1 paper · 1 filter
Ejafa Bassam, Dawei Zhu, Kaigui Bian
Knowledge distillation is a model compression technique in which a compact "student" network is trained to replicate the predictive behavior of a larger "teacher" network. In logit…