-
531.发明申请公开(公告)号：US20220139096A1
公开(公告)日：2022-05-05
申请号：US17578735
申请日：2022-01-19
Inventor： Pengyuan Lv , Chengquan Zhang , Kun Yao , Junyu Han
Abstract: A character recognition method, a model training method, a related apparatus and an electronic device are provided. The specific solution is: obtaining a target picture; performing feature encoding on the target picture to obtain a visual feature of the target picture; performing feature mapping on the visual feature to obtain a first target feature of the target picture, where the first target feature is a feature that has a matching space with a feature of character semantic information of the target picture; inputting the first target feature into a character recognition model for character recognition to obtain a first character recognition result of the target picture.
-
公开(公告)号：US20220138093A1
公开(公告)日：2022-05-05
申请号：US17576157
申请日：2022-01-14
Inventor： Tao YANG
IPC: G06F11/36
Abstract: The present disclosure provides a method and apparatus for continuous integration testing, and relates to the fields of cloud computing and testing technology. A specific implementation comprises: acquiring a to-be-tested script, and acquiring a plurality of pieces of environment information required for testing; selecting a test device conforming to hardware parameter information in the environment information from a test device cluster; initializing the test device according to non-hardware parameter information in the environment information, to obtain an initialized test device; and testing the to-be-tested script using the initialized test device, to generate a test result corresponding to the environment information.
-
公开(公告)号：US20220129753A1
公开(公告)日：2022-04-28
申请号：US17572921
申请日：2022-01-11
Inventor： Yuxiang LU , Jiaxiang LIU , Xuyi CHEN , Shikun FENG , Shuohuan WANG , Yu SUN , Shiwei HUANG , Jingzhou HE
Abstract: A pre-training method of a neural network model, an electronic device, and a medium. The pre-training data is inputted to the initial neural network model, and the initial neural network model is pre-trained in the first training mode, in the first training mode, the plurality of hidden layers share one hidden layer parameter, and the loss value of the initial neural network model is obtained, if the loss value of the initial neural network model is less than a preset threshold, the initial neural network model continues to be pre-trained in the second training mode, in the second training mode, each of the plurality of hidden layers has its own hidden layer parameter.
-
公开(公告)号：US20220113943A1
公开(公告)日：2022-04-14
申请号：US17455100
申请日：2021-11-16
Inventor： Guanglai Deng , Chao Tian
Abstract: The disclosure provides a method for multiply-add operations for a neural network. The method includes: determining types of respective pieces of data to be calculated based on a multiply-add operation request; in a condition of the type of each piece of the data to be calculated is a type of single-precision floating point, compressing mantissa of each pieces of the data to be calculated to obtain each compressed mantissa; splitting each compressed mantissa according to a preset rule and determining high digits and low digits of the compressed mantissa; and performing a multiply-add operation on each compressed mantissa based on the high digits and low digits of the compressed mantissa.
-
公开(公告)号：US11300966B2
公开(公告)日：2022-04-12
申请号：US16708514
申请日：2019-12-10
Inventor： Fan Yang , Xiaoxing Zhu , Xiang Liu , Shuang Zhang , Jingjing Xue
Abstract: The present disclosure provides a vehicle control method, including: acquiring current driving information of a vehicle; determining, according to the current driving information and a preset risk map, whether the vehicle drives to a risky road segment; and controlling a driving state of the vehicle accordingly when it is determined that the vehicle drives to the risky road segment. By presetting a risk map, defining road and roadside risky areas in the map, defining risky road segments of roads according to the road and roadside risky areas, determining in real time, whether the vehicle drives to a risky road segment, and controlling a driving state of the vehicle when it is determined that the vehicle drives to the risky road segment, so as to avoid risks that the risky areas cause to the driving of the vehicle, the driving safety of the vehicle can be effectively improved.
