It is recommended to create a Python virtual environment before proceeding to the installation. Please see the TensorFlow installation guide for more information.
To install the latest version, run the following:
# Installing with the `--upgrade` flag ensures you'll get the latest version.
pip install --user --upgrade tensorflow-model-optimization
For release details, see our release notes.
For the required version of TensorFlow and other compatibility information, see the API Compatibility Matrix section of the Overview page for the technique you intend to use. For instance, for pruning, the Overview page is here.
Since TensorFlow is not included as a dependency of the TensorFlow Model
Optimization package (in
setup.py), you must explicitly install the TensorFlow
tf-nightly-gpu). This allows us to maintain one
package instead of separate packages for CPU and GPU-enabled TensorFlow.
Installing from Source
You can also install from source. This requires the Bazel build system.
# To install dependencies on Ubuntu:
# sudo apt-get install bazel git python-pip
# For other platforms, see Bazel docs above.
git clone https://github.com/tensorflow/model-optimization.git
bazel build --copt=-O3 --copt=-march=native :pip_pkg
pip install --user --upgrade $PKGDIR/*.whl