Traditional methods for compressing neural networks (NN), such as pruning and distillation, primarily focus on reducing the size of the NN, often at the expense of accuracy. To overcome this limitation, we propose a novel refining NN method based on symbolic r... ...

请注册登录后继续浏览