-
Patent Agency Ranking
公开(公告)号：US20220107949A1
公开(公告)日：2022-04-07
申请号：US17553887
申请日：2021-12-17
Inventor： Zhen ZHEN , Zunyao XUE , Jiajie CHEN , Zhixun ZHENG , Biao FAN
IPC: G06F16/2453
Abstract: The present disclosure provides a method of optimizing a search system, which relates to a field of data processing, and in particular to a field of data search. The method is implemented to include: determining a first hit rate of a cache unit of the search system for a plurality of user queries; for each element in a first set of elements, determining at least one key element by: generating a plurality of first queries that correspond to the plurality of user queries; determining a second hit rate of the offline cache unit for the plurality of first queries; and determining the element as one of at least one key element in response to determining that a difference between the second hit rate and the first hit rate is less than a difference threshold; and optimizing the search system based on the at least one key element.
-
公开(公告)号：US20220103782A1
公开(公告)日：2022-03-31
申请号：US17643687
申请日：2021-12-10
Inventor： Chao Li
Abstract: The disclosure discloses a method for video frame interpolation. The method includes: obtaining a first visual semantic feature and first pixel information of a first frame, and obtaining a second visual semantic feature and second pixel information of a second frame; generating semantic optical flow information based on the first visual semantic feature and the second visual semantic feature; generating pixel optical flow information based on the first pixel information and the second pixel information; and generating an interpolation frame between the first frame and the second frame based on the semantic optical flow information and the pixel optical flow information, and interpolating the interpolation frame between the first frame and the second frame.
-
公开(公告)号：US20220101642A1
公开(公告)日：2022-03-31
申请号：US17545765
申请日：2021-12-08
Inventor： Qunyi XIE , Yangliu XU , Xiameng QIN , Chengquan ZHANG
Abstract: The disclosure discloses a method for character recognition, an electronic device, and a storage medium. The technical solution includes: obtaining a test sample image and a test sample character both corresponding to a test task; performing fine-tuning on a trained meta-learning model based on the test sample image and the test sample character to obtain a test task model; obtaining a test image corresponding to the test task; and generating a test character corresponding to the test image by inputting the test image into the test task model.
-
公开(公告)号：US20220101199A1
公开(公告)日：2022-03-31
申请号：US17531132
申请日：2021-11-19
Inventor： Hao LIU , Weijia ZHANG , Dejing DOU , Hui XIONG
IPC: G06N20/00
Abstract: A training method for a point-of-interest recommendation model and a method for recommending a point of interest are provided. An implementation solution includes: obtaining training data including a plurality of point-of-interest recommendation requests; determining initialization parameters of the point-of-interest recommendation model; for a first point-of-interest recommendation request among the plurality of point-of-interest recommendation requests, determining a current return for the first point-of-interest recommendation request by utilizing the point-of-interest recommendation model, and determining, based on a second point-of-interest recommendation request initiated after the first point-of-interest recommendation request is completed, a target return for the first point-of-interest recommendation request by utilizing the point-of-interest recommendation model; and adjusting the initialization parameters of the point-of-interest recommendation model based on a difference between the current return and the target return for the first point-of-interest recommendation request, to obtain final parameters of the point-of-interest recommendation model.
-
公开(公告)号：US20220100777A1
公开(公告)日：2022-03-31
申请号：US17548680
申请日：2021-12-13
Inventor： Bo JING
Abstract: Provided are a transaction request construction method and apparatus, a transaction request processing method and apparatus, a device and a storage medium, which relate to the field of blockchain technology and can be used for cloud computing and cloud services. A specific implementation includes: generating a standard transaction request according to a standard key of a service application party, to-be-processed request data, a target blockchain architecture to be accessed and a target blockchain identifier; and calling a transaction conversion service and converting the standard transaction request into a target transaction request under the target blockchain architecture according to the standard key, the target blockchain architecture and the target blockchain identifier; where the target transaction request is used for processing the to-be-processed request data. Transaction request construction and processing efficiency can be improved.
-
-
-
-
-
-
-
-
-
Search Results
1522
records
(took 0.038 seconds)
