Model Conversion
Stay organized with collections
Save and categorize content based on your preferences.
LiteRT requires models to be in a specialized format for on-device execution.
This step is commonly referred to as "conversion" and uses LiteRT conversion
components to convert from one model format to LiteRT's representation.
Multi-Framework Conversion Support
LiteRT supports converting models from PyTorch and TensorFlow.
See the following pages for more details:
Except as otherwise noted, the content of this page is licensed under the Creative Commons Attribution 4.0 License, and code samples are licensed under the Apache 2.0 License. For details, see the Google Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates.
Last updated 2026-06-02 UTC.
[null,null,["Last updated 2026-06-02 UTC."],[],[]]
