Summary of Coronetgan: Controlled Pruning Of Gans Via Hypernetworks, by Aman Kumar et al.
CoroNetGAN: Controlled Pruning of GANs via Hypernetworks
by Aman Kumar, Khushboo Anand, Shubham Mandloi, Ashutosh Mishra, Avinash Thakur, Neeraj Kasera, Prathosh A P
First submitted to arxiv on: 13 Mar 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Image and Video Processing (eess.IV)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary In this paper, researchers tackle the challenge of deploying Generative Adversarial Networks (GANs) on resource-constrained edge devices. GANs have shown impressive performance in generative computer vision applications, but their large number of parameters poses a significant hurdle for deployment. The authors propose CoroNet-GAN, a novel approach that compresses GANs using differentiable pruning via hypernetworks, allowing for controllable compression and reduced training time. The method is evaluated on various conditional GAN architectures (Pix2Pix and CycleGAN) and benchmark datasets (Edges-to-Shoes, Horse-to-Zebra, and Summer-to-Winter). Results show that CoroNet-GAN outperforms baselines in terms of FID scores and achieves high-fidelity images across all datasets. Additionally, the approach demonstrates better inference time on various smartphone chipsets and data-types, making it a feasible solution for edge device deployment. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about finding ways to make computer vision models like GANs work on devices that don’t have a lot of power or storage. These models are good at generating new images or videos, but they’re very big and need a lot of resources to run. The researchers came up with a new way to shrink these models called CoroNet-GAN, which makes them smaller and faster while still keeping their good performance. They tested this method on different kinds of computer vision tasks and showed that it works better than other methods. |
Keywords
* Artificial intelligence * Gan * Inference * Pruning
