降低分布式训练通信的梯度稀疏压缩方法
陈世达,刘强,韩亮

Gradient sparsification compression approach to reducing communication in distributed training
Shi-da CHEN,Qiang LIU,Liang HAN
图 6 不同策略下模型学习曲线及top-1准确率对比
Fig.6 Comparison of model learning curve and top-1 accuracy under different strategies