Mgbakwunye ahịrịokwu-BERT
Sentence-BERT (SBERT) na-emegharị BERT iji mepụta otu vector dị ogologo maka otu ahịrịokwu dum, yabụ enwere ike iji ihe ọ pụtara tụnyere myiri cosine ngwa ngwa.
Nchịkọta
It made semantic search and clustering over millions of sentences practical, turning a job that took BERT hours into milliseconds.
Ime miri emi
Plain BERT nwere ike iji ahịrịokwu abụọ atụnyere myirịta, mana ọ bụ naanị site n'inyekọta ha abụọ ọnụ site na netwọkụ, nke na-adị ngwa ngwa n'oke nha: iji ahịrịokwu 10,000 atụnyere ga-achọ ihe dị ka nde 50 ga-aga n'ihu. Sentence-BERT, nke Reimers na Gurevych webatara na 2019, na-edozi nke a site na iji netwọk siamese (ejima): ụlọ elu BERT abụọ nwere oke nha nke ọ bụla na-etinye otu ahịrịokwu n'adabereghị, mgbe ahụ, nkwụghachi ụgwọ (na-apụtakarị ịkwakọba ihe ntinye akara) na-enye otu vector kwa ahịrịokwu. A na-emegharị ihe nlereanya ahụ nke ọma ka ahịrịokwu ndị yiri ya na-adakọ ọnụ na oghere vector. Ugbu a, etinyere ahịrịokwu ọ bụla otu ugboro ka ọ bụrụ ntinye a na-emegharịgharị, na myirịta na-aghọ ngwaahịa ntụpọ dị ọnụ ala, na-eme ka ọchụchọ, mwepu, na nchịkọta n'ogo buru ibu.
Nghọta nka nka
A na-azụ SBERT ka ọ jiri ihe owuwu siamese na ebumnuche dị iche ma ọ bụ atọ. Ihe nrịbama asụsụ okike bụ ihe a na-ahụkarị: a na-ejikọta ụzọ abụọ na-atọ ụtọ ọnụ, na-emegiderịta onwe ha. Ụlọ elu abụọ ahụ na-ekekọrịta ihe ọ̀tụ̀tụ̀, ya mere ngbanwe bụ symmetric. Pụtara ịgbakọ n'elu vector nke ikpeazụ n'ozuzu na-aka nke ọma site na iji akara [CLS] naanị, na-emepụta ntinye ebe myirịta cosine na-adabere na nso nso a.
Mmetụta atụmatụ
Ọsọ na ọnụ ọgụgụ
Usoro ọrụ asụsụ nwere ike ịga ngwa ngwa n'achụghị nkwụsi ike.
Nweta na iru
Ọ na-agbasawanye ohere n'ofe asụsụ na ụdị nzikọrịta ozi.
Mkpebi doro anya
Ndị otu nwere ike itinyekwu oge na ikpe ebe akpaaka na-ejikwa nkwughachi.
Ọdịnihu nke Okwu-Mgbakwunye BERT
Ihe ngbanwe ụdị SBERT ugbu a na-akwado ọgbọ eweghachite-agbakwunyere, na-enye ọnọdụ dabara adaba na ụdị asụsụ buru ibu. Oghere ahụ na-aga n'ihu n'ụdị ntinye ntụziaka ka ukwuu, ihe ntinye asụsụ na ọtụtụ asụsụ, na ihe nnọchianya Matryoshka nke akụkụ ya nwere ike mezie maka ọsọ ọsọ. Ngwakọ pipeline na-ejikọ ngwa ngwa nnweta bi-encoder na ntinye ntinye koodu iji nwayọ nwayọ, na-ejikọta ọnụ ọgụgụ SBERT na nkenke dị elu na ndị ga-emeri elu.
Mmejuputa n'ezie n'ụwa
Igwe nchọta semantic tinyere ajụjụ na dọkụmentị niile wee weghachi vector kacha nso kama ịdabere na mkpirisi okwu.
