Concurrent access to arm NPU?

ML Embedded Evaluation Kit
#1

Is concurrent access to ARM NPU: U55 supported?
Anything we need to be aware of?
Thanks

#2

@slhskgsdlfjslg you can run parallel inference with TFLM framework but each parallel inference will need a specific arena and require an RTOS. you also need to consider that if a network fully maps to ethos-u, the memory bandwidth might become a limiting factor in such parallel execution.