Neural Network Panning: Screening the Optimal Sparse Network Before Training

Xiatao Kang, Ping Li, Jiayi Yao +1

Pruning on neural networks before training not only compresses the original models, but also accelerates the network training phase, which has substantial application value. The cu…