate layers (ablation) | OMGD (U-Net) · edges->shoes | FID: 26.62 -> 25.00 | Improved |
| add channel distillation of intermediate layers (ablation) | OMGD (U-Net) · Cityscapes | mIoU: 48.12 -> 48.91 | Improved |
| add channel distillation of intermediate layers (ablation) | OMGD (Res-Net) · horse->zebra | FID: 61.21 -> 51.92 | Improved |
| add channel distillation of intermediate layers (ablation) | OMGD (Res-Net) · summer->winter | FID: 75.47 -> 73.79 | Improved |
| train student with discriminator and GAN loss instead of discriminator-free (ablation) | OMGD (U-Net) · edges->shoes | FID: 25.00 -> 73.44 | Hurt |
| train student with discriminator and GAN loss instead of discriminator-free (ablation) | OMGD (U-Net) · Cityscapes | mIoU: 48.91 -> 31.79 | Hurt |
| add learn-best teacher selection strategy (ablation) | OMGD (Res-Net) · horse->zebra | FID: 63.15 -> 52.00 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- GAN-Compression — GAN Compression: Efficient Architectures for Interactive Conditional GANs (CVPR 2020) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Content-Aware GAN Compression (CVPR 2021 · similarity 0.92)
- AutoGAN-Distiller: Searching to Compress Generative Adversarial Networks (ICML 2020 · similarity 0.91)
- GAN Compression: Efficient Architectures for Interactive Conditional GANs (CVPR 2020 · similarity 0.91)
- Nickel and Diming Your GAN: A Dual-Method Approach to Enhancing GAN Efficiency via Knowledge Distillation (ECCV 2024 · similarity 0.91)
- Compressing Image-to-Image Translation GANs Using Local Density Structures on Their Learned Manifold (AAAI 2024 · similarity 0.88)
- E$^2$GAN: Efficient Training of Efficient GANs for Image-to-Image Translation (ICML 2024 · similarity 0.87)
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.
