公开(公告)号：US11531860B2

公开(公告)日：2022-12-20

申请号：US16174193

申请日：2018-10-29

申请人： Cambricon (Xi'an) Semiconductor Co., Ltd.

发明人： Qi Guo ,  Xunyu Chen ,  Yunji Chen ,  Tianshi Chen

IPC分类号： G06N3/04 ,  G06N3/06 ,  G06F7/50 ,  G06N3/063 ,  G06F7/501 ,  G06F9/30 ,  G06F7/523

摘要： Aspects for Long Short-Term Memory (LSTM) blocks in a recurrent neural network (RNN) are described herein. As an example, the aspects may include one or more slave computation modules, an interconnection unit, and a master computation module collectively configured to calculate an activated input gate value, an activated forget gate value, a current cell status of the current computation period, an activated output gate value, and a forward pass result.