公开(公告)号：US20250103959A1

公开(公告)日：2025-03-27

申请号：US18885339

申请日：2024-09-13

Applicant: Beijing Baidu Netcom Science Technology Co., Ltd.

Inventor： Liang Shen ,  Dianhai Yu ,  Weibao Gong ,  Jinle Zeng ,  Haifeng Wang

IPC: G06N20/00

Abstract: Provided is a performance optimization method for a model training device, an electronic device, and a storage medium, relating to the fields of deep learning, large model training, and distributed parallel strategies. The method includes: determining communication timing of a current model training device with respect to a target model block at a target sorting position, so as to be able to perform synchronously collective communication with other model training devices of a plurality of model training devices with respect to model blocks at the target sorting position; and performing the collective communication on a backward gradient of the target model block at the communication timing.

公开(公告)号：US20250139327A1

公开(公告)日：2025-05-01

申请号：US18895722

申请日：2024-09-25

Applicant: BEIJING BAIDU NETCOM SCIENCE TECHNOLOGY CO., LTD.

Inventor： Liang Shen ,  Jinle Zeng ,  Hongxiang Hao ,  Weibao Gong ,  Dianhai Yu ,  Haifeng Wang

IPC: G06F30/20

Abstract: A method for processing a model operator includes: determining an operator set for model networking, wherein the operator set comprises a plurality of operators; determining a storage amount occupied by an output tensor of each operator in the operator set and a computation time period consumed in a forward computation of each operator in the operator set; and determining a first operator participating in recomputation in a model from the operator set, based on the storage amounts and the computation time periods of the plurality of operators.

公开(公告)号：US20250029010A1

公开(公告)日：2025-01-23

申请号：US18895264

申请日：2024-09-24

Applicant: BEIJING BAIDU NETCOM SCIENCE TECHNOLOGY CO., LTD.

Inventor： Dianhai Yu ,  Gexiao Tian ,  Weibao Gong ,  Haifeng Wang ,  Yongsheng Xu ,  Jiabin Yang

IPC: G06N20/00

Abstract: A cluster-based training method includes: in response to a hardware fault in the training node, selecting a target standby node from the plurality of standby nodes, and obtaining a target training snapshot of the model training task in the training node, in which the target training snapshot includes training state data of the model training task; and initializing the target standby node based on a container image of a model training program in the training node and the training state data to replace the training node with the target standby node to continue executing the model training task.