Abstract
By quantizing weights with different precision for different parts of a network, mixed-precision quantization promises to reduce the hardware cost and......
小提示：本篇文献需要登录阅读全文，点击跳转登录
Full Text Link
