Support for event handling callback? #3198
Description
[FEATURE REQUEST]
On the STM32H7/RT1062, we can run huge models like YOLOV8-Small (10MB). However, they may take 10 seconds to run with TensorFlow. During this time, we need to process system events created during interrupts.
Could an event processing callback hook be added to the invoke loop that could be called after each operator?