Skip to main content

A package for building and training quantized and binary neural networks with Pytorch

Project description


BITorch is a library currently under development to simplify building quantized and binary neural networks with PyTorch. This is an early preview version of the library. If you wish to use it and encounter any problems, please create an issue. Our current roadmap contains:

  • Extending the model zoo with pre-trained models of state-of-the-art approaches
  • Adding examples for advanced training methods with multiple stages, knowledge distillation, etc.

All changes are tracked in the changelog.

Please refer to our wiki for a comprehensive introduction into the library or use the introduction notebook in examples/notebooks.


Similar to recent versions of torchvision, you should be using Python 3.8 or newer. Currently, the only supported installation is pip (a conda package is planned in the future).


If you wish to use a specific version of PyTorch for compatibility with certain devices or CUDA versions, we advise on installing the corresponding versions of pytorch and torchvision first (or afterwards), please consult pytorch's getting started guide.

Otherwise, simply run:

pip install bitorch

Note, that you can also request a specific PyTorch version directly, e.g. for CUDA 11.3:

pip install bitorch --extra-index-url

If you want to run the examples install the optional dependencies as well:

pip install "bitorch[opt]"

Local and Development Install Options

The package can also be installed locally for editing and development. First, clone the repository, then run:

pip install -e .         # without optional dependencies
pip install -e ".[opt]"  # with optional dependencies

Dali Preprocessing

If you want to use the Nvidia dali preprocessing library, e.g. with CUDA 11.x, (currently only supported for imagenet) you need to install the nvidia-dali-cuda110 package by running the following command:

pip install --extra-index-url --upgrade nvidia-dali-cuda110


Install the package and dev requirements locally for development:

pip install -e ".[dev]"


The tests can be run with pytest:


Code formatting and typing

For conveniently checking whether your code suites the required style (more details below), run


New code should be compatible with Python 3.X versions and be compliant with PEP8. To check the codebase, please run


The codebase has type annotations, please make sure to add type hints if required. We use mypy for type checking:

mypy --config-file mypy.ini

For code formatting we use black:

black . --check --verbose --diff --color  # check what changes the formatter would do
black .  # apply the formatter

In order to automatically apply the code formatting with every commit, you can also install pre-commit and use the pre-commit hook:

pre-commit install


We use Google's Python Docstring Format to document our code.

Documentation can be generated with

sphinx-build -b html docs/source/ docs/build/ -a

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

bitorch-0.3.0.tar.gz (114.5 kB view hashes)

Uploaded Source

Built Distribution

bitorch-0.3.0-py2.py3-none-any.whl (152.3 kB view hashes)

Uploaded Python 2 Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page