公开(公告)号：US11775803B2

公开(公告)日：2023-10-03

申请号：US18012938

申请日：2021-04-26

申请人： INSPUR SUZHOU INTELLIGENT TECHNOLOGY CO., LTD.

发明人： Haiwei Liu ,  Gang Dong ,  Yaqian Zhao ,  Rengang Li ,  Dongdong Jiang ,  Hongbin Yang ,  Lingyan Liang

IPC分类号： G06F12/10 ,  G06N3/08 ,  G06N3/0442 ,  G06F17/16

CPC分类号： G06N3/0442 ,  G06F17/16

摘要： A system for accelerating an RNN network including: a first cache, which is used for outputting Wx1 to WxN or Wh1 to WhN in parallel in N paths in a cyclic switching manner, and the degree of parallelism is k; a second cache, which is used for outputting xt or ht-1 in the cyclic switching manner; a vector multiplication circuit, which is used for, by using N groups of multiplication arrays, respectively calculating Wx1xt to WxNxt, or respectively calculating Wh1ht-1 to WhNht-1; an addition circuit, which is used for calculating Wx1xt+Wh1ht-1+b1 to WxNxt+WhNht-1+bN; an activation circuit, which is used for performing an activation operation according to an output of the addition circuit; a state updating circuit, which is used for acquiring ct-1, calculating ct and ht, updating ct-1, and sending ht to the second cache; a bias data cache; a vector cache; and a cell state cache.

公开(公告)号：US11748970B2

公开(公告)日：2023-09-05

申请号：US17794110

申请日：2020-11-16

申请人： INSPUR SUZHOU INTELLIGENT TECHNOLOGY CO., LTD.

发明人： Qichun Cao ,  Yaqian Zhao ,  Gang Dong ,  Lingyan Liang ,  Wenfeng Yin

IPC分类号： G06V10/00 ,  G06V10/28

CPC分类号： G06V10/28

摘要： A hardware environment-based data quantization method includes: parsing a model file under a current deep learning framework to obtain intermediate computational graph data and weight data that are independent of a hardware environment; performing calculation on image data in an input data set through a process indicated by an intermediate computational graph to obtain feature map data; separately performing uniform quantization on the weight data and the feature map data of each layer according to a preset linear quantization method, and calculating a weight quantization factor and a feature map quantization factor (S103); combining the weight quantization factor and the feature map quantization factor to obtain a quantization parameter that makes hardware use shift instead of division; and finally, writing the quantization parameter and the quantized weight data to a bin file according to a hardware requirement so as to generate quantized file data (S105).

公开(公告)号：US12045729B2

公开(公告)日：2024-07-23

申请号：US18005620

申请日：2021-01-25

申请人： INSPUR SUZHOU INTELLIGENT TECHNOLOGY CO., LTD.

发明人： Wenfeng Yin ,  Gang Dong ,  Yaqian Zhao ,  Qichun Cao ,  Lingyan Liang ,  Haiwei Liu ,  Hongbin Yang

IPC分类号： G06N3/09 ,  G06N3/0895 ,  G06N3/0985

CPC分类号： G06N3/0985 ,  G06N3/0895

摘要： A neural network compression method whereby forward inference is performed on target data by using a target parameter sharing network to obtain an output feature map of the last convolutional module, a channel related feature is extracted from the output feature map, the extracted channel related feature and a target constraint condition are input into a target meta-generative network, and an optimal network architecture under the target constraint condition is predicted by using the target meta-generative network to obtain a compressed neural network model.