Introduction

This is the source code of the paper entitled Accelerating Envolutionary Neural Architecture via Multi-Fidelity Evaluation

Notes

The source code inlcluding searching, training and best models found by MFENAS will be released once the paper is accepted.

-------------------------------------------- Current Version Notes------------------------------------------

The current version is just a baseline approach for reference, the detailed explanation and useage will be given when this paper is accepted. Of course, you can study this source code yourself to help your work.

To plot the architecture of found cells, you have to install pygraphviz package and use Plot_network function in utils.py.

Dependency Install

pytorch>=1.4.0
pygraphviz # used for plotting neural architectures

Usage

# Search process
python EMO.py

#Best solution
Baseline:  [[1, 1, 0, 1, 0, 0, 3, 1, 1, 0, 1, 1, 0, 0, 0, 1, 0, 6, 1, 1, 1, 0, 0, 1, 0, 1, 1, 0, 1, 1, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 3, 1, 0, 0, 0, 1, 0, 0, 1, 0, 8, 1, 0, 1, 0, 1, 0, 0, 1, 0, 0, 10, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 3], [1, 0, 8, 0, 1, 0, 9, 0, 1, 0, 0, 8, 1, 1, 0, 0, 0, 7, 1, 0, 0, 0, 0, 0, 8, 1, 1, 0, 0, 0, 1, 0, 3]], channel=46 for CIFAR-10/100

MFENAS: [[1, 1, 3, 1, 0, 0, 10, 1, 1, 1, 0, 3, 1, 0, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 7, 0, 1, 1, 0, 0, 0, 1, 11, 0, 1, 0, 0, 1, 0, 0, 0, 8, 1, 1, 0, 0, 0, 1, 0, 0, 0, 3, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 9, 1, 1, 0, 1, 0, 0, 0, 0, 1, 0, 0, 3, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 3, 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0], [0, 1, 8, 1, 1, 0, 6, 1, 1, 1, 0, 9, 0, 1, 1, 0, 0, 1, 1, 0, 1, 0, 1, 1, 6, 0, 0, 1, 0, 0, 1, 1, 7, 1, 1, 0, 0, 0, 0, 0, 0, 6, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 1, 0, 1, 1, 0, 1, 8, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 8, 0, 1, 1, 0, 1, 0, 1, 0, 0, 0, 0, 1, 5]],channel =44 for CIFAR-10/100



# Training process (mannually input solution and change the number of channels)
python train_cifar.py # validate on CIFAR-10 and CIFAR-100 datasets
python train_imagenet.py # validate on ImageNet dataset

# Plotting
Given a solution in EMO.py
the solution's normal cell and reduction cell can be plotted by
executing 'utils.Plot_network(solution.dag[0], path)' and 'utils.Plot_network(solution.dag[1], path)',
where 'path' is the path to save figures.

Citation

If you find this work helpful in your research, please use the following BibTex entry to cite our paper.

@article{yang2022accelerating,
  title={Accelerating Evolutionary Neural Architecture Search via Multi-Fidelity Evaluation},
  author={Yang, Shangshang and Tian, Ye and Xiang, Xiaoshu and Peng, Shichen and Zhang, Xingyi},
  journal={IEEE Transactions on Cognitive and Developmental Systems},
  year={2022}
}

Acknowledgement

Thanks for the help of NAO, NSGA-Net and ACE-NAS.