Overcomplete 0.0.2

Toolbox for experimenting with (Overcomplete) Dictionary Learning for Vision model

Overcomplete is a compact research library in Pytorch designed to study (Overcomplete)-Dictionary learning methods to extract concepts from large Vision models. In addition, this repository also introduces various visualization methods, attribution and metrics. However, Overcomplete emphasizes experimentation.

🚀 Getting Started with Overcomplete

Overcomplete requires Python 3.8 or newer and several dependencies, including Numpy. It supports both only Torch. Installation is straightforward with Pypi:

pip install overcomplete

With Overcomplete installed, you can dive into any optimisation based dictionary learning method to extract visual features. The API is designed to be intuitive, requiring only a few hyperparameters to get started.

Example usage:

import torch
import overcomplete
todo

Notebooks

• Starter Notebook:

Citation

@article{todo,
}

Authors

• Thomas Fel - [e-mail], PhD Student, Brown University & DEEL (ANITI)
• Remi Cadène - todo