[NN] Inference Time of Object Detection ***

 

 

 

Ref: https://coral.ai/docs/edgetpu/benchmarks/

Model architecture	Desktop CPU 1	Desktop CPU 1 + USB Accelerator (USB 3.0) with Edge TPU	Embedded CPU 2	Dev Board 3 with Edge TPU

 

 

 

Ref: Why we don’t use Lambda for serverless machine learning

Lambda doesn’t support GPU inference

 

 

训练比较重要，尤其是多GPU快速训练。