首页> 外国专利> neural network quantization METHOD AND APPARATUS FOR NEURAL NETWORK QUANTIZATION

neural network quantization METHOD AND APPARATUS FOR NEURAL NETWORK QUANTIZATION

机译:神经网络量化的神经网络量化方法和装置

摘要

The present invention relates to a method and apparatus for neural network quantization. The method for neural network quantization includes the steps of: configuring a multidimensional vector representing a network parameter from a trained neural network model; obtaining a shared quantized vector as a cluster center by quantizing the multidimensional vector; finely controlling the shared quantized vector; and performing encoding by using the shared quantized vector. Accordingly, the present invention can perform deep neural networks.
机译:本发明涉及一种用于神经网络量化的方法和设备。用于神经网络量化的方法包括以下步骤:从训练的神经网络模型配置表示网络参数的多维矢量;通过对多维矢量进行量化,得到共享的量化矢量作为聚类中心;精细控制共享量化矢量;通过使用共享的量化矢量进行编码。因此,本发明可以执行深度神经网络。

著录项

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号