SERF: efficient scheduling for fast deep neural network serving via judicious parallelism.

Feng Yan, Yuxiong He, Olatunji Ruwase, Evgenia Smirni

Browse the full SC paper archive.