Neural network training in a distributed system
Methods and systems for performing a training operation of a neural network are provided. In one example, a method comprises: performing backward propagation computations for a second layer of a neural network to generate second weight gradients; splitting the second weight gradients into portions;...
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Patent |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext bestellen |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | |
|---|---|
| container_issue | |
| container_start_page | |
| container_title | |
| container_volume | |
| creator | Hah, Thiam Khean Heaton, Richard John Vivekraja, Vignesh Huang, Randy Renfu Diamant, Ron |
| description | Methods and systems for performing a training operation of a neural network are provided. In one example, a method comprises: performing backward propagation computations for a second layer of a neural network to generate second weight gradients; splitting the second weight gradients into portions; causing a hardware interface to exchange a first portion of the second weight gradients with the second computer system; performing backward propagation computations for a first layer of the neural network to generate first weight gradients when the exchange of the first portion of the second weight gradients is underway, the first layer being a lower layer than the second layer in the neural network; causing the hardware interface to transmit the first weight gradients to the second computer system; and causing the hardware interface to transmit the remaining portions of the second weight gradients to the second computer system. |
| format | Patent |
| fullrecord | <record><control><sourceid>epo_EVB</sourceid><recordid>TN_cdi_epo_espacenet_US11941528B2</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>US11941528B2</sourcerecordid><originalsourceid>FETCH-epo_espacenet_US11941528B23</originalsourceid><addsrcrecordid>eNrjZND3Sy0tSsxRyEstKc8vylYoKUrMzMvMS1fIzFNIVEjJLC4pykwqLUlNUSiuLC5JzeVhYE1LzClO5YXS3AyKbq4hzh66qQX58anFBYnJqUCj4kODDQ0tTQxNjSycjIyJUQMAScArVQ</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>patent</recordtype></control><display><type>patent</type><title>Neural network training in a distributed system</title><source>esp@cenet</source><creator>Hah, Thiam Khean ; Heaton, Richard John ; Vivekraja, Vignesh ; Huang, Randy Renfu ; Diamant, Ron</creator><creatorcontrib>Hah, Thiam Khean ; Heaton, Richard John ; Vivekraja, Vignesh ; Huang, Randy Renfu ; Diamant, Ron</creatorcontrib><description>Methods and systems for performing a training operation of a neural network are provided. In one example, a method comprises: performing backward propagation computations for a second layer of a neural network to generate second weight gradients; splitting the second weight gradients into portions; causing a hardware interface to exchange a first portion of the second weight gradients with the second computer system; performing backward propagation computations for a first layer of the neural network to generate first weight gradients when the exchange of the first portion of the second weight gradients is underway, the first layer being a lower layer than the second layer in the neural network; causing the hardware interface to transmit the first weight gradients to the second computer system; and causing the hardware interface to transmit the remaining portions of the second weight gradients to the second computer system.</description><language>eng</language><subject>CALCULATING ; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS ; COMPUTING ; COUNTING ; PHYSICS</subject><creationdate>2024</creationdate><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&date=20240326&DB=EPODOC&CC=US&NR=11941528B2$$EHTML$$P50$$Gepo$$Hfree_for_read</linktohtml><link.rule.ids>230,308,776,881,25542,76289</link.rule.ids><linktorsrc>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&date=20240326&DB=EPODOC&CC=US&NR=11941528B2$$EView_record_in_European_Patent_Office$$FView_record_in_$$GEuropean_Patent_Office$$Hfree_for_read</linktorsrc></links><search><creatorcontrib>Hah, Thiam Khean</creatorcontrib><creatorcontrib>Heaton, Richard John</creatorcontrib><creatorcontrib>Vivekraja, Vignesh</creatorcontrib><creatorcontrib>Huang, Randy Renfu</creatorcontrib><creatorcontrib>Diamant, Ron</creatorcontrib><title>Neural network training in a distributed system</title><description>Methods and systems for performing a training operation of a neural network are