thuml / Transferable-Adversarial-Training

Code release for Transferable Adversarial Training: A General Approach to Adapting Deep Classifiers (ICML2019)

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Transferable Adversarial Training: A General Approach to Adapting Deep Classifiers (ICML 2019)

Code release for Transferable Adversarial Training: A General Approach to Adapting Deep Classifiers (ICML 2019)

Dataset

VisDA-2017

Multi-Domain Sentiment

Requirements

  • python 2.7
  • PyTorch 0.4
  • Tensorflow >= 1.0
  • Tensorlayer >= 1.11
  • Tensorboard

Usage

  • download datasets
  • extract feature representations
  • train python XXX.py
  • monitor tensorboard --logdir .

Citation

please cite:

@InProceedings{TAT_2019_ICML,
author = {Liu, Hong and Long, Mingsheng and Wang, Jianmin and Jordan, Michael I.},
title = {Transferable Adversarial Training: A General Approach to Adapting Deep Classifiers},
booktitle = {Proceedings of the 36th International Conference on Machine Learning},
month = {June},
year = {2019}
}

Reference codes

https://github.com/thuml/easydl

Contact

About

Code release for Transferable Adversarial Training: A General Approach to Adapting Deep Classifiers (ICML2019)

License:MIT License


Languages

Language:Python 100.0%