公开(公告)号：US20220253372A1

公开(公告)日：2022-08-11

申请号：US17627090

申请日：2020-10-28

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

Inventor： Yongkang XIE ,  Ruyue MA ,  Zhou XIN ,  Hao CAO ,  Kuan SHI ,  Yu ZHOU ,  Yashuai LI ,  En SHI ,  Zhiquan WU ,  Zihao PAN ,  Shupeng LI ,  Mingren HU ,  Tian WU

IPC: G06F11/36 ,  G06F8/35 ,  G06F8/41 ,  G06F8/20 ,  G06F9/445

Abstract: An apparatus and a method for executing a customized production line using an artificial intelligence development platform, a computing device and a computer readable storage medium are provided. The apparatus includes: a production line executor configured to generate a native form of the artificial intelligence development platform based on a file set, the native form to be sent to a client accessing the artificial intelligence development platform so as to present a native interactive page of the artificial intelligence development platform; and a standardized platform interface configured to provide an interaction channel between the production line executor and the artificial intelligence development platform. The production line executor is further configured to generate an intermediate result by executing processing logic defined in the file set and to process the intermediate result by interacting with the artificial intelligence development platform via the standardized platform interface.

公开(公告)号：US20220309395A1

公开(公告)日：2022-09-29

申请号：US17604670

申请日：2020-09-16

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

Inventor： Tuobang WU ,  En SHI ,  Yongkang XIE ,  Xiaoyu CHEN ,  Lianghuo ZHANG ,  Jie LIU ,  Binbin XU

IPC: G06N20/00 ,  G06F16/242 ,  G06F16/25

Abstract: The present disclosure discloses a method and an apparatus for adapting a deep learning model, an electronic device and a medium, which relates to technology fields of artificial intelligence, deep learning, and cloud computing. The specific implementation plan is: obtaining model information of an original deep learning model and hardware information of a target hardware to be adapted; querying a conversion path table according to the model information and the hardware information to obtain a matched target conversion path; and converting, according to the target conversion path, the original deep learning model to an intermediate deep learning model in the conversion path, and converting the intermediate deep learning model to the target deep learning model. Therefore, the deep learning model conversion is performed based on the model conversion path determined by the model information of the original deep learning model and the hardware information of the target hardware, which realizes converting any type of original deep learning model into the target deep learning model adapted to any target hardware, and solves the problem that the deep learning model is difficult to be applied to different hardware terminals.

公开(公告)号：US20210216805A1

公开(公告)日：2021-07-15

申请号：US17205773

申请日：2021-03-18

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

Inventor： Xiangxiang LV ,  En SHI ,  Yongkang XIE

IPC: G06K9/46 ,  G06N3/08 ,  G06N3/04 ,  G06T5/20

Abstract: The present application discloses an image recognition method, apparatus, an electronic device and a storage medium, and relates to the field of neural networks and depth learning. An implementation solution may be as follows: loading a first image recognition model; inputting an image to be recognized into a first image recognition model; predicting the image to be recognized by using a first image recognition model to obtain an output result of a network layer of the first image recognition model; and performing post-processing on the output result of the network layer of the first image recognition model, to obtain an image recognition result.

公开(公告)号：US20220067375A1

公开(公告)日：2022-03-03

申请号：US17200445

申请日：2021-03-12

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

Inventor： Penghao ZHAO ,  Haibin ZHANG ,  Shupeng LI ,  En SHI ,  Yongkang XIE

IPC: G06K9/00 ,  G06K9/46 ,  G06T3/40 ,  G06N20/00

Abstract: A method includes: determining at least one typical object ratio from a first training data set by counting ratios of objects in training pictures of the first training data set; determining at least one picture scaling size based at least on the at least one typical object ratio; scaling the training pictures of the first training data set according to the at least one picture scaling size; obtaining a second training data set by slicing the scaled training pictures; training an object detection model using the second training data set; and performing object detection on a to-be-detected picture using the trained object detection model. The object detection method according to the embodiments of the present disclosure can be used to complete, without manual intervention, a task of detecting an extremely small object.