TensorFlow Model Optimization Toolkit
A TensorFlow toolkit that optimizes Keras and TensorFlow models for deployment through techniques such as quantization and pruning for sparse weights.
TensorFlow Model Optimization Toolkit is a suite of tools, aimed at both novice and advanced users, for optimizing machine learning models for deployment and execution. Supported techniques include quantization and pruning for sparse weights, with APIs built specifically for Keras, and it is distributed as a set of stable Python packages installable alongside TensorFlow.
Different subpackages, covering clustering, quantization and sparsity, are maintained by TensorFlow's own optimization team and by Arm ML Tooling for the clustering component. The project's website hosts an overview of the individual tools, documented optimization gains, a roadmap, tutorials and API documentation, and contributions follow TensorFlow's standard contribution guidelines and code of conduct.