provided. In one example, a method comprises: performing backward propagation computations for a second layer of a neural network to generate second weight gradients; splitting the second weight gradients into portions; causing a hardware interface to exchange a first portion of the second weight gradients with the second computer system; performing backward propagation computations for a first layer of the neural network to generate first weight gradients when the exchange of the first portion of the second weight gradients is underway, the first layer being a lower layer than the second layer in the neural network; causing the hardware interface to transmit the first weight gradients to the second computer system; and causing the hardware interface to transmit the remaining portions of the second weight gradients to the second computer system.</description><subject>CALCULATING</subject><subject>COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS</subject><subject>COMPUTING</subject><subject>COUNTING</subject><subject>PHYSICS</subject><fulltext>true</fulltext><rsrctype>patent</rsrctype><creationdate>2024</creationdate><recordtype>patent</recordtype><sourceid>EVB</sourceid><recordid>eNrjZND3Sy0tSsxRyEstKc8vylYoKUrMzMvMS1fIzFNIVEjJLC4pykwqLUlNUSiuLC5JzeVhYE1LzClO5YXS3AyKbq4hzh66qQX58anFBYnJqUCj4kODDQ0tTQxNjSycjIyJUQMAScArVQ</recordid><startdate>20240326</startdate><enddate>20240326</enddate><creator>Hah, Thiam Khean</creator><creator>Heaton, Richard John</creator><creator>Vivekraja, Vignesh</creator><creator>Huang, Randy Renfu</creator><creator>Diamant, Ron</creator><scope>EVB</scope></search><sort><creationdate>20240326</creationdate><title>Neural network training in a distributed system</title><author>Hah, Thiam Khean ; Heaton, Richard John ; Vivekraja, Vignesh ; Huang, Randy Renfu ; Diamant, Ron</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-epo_espacenet_US11941528B23</frbrgroupid><rsrctype>patents</rsrctype><prefilter>patents</prefilter><language>eng</language><creationdate>2024</creationdate><topic>CALCULATING</topic><topic>COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS</topic><topic>COMPUTING</topic><topic>COUNTING</topic><topic>PHYSICS</topic><toplevel>online_resources</toplevel><creatorcontrib>Hah, Thiam Khean</creatorcontrib><creatorcontrib>Heaton, Richard John</creatorcontrib><creatorcontrib>Vivekraja, Vignesh</creatorcontrib><creatorcontrib>Huang, Randy Renfu</creatorcontrib><creatorcontrib>Diamant, Ron</creatorcontrib><collection>esp@cenet</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Hah, Thiam Khean</au><au>Heaton, Richard John</au><au>Vivekraja, Vignesh</au><au>Huang, Randy Renfu</au><au>Diamant, Ron</au><format>patent</format><genre>patent</genre><ristype>GEN</ristype><title>Neural network training in a distributed system</title><date>2024-03-26</date><risdate>2024</risdate><abstract>Methods and systems for performing a training operation of a neural network are provided. In one example, a method comprises: performing backward propagation computations for a second layer of a neural network to generate second weight gradients; splitting the second weight gradients into portions; causing a hardware interface to exchange a first portion of the second weight gradients with the second computer system; performing backward propagation computations for a first layer of the neural network to generate first weight gradients when the exchange of the first portion of the second weight gradients is underway, the first layer being a lower layer than the second layer in the neural network; causing the hardware interface to transmit the first weight gradients to the second computer system; and causing the hardware interface to transmit the remaining portions of the second weight gradients to the second computer system.</abstract><oa>free_for_read</oa></addata></record> |
| fulltext | fulltext_linktorsrc |
| identifier | |
| ispartof | |
| issn | |
| language | eng |
| recordid | cdi_epo_espacenet_US11941528B2 |
| source | esp@cenet |
| subjects | CALCULATING COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS COMPUTING COUNTING PHYSICS |
| title | Neural network training in a distributed system |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-12T21%3A12%3A34IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-epo_EVB&rft_val_fmt=info:ofi/fmt:kev:mtx:patent&rft.genre=patent&rft.au=Hah,%20Thiam%20Khean&rft.date=2024-03-26&rft_id=info:doi/&rft_dat=%3Cepo_EVB%3EUS11941528B2%3C/epo_EVB%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true |