NeuralCompression

What's New

• April 2023 (video compression) - Added a pytorch implementation of the video compression transformer (VCT)
• November 2022 (image compression) - Added Bits-Back coding with diffusion models!
• July 2021 (image compression) - Released implemenation of Scale Hyperprior
• July 2021 (video compression) - Released implementation of DVC

About

NeuralCompression is a Python repository dedicated to research of neural networks that compress data. The repository includes tools such as JAX-based entropy coders, image compression models, video compression models, and metrics for image and video evaluation.

NeuralCompression is alpha software. The project is under active development. The API will change as we make releases, potentially breaking backwards compatibility.

Installation

NeuralCompression is a project currently under development. You can install the repository in development mode.

PyPI Installation

First, install PyTorch according to the directions from the PyTorch website. Then, you should be able to run

pip install neuralcompression

to get the latest version from PyPI.

Development Installation

First, clone the repository and navigate to the NeuralCompression root directory and install the package in development mode by running:

pip install --editable ".[tests]"

If you are not interested in matching the test environment, then you can just apply pip install -e ..

Repository Structure

We use a 2-tier repository structure. The neuralcompression package contains a core set of tools for doing neural compression research. Code committed to the core package requires stricter linting, high code quality, and rigorous review. The projects folder contains code for reproducing papers and training baselines. Code in this folder is not linted aggressively, we don't enforce type annotations, and it's okay to omit unit tests.

The 2-tier structure enables rapid iteration and reproduction via code in projects that is built on a backbone of high-quality code in neuralcompression.

neuralcompression

• neuralcompression - base package
  • data - PyTorch data loaders for various data sets
  • entropy_coders - lossless compression algorithms in JAX
    • craystack - an implementation of the rANS algorithm with the craystack API
  • functional - methods for image warping, information cost, flop counting, etc.
  • layers - building blocks for compression models
  • metrics - torchmetrics classes for assessing model performance
  • models - complete compression models

projects

• projects - recipes and code for reproducing papers
  • scale_hyperprior_lightning Scale Hyperprior (Balle et al., 2018)
  • deep_video_compression DVC (Lu et al., 2019)

Tutorial Notebooks

This repository also features interactive notebooks detailing different parts of the package, which can be found in the tutorials directory. Existing tutorials are:

• Walkthrough of the neuralcompression flop counter (view on Colab).
• Using neuralcompression.metrics and torchmetrics to calculate rate-distortion curves (view on Colab).

Getting Started

For an example of package usage, see the Scale Hyperprior for an example of how to train an image compression model in PyTorch Lightning. See DVC for a video compression example.

Contributions

Please read our CONTRIBUTING guide and our CODE_OF_CONDUCT prior to submitting a pull request.

We test all pull requests. We rely on this for reviews, so please make sure any new code is tested. Tests for neuralcompression go in the tests folder in the root of the repository. Tests for individual projects go in those projects' own tests folder.

We use black for formatting, isort for import sorting, flake8 for linting, and mypy for type checking.

License

NeuralCompression is MIT licensed, as found in the LICENSE file.

Cite

If you find NeuralCompression useful in your work, feel free to cite

@misc{muckley2021neuralcompression,
    author={Matthew Muckley and Jordan Juravsky and Daniel Severo and Mannat Singh and Quentin Duval and Karen Ullrich},
    title={NeuralCompression},
    howpublished={\url{https://github.com/facebookresearch/NeuralCompression}},
    year={2021}
}