TernGrad: Ternary Gradients to Reduce Communication in Distributed Deep Learning

Devs

TernGrad: Ternary Gradients to Reduce Communication in Distributed Deep Learning | Read Paper on Bytez