Member-only story

Day 11: 2020.04.23
Paper: Identity Mappings in Deep Residual Networks
Category: Model/CNN/Deep Learning/Image Recognition

A follow-up on the paper yesterday.

In this paper, the authors tested multiple adjustment on the original residual network idea. It turns out that the “full pre-activation” one is with the quickest learning rate and the lowest error rate.

--

--

Chun-kit Ho
Chun-kit Ho

Written by Chun-kit Ho

cloud architect@ey | full-stack software engineer | social innovation | certified professional solutions architect in aws & gcp

No responses yet