FAQs - ST Neural ART NPU

Generic

Is it possible to deploy a model exclusively on the ARM Cortex® M55 core of an STM32N6xx device?

YES - in this case the ‘ –st-neural-art [ profile@conf.json]’ option should be not used. The classical flow for a STM32xx target can be used.

Is it possible to use directly the ST Neural ART compiler?

YES - When the intermediate ONNX and associated JSON files are generated, it is possible to directly call the ST Neural ART compiler.