Sistemu ọgbọ eweghachite na-eji ntinye SBERT weta akụkụ dị mkpa iji mebie azịza chatbot.
Ngwá ọrụ nkwado ndị ahịa na-achịkọta tiketi mbata site na itinye myirịta na otu oyiri ma ọ bụ ihe metụtara ya na-akpaghị aka.
Ọbá akwụkwọ Python nke ahịrịokwu-transformers na-enye ụdị SBERT a zụrụ azụ maka ịkọwapụta okwu ngwuputa na iwepụta ederede yiri ya.
Ihe ize ndụ & okporo ụzọ nche
Eziokwu ndị e chepụtara echepụta nwere ike jiri nwayọ tinye akụkọ, nkwado nkwado, ma ọ bụ nsonaazụ nyocha.
Mmetụta ngwa ngwa nwere ike ịmepụta nsonaazụ na-ekwekọghị ekwekọ n'ofe arịrịọ ndị yiri ya.
Enwere ike ikpughe data ederede nwere mmetụta ma ọ bụrụ na njikwa ohere adịghị ike.
Map mmejuputa
Kọwaa usoro mmepụta, ụda, na ụkpụrụ ịdịmma tupu ibugharị.
Weghachite nzaghachi site na isi mmalite ntụkwasị obi mgbe ọ bụla izi ezi dị mkpa.
Debe ebe nleba anya mmadụ maka mpụta dị elu.
Sochie ụkpụrụ ọdịda ma na-azụghachi mkpali ma ọ bụ usoro ọrụ mgbe niile.
Nọgide na-eme nchọpụta
Free newsletter
Get the daily AI briefing
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Take the Sentence-BERT Embeddings quiz
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
Ntuziaka na-esote
Ntinye akwụkwọ Hypothetical HyDE
Ajụjụ a na-ajụkarị
What is Sentence-BERT Embeddings?
Sentence-BERT (SBERT) na-emegharị BERT iji mepụta otu vector dị ogologo maka otu ahịrịokwu dum, yabụ enwere ike iji ihe ọ pụtara tụnyere myiri cosine ngwa ngwa. O mere nyocha na nchikota ọtụtụ nde ahịrịokwu bara uru, na-atụgharị ọrụ were awa BERT n'ime milliseconds.
Kedu isi nsogbu na BERT larịị ka Sentence-BERT na-edozi?
BERT dị larịị ga-ahazirịrị ụzọ abụọ nke ahịrịokwu ọnụ, yabụ ntụnyere n'ụzọ buru ibu enweghị ike ịgbakọ; SBERT na-etinye akara ahịrịokwu ọ bụla otu ugboro n'ime vector enwere ike iji ya rụọ ọrụ.
Kedu nhazi netwọkụ ka Sentence-BERT na-eji?
SBERT na-eji usoro siamese (ejima) ebe abụọ BERT encoders na-ekekọrịta nha na nke ọ bụla na-etinye otu ahịrịokwu n'onwe ya.
Kedu atụmatụ ịchịkọta ọnụ ka a na-atụkarị aro maka ntinye SBERT?
Pụtara ịgbakọ n'elu vector token ikpeazụ na-akacha nke ọma site na iji akara [CLS] naanị maka ntinye ahịrịokwu.
Ozugbo agbakwunyere ahịrịokwu, kedu ka esi elele myirịta ha?
Isi ihe SBERT niile bụ na myirịta na-ebelata ka ọ bụrụ ntụnyere ngwaahịa cosine/ ntụpọ dị ọnụ ala n'etiti vectors agbagoro agbagoro.
Kedu ụdị dataset a na-ejikarị emezi SBERT?
A na-eji data NLI eme ihe n'ọtụtụ ebe: a na-ejikọta ụzọ abụọ entailment ọnụ na emegiderịta onwe ha, na-akpụzi oghere ntinye bara uru.