1 code implementation • 11 Sep 2017 • Xuan Peng, Xunzhang Gao, Xiang Li
To break this dependency between neighboring hidden units and speed up the convergence of training, a novel training strategy is proposed.
