cvlab-yonsei/az-nas

An official implementation of "AZ-NAS: Assembling Zero-Cost Proxies for Network Architecture Search" (CVPR 2024) in PyTorch.

Pytorch implementation of AZ-NAS

This repository contains the official implementation of the paper "AZ-NAS: Assembling Zero-Cost Proxies for Network Architecture Search" presented at CVPR 2024.

For detailed information, please visit the project website or read the paper on arXiv.

Requirements

To reproduce our results, we provide Dockerfile to build a Docker image.

Refer to the Dockerfile for details about the environment setup.

Code

Instructions and example code for each search space can be found in the README file within the corresponding folder.

Acknowledgement

Our implementation is based on several open-source projects.

We express our gratitude to the authors and contributors of these projects.

If you use any of their assets, please cite the corresponding papers appropriately:

• NATS-Bench
• AutoDL-Projects
• ZenNAS
• ZiCo
• AutoFormer
• TFTAS

Citation

If you find our work useful in your research, please cite our paper:

@inproceedings{lee2024assembling,
  title={{AZ-NAS}: Assembling Zero-Cost Proxies for Network Architecture Search},
  author={Lee, Junghyup and Ham, Bumsub},
  booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},
  year={2024}
}

az-nas's Issues

Could you describe more details?

Hello, thank you for sharing your project.

Do you have a plan to update README file to explain your work?

Request for requirements.txt

Hi,

I noticed your project is related to the ZiCo paper, which I find intriguing. Could you share the requirements.txt to help set up the environment correctly? This would contribute to the project's accessibility and popularity.

Thanks a lot!