litert-converter

What it is and what it does

litert-converter is a model conversion tool that transforms machine learning models into LiteRT format, the official runtime for running inference on mobile and embedded devices. It bridges the gap between model development (typically in TensorFlow or similar frameworks) and deployment on resource-constrained hardware like Android and iOS, handling the translation to a format optimized for low latency and small binary footprint.

The package depends on a substantial ML stack—numpy, protobuf, flatbuffers, lark, xdsl, and ml_dtypes—to parse, transform, and serialize models. It is actively maintained and recently released, with wheels pre-built for Python 3.10 through 3.14 on macOS arm64 and Linux x86_64, though this platform specificity means Windows users or those on other architectures will need to build from source or find alternative solutions.

Use it for:

• Convert a trained TensorFlow model to LiteRT format for deployment on an Android app.
• Optimize a neural network for inference on iOS devices with strict latency and memory constraints.
• Prepare a machine learning model for edge deployment on embedded Linux systems.
• Automate model conversion in a CI/CD pipeline targeting multiple mobile platforms.
• Reduce model binary size for distribution on devices with limited storage.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Converts machine learning models to LiteRT format for deployment on mobile and embedded devices with low latency and minimal binary size.

Yes, if you are targeting mobile or embedded device deployment and your model is compatible with LiteRT. The package is actively maintained, permissively licensed, and directly supported by the TensorFlow ecosystem. Install friction is moderate due to platform-specific wheels and a large dependency tree, but this is typical for ML tooling. Verify that your target platform (macOS arm64 or Linux x86_64) and Python version (3.10–3.14) are supported before committing.

Install

litert-converter on PyPI

pip install litert-converter

uv add litert-converter

poetry add litert-converter

Installing litert-converter

Before you install

Medium install friction due to compiled wheels for specific Python versions (3.10–3.14) and platforms (macOS arm64, Linux x86_64). Active maintenance with a recent release (4 days old). Nine runtime dependencies including numpy, protobuf, and xdsl add complexity but are standard ML tooling.

License in practice

Apache 2.0 permissive license allows commercial and private use with minimal restrictions; you must include a copy of the license and state significant changes.

Quickstart

pip install litert-converter

from litert_converter import convert

converted_model = convert(source_model_path, target_format='litert')

Requires Python 3.10–3.14 on macOS (arm64) or Linux (x86_64); no Windows wheels available in this release.

Verify before relying

• Whether the package supports model formats beyond TensorFlow Lite (e.g., ONNX, PyTorch).
• Whether conversion preserves model accuracy or applies quantization by default.
• API stability and backward compatibility guarantees across minor versions.
• Performance benchmarks or latency targets for converted models on target devices.

Package facts

Evidence: litert_converter-0.3.1-cp310-cp310-macosx_12_0_arm64.whl; litert_converter-0.3.1-cp310-cp310-manylinux_2_27_x86_64.whl; litert_converter-0.3.1-cp311-cp311-macosx_12_0_arm64.whl; litert_converter-0.3.1-cp311-cp311-manylinux_2_27_x86_64.whl; litert_converter-0.3.1-cp312-cp312-macosx_12_0_arm64.whl; litert_converter-0.3.1-cp312-cp312-manylinux_2_27_x86_64.whl; litert_converter-0.3.1-cp313-cp313-macosx_12_0_arm64.whl; litert_converter-0.3.1-cp313-cp313-manylinux_2_27_x86_64.whl; litert_converter-0.3.1-cp314-cp314-macosx_12_0_arm64.whl; litert_converter-0.3.1-cp314-cp314-manylinux_2_27_x86_64.whl

Keywords: litert, tflite, tensorflow, tensor, machine, learning

Tags

More Software Development packages