The TensorFlow Model Optimization Toolkit is a suite of tools that users, both novice and advanced, can use to optimize machine learning models for deployment and execution.
Supported techniques include quantization and pruning for sparse weights. There are APIs built specifically for Keras.
For an overview of this project and individual tools, the optimization gains, and our roadmap refer to tensorflow.org/model_optimization. The website also provides various tutorials and API docs.
The toolkit provides stable Python APIs.
For installation instructions, see tensorflow.org/model_optimization/guide/install.
If you want to contribute to TensorFlow Model Optimization, be sure to review the contribution guidelines. This project adheres to TensorFlow's code of conduct. By participating, you are expected to uphold this code.
We use GitHub issues for tracking requests and bugs.
| Subpackage | Maintainers |
|---|---|
| tfmot.clustering | Arm ML Tooling |
| tfmot.quantization | TensorFlow Model Optimization |
| tfmot.sparsity | TensorFlow Model Optimization |
As part of TensorFlow, we're committed to fostering an open and welcoming environment.
- TensorFlow Blog: Stay up to date on content from the TensorFlow team and best articles from the community.