Orbital edge computing: nano satellite constellations as a computer system

Re:Orbital edge computing: nano satellite constellations as ... by datameta on Monday October 12, 2020 @12:28PM (#24756134)

One can run astonishingly accurate near-SotA ML models on even the rad hardened space rated systems with scant resources. For example after 8-bit quantization, pruning, and compression a wake-word NN can take just 14KB. An RPi tier SBC can do quite a bit of inferencing!

[ Parent | on HN ]

Re:Orbital edge computing: nano satellite constellations as ... by d_silin on Monday October 12, 2020 @02:04PM (#24757405)

Hey, do you have any sources for that to read?

[ Parent | on HN ]

Re:Orbital edge computing: nano satellite constellations as ... by datameta on Monday October 12, 2020 @02:58PM (#24758054)

Sure. I saw talks by NXP and Renesas at the ARM Dev Summit last week that demonstrated the use of the Glow ML compiler [0] and CMSIS-NN Library [1] to squeeze a Wake Word NN [2] from 20KB to 14KB (with much greater reduction possible on larger ML models). Both Tensorflow Lite and Pytorch is supported in this toolchain.

[0] https://github.com/pytorch/glow

[1] https://arm-software.github.io/CMSIS_5/NN/html/index.html

[2] https://github.com/tensorflow/tensorflow/tree/master/tensorf...

[ Parent | on HN ]

Re:Orbital edge computing: nano satellite constellations as ... by d_silin on Monday October 12, 2020 @03:04PM (#24758140)

Thanks, I’ll have a look. Didn’t realize ML models that small can work.

[ Parent | on HN ]

Re:Orbital edge computing: nano satellite constellations as ... by monocasa on Monday October 12, 2020 @04:40PM (#24759200)

TinyML is a decent book as an intro to ML models in that size range.

[ Parent | on HN ]