-
531.发明申请公开(公告)号：US20210397791A1
公开(公告)日：2021-12-23
申请号：US17203680
申请日：2021-03-16
Inventor： Danxiang Zhu
IPC: G06F40/289 , G06K9/62 , G06F40/205 , G06N5/02
Abstract: The present disclosure provides a method for training language model, and associated apparatus, electronic device and readable storage medium, which relates to the technical field of deep learning and the technical field of natural language processing. A specific implementation solution is as follows: sampling a paragraph of text from each article in a plurality of articles respectively, to obtain multiple paragraphs of text; concatenating the multiple paragraphs of text to obtain a concatenated text; inputting the concatenated text into a language model, a prediction value of the number of articles being output via the language model; training the language model based on the actual number of articles in the plurality of articles and a prediction value of the number of articles, until a preset training completion condition is satisfied. In the present disclosure, the classification of the entire paragraph of text content by the language model may be implemented and the effect of recognizing the text content by the language model may be enhanced by training the language model using texts sampled from the plurality of articles.
-
公开(公告)号：US20210397772A1
公开(公告)日：2021-12-23
申请号：US17132929
申请日：2020-12-23
Inventor： Shusen Liu , Runyao Duan , Ningfeng Wang , Danxiang Wu
IPC: G06F30/3308 , G06N10/00
Abstract: A quantum circuit simulation method and device, an apparatus, and a storage medium are provided, which are related to a field of quantum simulation computation. The specific implementation is: obtaining, based on a quantum circuit containing n qubits, an n-order pure state corresponding to the quantum circuit; determining, based on the quantum circuit, a (k, k)-order gate tensor representing a quantum gate, on which a contraction processing is to be performed with the n-order state tensor; transforming the contraction processing between the n-order state tensor and the (k, k)-order gate tensor into a processing between matrices which can be expressed in a classic computer and reduce a computation amount in the classic computer, to obtain a processing result; and using the processing result as a result of the contraction processing between the n-order state tensor and the (k, k)-order gate tensor, to complete simulation of the quantum circuit.
-
公开(公告)号：US20210397646A1
公开(公告)日：2021-12-23
申请号：US17445000
申请日：2021-08-13
Inventor： Yixin SHI , Zihao PAN , Yuhan SHEN , Xueying LV , Baohua LAI , Zewu WU , Zeyu CHEN , Qiao ZHAO
IPC: G06F16/538 , G06F9/451
Abstract: The disclosure provides a method and an apparatus for processing visual data of a deep model, and an electronic device. The method includes: determining a user identifier in a request of displaying a workspace in response to acquiring the request; acquiring a set of historical visual data associated with the user identifier; and displaying the set of historical visual data in a page of a workspace of a user corresponding to the user identifier.
-
公开(公告)号：US20210397628A1
公开(公告)日：2021-12-23
申请号：US17112233
申请日：2020-12-04
Inventor： Yiyuan ZHANG , Juntao TONG
Abstract: The present disclosure discloses a method and apparatus for merging data of building blocks, a device and a storage medium, and relates to the field of electronic map and big data. A specific implementation comprises: acquiring data of building blocks of a target work region; dividing the target work region containing the data of building blocks into at least one target region based on a preset geographic isolation attribute; dividing each target region into at least one grid region based on a direction feature of a building block obtained by performing a direction feature extraction on the data of building blocks; and determining a relative positional relationship between corresponding building blocks based on data of building blocks in each grid region, and merging data of building blocks of at least two building blocks having a preset positional relationship.
-
535.发明申请公开(公告)号：US20210390943A1
公开(公告)日：2021-12-16
申请号：US17111238
申请日：2020-12-03
Inventor： Zhengkun GAO , Junteng ZHANG , Wenfu WANG , Tao SUN
IPC: G10L13/047 , G10L13/10 , G10L13/06
Abstract: The present disclosure discloses a method and apparatus for training a model, a method and apparatus for synthesizing a speech, a device and a storage medium, and relates to the field of natural language processing and deep learning technology. The method for training a model may include: determining a phoneme feature and a prosodic word boundary feature of sample text data; inserting a pause character into the phoneme feature according to the prosodic word boundary feature to obtain a combined feature of the sample text data; and training an initial speech synthesis model according to the combined feature of the sample text data, to obtain a target speech synthesis model.
