be easily connected to the MCU. And sensor's data could be easily processed by MCU. But ML is computative-cost thing (as far as I know) so I have doubts if CortexM might handle it. Maybe series F7 or H7 would be enough. Anyway it depends on the model and how it should be handled. As I understand you have no details about nanoedge (what model they use, algorithms etc)?
I have more information and will send their documentation later as I'm at the moment on the road
Ml costly only on trained path. After trained computative cost may be reduced
Обсуждают сегодня