Does the EdgeAI module support running multiple models in parallel?
For example, an anomaly detector running on one data source and a classification model running on a second data source.
The documentation for nrf_edgeai_init says that the function "Set up the internal components of the Edge AI runtime" and "should be called first and once", but it takes a user model as a parameter.
If multiple models are supported, what are the intended function calls? Call the function once on each model, or don't call it on one model?
If multiple models are not supported, is this a known use-case that will be supported in the future?
If nrf_edgeai_init is only operating on the model data structure and not configuring any global/static state this would answer the question, but the comment about setting up the runtime indicates this is not the case. Since the module is provided as a static library, it is not possible to investigate the implementation myself.