code for "Learning to Learn by jointly optimizing neural architecture and weights" CVPR2022

1. dataset prepare

download datasets from mini-imagenet, omniglot. put the datasets under the datasets folder like:

datasets
    |
mini-imagenet - omniglot - FC100

2. init the envirnment

conda git branch -M mainenv create >f torch.yaml

3. use the code

cd scrips sh search_imagenet.sh

4. cite our work

@inproceedings{ding2022learning,
      title={Learning to Learn by Jointly Optimizing Neural Architecture and Weights}, 
      author={Ding, Yadong and Wu, Yu and Huang, Chengyue and Tang, Siliang and Yang, Yi and Wei, Longhui and Zhuang, Yueting and Tian, Qi},
      booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
      year={2022}
}

5. contact us

Yadong Ding: yadong97@outlook.com