Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/agupt013/akt
https://github.com/agupt013/akt
Last synced: 2 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/agupt013/akt
- Owner: agupt013
- Created: 2020-09-25T03:19:06.000Z (over 4 years ago)
- Default Branch: master
- Last Pushed: 2022-03-17T22:48:04.000Z (almost 3 years ago)
- Last Synced: 2024-08-01T22:37:56.880Z (5 months ago)
- Language: Python
- Size: 49.8 KB
- Stars: 6
- Watchers: 3
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Adversarial Knowledge Transfer from Unlabeled Data
This repository is the official implementation our paper titled "[Adversarial Knowledge Transfer from Unlabeled Data](https://arxiv.org/abs/2008.05746)" accepted to [ACM-MM 2020](https://2020.acmmm.org/).
Please refer [project page](https://akashagupta.com/publication/acm2020_akt/project.html) for more details.
## Implementation Details
Our implementation is in PyTorch [1] with python 3.6.7. We train all our
model on GeForce RTX 2080 Ti GPUs. This implementation currently uses
one gpu and can be modified to use multiple gpus for larger batch size.
## Python Packages
Please refer to the requirements.txt file for all the packages we
used to create the environment for training our models. We create an
environment in anaconda.
## Datasets
This is a working code of our proposed method for PASCAL-VOC/ImageNet
experiment. We use PASCAL-VOC[2] dataset as the labeled target dataset
and ImageNet[3] as unlabeled source dataset.
## Usage
### To train a model on PASCAL-VOC and ImageNet experiment WITH GPU
python train.py --pascal_path \
--imgnet_path \
--gpu
python train.py --pascal_path /datasets/pascal-voc-2007/ \
--imgnet_path /datasets/imagenet-dataset/ \
--gpu 0
## To test your trained model WITH GPU
python train.py --pascal_path \
--model \
--gpu \
--test 1
python train.py --pascal_path /datasets/pascal-voc-2007/ \
--model ./checkpoints/best-model.pth \
--gpu 0 \
--test 1
### To train a model on PASCAL-VOC and ImageNet experiment WITHOUT GPU
python train.py --pascal_path \
--imgnet_path
python train.py --pascal_path /datasets/pascal-voc-2007/ \
--imgnet_path /datasets/imagenet-dataset/
### To test your trained model WITHOUT GPU
python train.py --pascal_path \
--model \
--test 1
python train.py --pascal_path /datasets/pascal-voc-2007/ \
--model ./checkpoints/best-model.pth \
--test 1
## Citation
```
@inproceedings{gupta2020adversarial,
title={Adversarial Knowledge Transfer from Unlabeled Data},
author={Gupta, Akash and Panda, Rameswar and Paul, Sujoy and Zhang, Jianming and Roy-Chowdhury, Amit K},
booktitle={Proceedings of the 28th ACM International Conference on Multimedia},
pages={2175--2183},
year={2020}
}
```
## Contact
Please contact the first author Akash Gupta ([[email protected]]([email protected])) for any questions.
## References
1. Paszke, Adam, Sam Gross, Soumith Chintala, Gregory Chanan, Edward
Yang, Zachary DeVito, Zeming Lin, Alban Desmaison, Luca Antiga,
and Adam Lerer. "Automatic differentiation in pytorch." (2017).
1. Everingham, Mark, Luc Van Gool, Christopher KI Williams, John Winn,
and Andrew Zisserman. "The pascal visual object classes (voc)
challenge." International journal of computer vision 88, no. 2
(2010): 303-338.
1. Deng, Jia, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li
Fei-Fei. "Imagenet: A large-scale hierarchical image database."
In 2009 IEEE conference on computer vision and pattern recognition,
pp. 248-255. Ieee, 2009.