Foldiak model

Reference Foliak model implementation for learning implemented in PyTorch with GPU support. From the paper:

@article{foldiak1990forming,
  title={Forming sparse representations by local anti-Hebbian learning},
  author={F{\"o}ldiak, Peter},
  journal={Biological cybernetics},
  volume={64},
  number={2},
  pages={165--170},
  year={1990},
  publisher={Springer}
}

Setup

1. Clone the repo.
2. Navigate to the directory containing the repo directory.
3. Run pip install -e foldiak
4. Navigate into the repo and install the requirements using pip install -r requirements.txt
5. Try running the demo notebook: examples/bars_example.ipynb

Note: To run the notebook you need the sparsecoding library installed. Its great and definitely worth downloading.

Note: If you are using a jupyter notebook, you will need to restart the jupyter kernel each time you change a source file.