hey,
What kind of inference framework you use for your demo of “object” detection video ?
for the edge 1S pro that ship in January, is tenser flow will work at 3 TOPS without major bugs? (I mean if other stuff are not working beside NPU and ethernet communication I don’t care… for a prototype)
Does the performance of the NPU/ H264 decoder are better on android or linux on this board?
thanks in advance.