-
Patent Agency Ranking
536.发明申请公开(公告)号：US20210390296A1
公开(公告)日：2021-12-16
申请号：US17199072
申请日：2021-03-11
Inventor： Mengyi En , Shanshan Liu , Xuan Li , Chengquan Zhang , Hailun Xu , Xiaoqiang Zhang
IPC: G06K9/00 , G06F40/279 , G06N3/08
Abstract: The present application discloses a method and an apparatus for optical character recognition, an electronic device and a storage medium, and relates to the fields of artificial intelligence and deep learning. The method may include: determining, for a to-be-recognized image, a text bounding box of a text area therein, and extracting a text area image from the to-be-recognized image according to the text bounding box; determining a bounding box of text lines in the text area image, and extracting a text-line image from the text area image according to the bounding box; and performing text sequence recognition on the text-line image, and obtaining a recognition result. The application of the solution in the present application can improve a recognition speed and the like.
-
公开(公告)号：US20210390124A1
公开(公告)日：2021-12-16
申请号：US17133886
申请日：2020-12-24
Inventor： Wei LIU
IPC: G06F16/29 , G06T11/60 , G06F16/9537 , G01C21/00 , G01C21/36
Abstract: The disclosure discloses a method for processing map data, a device, and a storage medium. The detailed implementing solution includes: determining POI data of bound points of interest (POI) of a road to be marked in response to obtaining a road marking instruction; selecting target POI data from the POI data of the bound POIs based on a scene where a user is located; and rendering a map interface based on road information of the road to be marked and the target POI data, and displaying the map interface rendered.
-
公开(公告)号：US20210390029A1
公开(公告)日：2021-12-16
申请号：US17397059
申请日：2021-08-09
Inventor： Yi ZHAO , Chaoping JI
Abstract: The present disclosure discloses a method and apparatus for testing a stress based on a cloud service, and relates to the field of cloud computing technology, and further to the field of cloud operation and maintenance technology. A particular implementation comprises: acquiring, based on first stress test information of a business system in a cloud service, a number of expected stress test nodes corresponding to the business system; creating edge computing nodes, a number of the edge computing nodes being identical to the number of the expected stress test nodes; and performing a stress test on the business system by using the edge computing nodes, to acquire second stress test information of the business system.
-
公开(公告)号：US20210383605A1
公开(公告)日：2021-12-09
申请号：US17412977
申请日：2021-08-26
Inventor： Haotian PENG , Ruizhi CHEN
Abstract: Provided are a driving method and apparatus of an avatar, a device and a medium. The method includes acquiring the skin weight of each skin vertex associated with the current bone node in the skinned mesh model of the avatar; acquiring target avatar data of the skinned mesh model when a picture to be converted is converted into the avatar; determining the bone driving factor of the skinned mesh model based on the skin weight, basic avatar data of the skinned mesh model and the target avatar data; and driving the skinned mesh model based on a bone driving factor of each bone node.
-
公开(公告)号：US20210383121A1
公开(公告)日：2021-12-09
申请号：US17115263
申请日：2020-12-08
Inventor： Chengxiang LIU , Hao LIU , Bolei HE
Abstract: A method for generating a tag of a video, an electronic device, and a storage medium are related to a field of natural language processing and deep learning technologies. The detailed implementing solution includes: obtaining multiple candidate tags and video information of the video; determining first correlation information between the video information and each of the multiple candidate tags; sorting the multiple candidate tags based on the first correlation information to obtain a sort result; and generating the tag of the video based on the sort result.
-
-
-
-
-
-
-
-
-
Search Results
1986
records
(took 0.043 seconds)
