This is the codes of paper "EGANS: Evolutionary Generative Adversarial Network Search for Zero-Shot Learning" accepted to IEEE Transactions on Evolutionary Computation (TEC).
The code implementation of EGANS mainly based on PyTorch. All of our experiments run and test in Python 3.10.6.
We use CLSWGAN as the baseline to verify our method and conduct experiments on four popular ZSL benchmarks: CUB, SUN, FLO and AWA2 following the data split of xlsa17.
$ python clswgan_G_search.py
$ python clswgan_D_search.py
$ python clswgan_retrain.py
If this work is helpful for you, please cite our paper.
@ARTICLE{10225587,
author={Chen, Shiming and Chen, Shuhuang and Hou, Wenjin and Ding, Weiping and You, Xinge},
journal={IEEE Transactions on Evolutionary Computation},
title={EGANS: Evolutionary Generative Adversarial Network Search for Zero-Shot Learning},
year={2023},
volume={},
number={},
pages={1-1},
keywords={Computer architecture;Generators;Generative adversarial networks;Visualization;Training;Semantics;Optimization;Evolutionary neural architecture search;zero-shot learning;generative adversarial networks},
doi={10.1109/TEVC.2023.3307245}}
If you have any questions, please drop email to [email protected] or [email protected].
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent