Installing with pip
The Model Card Toolkit is hosted on PyPI, and requires Python 3.7 or later. Installing with pip is the best approach for most use cases.
Installing the basic, framework agnostic package:
pip install model-card-toolkit
If you are generating model cards for TensorFlow models, install the optional TensorFlow dependencies to use Model Card Toolkit's TensorFlow utilities:
pip install model-card-toolkit[tensorflow]
You may need to append the --use-deprecated=legacy-resolver flag when running
versions of pip starting with 20.3.
Installing from source
Installing from source is best if you would like to contribute code to the project or if you want to use the latest features. Follow the instructions in the contributing guide to learn how to set up a local development environment. Follow the steps below to install a wheel from source.
Before you can install Model Card Toolkit from source, you need to install Bazel>=2.0.0, which powers the protobuf stub code generation.
First, clone the GitHub repository:
git clone https://github.com/tensorflow/model-card-toolkit.git
cd model-card-toolkit
Build the pip package from source:
python setup.py sdist bdist_wheel
Finally, install your locally built package:
pip install --upgrade ./dist/model_card_toolkit-*.whl
Troubleshooting installation problems
If you run into issues with Model Card Toolkit installation, please
file an issue
with details on your operating system version, Python version, pip version, and
locally-installed packages. You can find your locally-installed packages with
pip freeze.