Power efficient machine learning models deployment on edge iot devices

HIGHLIGHTS

SUMMARY

This led to the calculation of 79.37 uWh power consumption per inference and 18.82 uWh per inference calculation of 79.37 uWh power consumption per inference and 18.82 uWh per inference pure power consumption. pure power consumption. The per inference pure power consumption SRAM Pure was more than 100 times bigger (0.14 uWh/inference toTotal 18.82 uWh/inference) and almost Virtual Inference Inference NoOp (Tensor Model Power Per Inference 80 times bigger asPower a total power per inference. Comparing results with the results the uncompressed LeNet-5 model used . . .

 



If you want to have access to all the content you need to log in!

Thanks :)

Username or Email

Password

Remember me

Lost your password?

If you don't have an account, you can create one here.