sparsity-in-binary-neural-nets

Drexel AI sparsity paper: https://arxiv.org/pdf/2101.06518.pdf

Datasets for Binary Classification https://jamesmccaffrey.wordpress.com/2018/03/14/datasets-for-binary-classification/

Development

• Install conda to your computer (once during first installation)
• Create an environment from environment.yml by
  conda env create -f environment.yml (only once during first installation)
• Activate the environment (every time you start developing) conda activate sparsity
• [OPTIONAL] If you need to install a new package, you can do pip install package.
• Finally, do conda deactivate when you are done.

Run

python main.py User arguments can be passed such as: python main.py -verbose=1 -batch_size=12 For a full list of arguments see, pipeline/argument_parser.py file.

If you end up using our work, please cite as below:

@article{alparslan2021evaluating,
  title={Evaluating Online and Offline Accuracy Traversal Algorithms for k-Complete Neural Network Architectures},
  author={Alparslan, Yigit and Moyer, Ethan Jacob and Kim, Edward},
  journal={arXiv preprint arXiv:2101.06518},
  year={2021}
}