公开(公告)号：US20210004625A1

公开(公告)日：2021-01-07

申请号：US17025419

申请日：2020-09-18

Applicant: Huawei Technologies Co., Ltd.

Inventor： Changzheng Zhang ,  Xin Jin ,  Dandan Tu

IPC: G06K9/32 ,  G06T7/11 ,  G06K9/62 ,  G06K9/46

Abstract: In an object detection model training method, a classifier that has been trained in a first phase is duplicated to at least two copies, and in a training in a second phase, each classifier obtained through duplication is configured to detect to-be-detected objects with different sizes, and train an object detection model based on a detection result.

公开(公告)号：US20210012136A1

公开(公告)日：2021-01-14

申请号：US17036903

申请日：2020-09-29

Applicant: Huawei Technologies Co., Ltd.

Inventor： Changzheng Zhang ,  Xin Jin ,  Dandan Tu

IPC: G06K9/32 ,  G06K9/62 ,  G06K9/46 ,  G06N3/04

Abstract: An object detection model training method performed by a computing device, includes obtaining a system parameter including at least one of a receptive field of a backbone network, a size of a training image, a size of a to-be-detected object in the training image, a training computing capability, or a complexity of the to-be-detected object, determining a configuration parameter based on the system parameter, establishing a variable convolution network based on the configuration parameter and a feature map of the backbone network, recognizing the to-be-detected object based on a feature of the variable convolution network, and training the backbone network and the variable convolution network, where a convolution core used by any variable convolution layer may be offset in any direction in a process of performing convolution.

公开(公告)号：US20190332944A1

公开(公告)日：2019-10-31

申请号：US16425012

申请日：2019-05-29

Applicant: Huawei Technologies Co., Ltd.

Inventor： Xiaolong Bai ,  Changzheng Zhang ,  Mingzhen Xia

IPC: G06N3/08 ,  G06N3/04

Abstract: A training method, apparatus, and chip for a neural network model includes determining a model training mode of each layer based on an estimated data volume in a model parameter set and an estimated data volume of output data of the layer, obtaining second output data that is obtained by m worker modules by training a (j−1)th layer, and directly obtaining by a worker module a global gradient of a model parameter by training the model parameter based on the second output data when a model parallel training mode is used for a jth layer.