Neural network memory allocation method and device and electronic equipment

The embodiment of the invention provides a neural network memory allocation method and device and electronic equipment, and the method comprises the steps: determining a network layer meeting a reusable condition from other network layers in a neural network as a reusable network layer corresponding...

Full description

Saved in:
Bibliographic Details
Main Author HUANG CHONGJI
Format Patent
LanguageChinese
English
Published 29.09.2023
Subjects
Online AccessGet full text

Cover

Loading…
Abstract The embodiment of the invention provides a neural network memory allocation method and device and electronic equipment, and the method comprises the steps: determining a network layer meeting a reusable condition from other network layers in a neural network as a reusable network layer corresponding to a to-be-allocated network layer for each to-be-allocated network layer without memory allocation in the neural network; determining the size relationship between the reusable memory block and the required memory amount of the network layer to be distributed; if the reusable memory block is smaller than the required memory amount, adding a supplementary memory block of which the initial address is adjacent to the end address of the reusable memory block, and allocating the reusable memory block and the supplementary memory block to the network layer to be allocated; and if the reusable network layer does not exist in the allocated network layer, allocating the unallocated memory blocks which are not allocated to
AbstractList The embodiment of the invention provides a neural network memory allocation method and device and electronic equipment, and the method comprises the steps: determining a network layer meeting a reusable condition from other network layers in a neural network as a reusable network layer corresponding to a to-be-allocated network layer for each to-be-allocated network layer without memory allocation in the neural network; determining the size relationship between the reusable memory block and the required memory amount of the network layer to be distributed; if the reusable memory block is smaller than the required memory amount, adding a supplementary memory block of which the initial address is adjacent to the end address of the reusable memory block, and allocating the reusable memory block and the supplementary memory block to the network layer to be allocated; and if the reusable network layer does not exist in the allocated network layer, allocating the unallocated memory blocks which are not allocated to
Author HUANG CHONGJI
Author_xml – fullname: HUANG CHONGJI
BookMark eNrjYmDJy89L5WTw9kstLUrMUchLLSnPL8pWyE3NzS-qVEjMyclPTizJzM8DipRk5KcoJOalKKSklmUmp4KZqTmpySVF-XmZyQqphaWZBbmpeSU8DKxpiTnFqbxQmptB0c01xNlDN7UgPz61uCAxORVoTbyzn6GhmYWRoYGBiaMxMWoAtdI3hQ
ContentType Patent
DBID EVB
DatabaseName esp@cenet
DatabaseTitleList
Database_xml – sequence: 1
  dbid: EVB
  name: esp@cenet
  url: http://worldwide.espacenet.com/singleLineSearch?locale=en_EP
  sourceTypes: Open Access Repository
DeliveryMethod fulltext_linktorsrc
Discipline Medicine
Chemistry
Sciences
Physics
DocumentTitleAlternate 一种神经网络内存分配方法、装置及电子设备
ExternalDocumentID CN116821004A
GroupedDBID EVB
ID FETCH-epo_espacenet_CN116821004A3
IEDL.DBID EVB
IngestDate Fri Jul 19 13:02:28 EDT 2024
IsOpenAccess true
IsPeerReviewed false
IsScholarly false
Language Chinese
English
LinkModel DirectLink
MergedId FETCHMERGED-epo_espacenet_CN116821004A3
Notes Application Number: CN202310745461
OpenAccessLink https://worldwide.espacenet.com/publicationDetails/biblio?FT=D&date=20230929&DB=EPODOC&CC=CN&NR=116821004A
ParticipantIDs epo_espacenet_CN116821004A
PublicationCentury 2000
PublicationDate 20230929
PublicationDateYYYYMMDD 2023-09-29
PublicationDate_xml – month: 09
  year: 2023
  text: 20230929
  day: 29
PublicationDecade 2020
PublicationYear 2023
RelatedCompanies HANGZHOU HIKVISION DIGITAL TECHNOLOGY CO., LTD
RelatedCompanies_xml – name: HANGZHOU HIKVISION DIGITAL TECHNOLOGY CO., LTD
Score 3.626461
Snippet The embodiment of the invention provides a neural network memory allocation method and device and electronic equipment, and the method comprises the steps:...
SourceID epo
SourceType Open Access Repository
SubjectTerms CALCULATING
COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
COMPUTING
COUNTING
ELECTRIC DIGITAL DATA PROCESSING
PHYSICS
Title Neural network memory allocation method and device and electronic equipment
URI https://worldwide.espacenet.com/publicationDetails/biblio?FT=D&date=20230929&DB=EPODOC&locale=&CC=CN&NR=116821004A
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
link http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwfV3dS8MwED_m_HzTquj8IIL0rWjTL_NQxKUtQ7EbMmVvo19jE-2m7RD9671knfVF38IFQnLwy93lfrkDOI8TK8Jox9ZEkzTNNFJLixCQGh0ZmZPE6EPHkuUb2p1H83ZgDRrwvPwLI-uEfsjiiIioBPFeyvt6Vj9ieZJbWVzEExRNr4O-66lVdIz-NJp71Wu7fq_rdbnKuctDNXxwdd2-oqI62s0KrKIb7Qg0-E9t8Stl9tukBNuw1sPV8nIHGl9jBTb5svOaAhv3VcJbgXXJ0EwKFFYoLHbhTlTUiF5IvqBwk1fBlv0kIoW-2DtZ9IUmUZ6SNBN3gRzWLW9I9jafSKbQHpwFfp93NNzf8EcZQx7WRzH2oZlP8-wAiO04TpQymTo0E6rHWcrsEbukjKZOwqJDaP29Tuu_ySPYEooVHAnKjqFZvs-zEzTEZXwqNfgNZ26M9w
link.rule.ids 230,309,783,888,25576,76876
linkProvider European Patent Office
linkToHtml http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwfV3dS8MwED_m_JhvWhWdXxGkb0Wbra19KOLSjeq2bkiVvY1-DSfaTdsh-td7STvri76FC4Tk4Je7y_1yB3AehJqP0Y6u8CZpSrMRaYqPgFTopBEbYYA-dCBYvq7uPDTvRtqoAs_LvzCiTuiHKI6IiAoR75m4r-flI5YtuJXpRTBF0ey641m2XETH6E-juZftltUeDuwBkxmzmCu795aq6leUV0e7WYFVdLENjob2Y4v_Spn_NimdLVgb4mpJtg2VrycJamzZeU2CjX6R8JZgXTA0wxSFBQrTHejyihr-C0lyCjd55WzZT8JT6PneSd4XmvhJRKKY3wViWLa8IfHbYiqYQrtw1ml7zFFwf-MfZYyZWx6lsQfVZJbE-0B0wzD8yBSpw2ZI1SCOTH1iXlKTRkZo-gdQ_3ud-n-Tp1BzvH5v3Lt1u4ewyZXM-RLUPIJq9r6Ij9EoZ8GJ0OY3L8eP6g
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Apatent&rft.title=Neural+network+memory+allocation+method+and+device+and+electronic+equipment&rft.inventor=HUANG+CHONGJI&rft.date=2023-09-29&rft.externalDBID=A&rft.externalDocID=CN116821004A