Machine Learning

meta-orchestraos-ml integrates the NXP machine-learning stack with OrchestraOS for supported i.MX8M Plus targets. Its reference configuration targets EIG-M and supports TensorFlow Lite inference through the OpenVX delegate, which provides access to the platform’s neural-processing hardware.

The layer contains:

• Integration with the NXP machine-learning package group.

• A reference image configuration including the ML stack and OpenCV.

• ONNX Runtime dependency and compiler compatibility patches.

• Package-selection adjustments for the reference ML image.

Hardware acceleration depends on the selected BSP, inference runtime, and model operators. Applications supply their models, preprocessing, and inference pipeline; the layer provides the runtime integration on which they build.

Dependencies, image composition, build instructions, and compatibility notes are maintained in README.md in the meta-orchestraos-ml repository. The layer is available under a separate commercial license; contact sales@cthings.co for access and licensing information.