firmware-stm32
Installation
SKILL.md
Edge Impulse on STM32
Integrate the complete C++ library exported from the project's Deployment page. Do not substitute a generic model or copy only the TFLite file.
Inspect first
- Identify the STM32 part, CubeIDE/CubeMX version, RTOS use, sensor and bus, model input type, sample interval, and RAM/flash limits.
- Inspect the export for
edge-impulse-sdk/,model-parameters/,tflite-model/, and the generated inferencing header. - Inspect the existing project's source groups, include paths, linker script, cache setup, and RTOS heap before editing.
Integrate the library
- Copy the exported SDK, model parameters, and model sources into a dedicated project directory.
- Add all required C/C++ sources and include paths. Compile inference callers as C++.
- Include the generated project header when present; otherwise include:
#include "edge-impulse-sdk/classifier/ei_run_classifier.h"