Mbind yuñ dugal
Text embedding yi dañuy soppi baat yi, frase yi, wala dokimaa yi ñu nekk limu nimero (vecteur) yuy jàpp luñuy tekki, suko defee mbind yi am lu ñuy nuru luñuy mujjee jegewaale ci espace bi.
Résumé
They are the foundation for semantic search, recommendations, clustering, and the retrieval behind many AI assistants.
Plongeur bu xóot
Ordinatër yi mënu ñu xalaat ci mbind mu ñor, moo tax dañuy soppi làkk wi ci vecteur yu am guddaay bu takku, lu ci bari ay téemeeri wala lu ëpp junni dimension. Li gëna am solo mooy distance bi ci espace vecteur bi dafay wane lu muy tekki: suuf su "kontaan" ak "kontaan" dañu jege seen biir, fekk "kontaan" ak "asphalte" dañu sori seen biir. Baatu-baatu njëkk yu melni Word2Vec ak GloVe dañu jox baat bu nekk benn vecteur bu takku, muy lu siiw lool ci analogie yu melni buur dindi góor ak jigéen ñu wàcci ci wetu reine. Seenug gàttal mooy baat bu melni "bànk" dafa am benn vecteur muy tekki tefesu dex wala bànk xaalis. Embeddings contextuel yu bees yu bawoo ci model transformateur yi defar nañu lii ci jox baat bi vecteur bu wuute ci frase bi. Modèlu embedding frase ak dokimaa dem nañu gëna sori, dajale passage yépp ci benn vecteur bu bari luñu ci mëna seetee wala boole.
Gis-gis xarala
Embedding vecteur bu dëgër la, te dañuy faral di natt nuru gi ci nuru cosine, muy méngale angle bi am ci diggante ñaari vecteur, guddaay bi du ci dara. Word2Vec jàng vecteur yi ci xam baat yi jege, moo tax baat yiy lëkkaloo dañuy booloo. Embeddings frase yu bees yi ñu ngi bawoo ci encodeur transformateur, ñuy faral di boole ay token ci benn vecteur ba noppi ñu tàggat leen ci ay mébet yu wuute yuy boole ay paraphrase ak di tàqale ay mbind yu amul benn lëkkaloo. Vecteur yi ciy génn ñooy li ñuy denc ci base de done vecteur yi, ñu méngale ko ci seetlu semantik ak ci génne giñ yokk.
njeextalu pexe
Gaawaay ak yaatuwaay
Liggéeyukaay yi ci làkk yi mën nañu gëna gaaw te duñu yàq deggoo gi.
Dugg ak yegg
Dafay yaatal jëfandikoo gi ci làkk yi ak ci anam yi ñuy jokkoo.
dogal yu gëna leer
Ekip yi mën nañu gëna yàgg ci àtte ci jamono ji otomatisation di liggéey ci baamtu.
Ëlëgu dugal mbind
Embeddings mingi nekk interface universel ngir IA: benn espace vecteur dafay gëna yaatu ci bind, nataal, audio, ak kod, loolu mooy tax ñu mëna seetlu ci anam yu bari. Xaarandil model yiy samp këyitu dokimaa yu gëna gudd ci anam wu jaar yoon, samp yu làkk yu bari yuy méngale luñuy tekki ci làkk yi, ak model yu gëna ndaw, yu gëna gaaw yuy daw ci aparey bi ngir nëbbëtu. Jëf yuñ miin lu melni normalisasioŋ ak embedding yuñ mëna dagg ci anam Matryoshka, loolu dina tax nga gaawa gàttal benn vecteur ngir denc dencukaay bi te baña ñàkk kalite bu bari, ñu ngi tasaaroo. Lu jamono ji ñuy yokk ci seet gi di magg, kalite bi ñuy dugal ci saasi dafay wane ni assistant IA yi di jaar yoon, di wéy di nekk barab bu am doole te am njeexital.
Doxal ci àdduna dëgg
Seetug semantik bu am doole suko defee laaj bi méngoo ak këyit yi ci luñu tekki, du ci baatu-caabi yi
Dajale ay junni jàngat ci ay theme ci dajale jàngat yu jege seen biir
Xalaatal xët wala produit yu noonu mel ci seet mbir yu vecteur yiñ boole gëna jege benn bi jëfandikukat bi bëgg
Gis ñaari yoon wala ñu jege ñaari yoon ñuy jàppale ci natt ni seeni embeddings jegee
Risk yi ak balustrade yi
Lépp lu jaarul yoon mën na dugg ci rapoor yi, jàppale ci liggéey bi, wala ci njariñu gëstu bi.
Sensibilite bu gaaw mën na jur njariñ yu wuute ci laajte yu noonu mel.
Done yu am solo mën nañu feeñ sudee seytu jëfandikoo gi néew doole.
Roadmap ngir samp gi
Mandargal formaa génne gi, melokaan bi, ak standard kalite yi laata ngay dugal ko.
Tontu yu am solo ak balluwaay yu wóor saa yu dëggu bi di am solo.
Fexeel am barabu xool nit ñi ngir am njariñ yu am solo.
Toppal anami gacce yi ak di faral di tàggataat ay laaj wala def-liggéey.
Weyal di banneexu
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 Text 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
Gis bi ci topp
Rotary Position
Laaj yi ñuy faral di laaj
What is Text Embeddings?
Text embedding yi dañuy soppi baat yi, frase yi, wala dokimaa yi ñu nekk limu nimero (vecteur) yuy jàpp luñuy tekki, suko defee mbind yi am lu ñuy nuru luñuy mujjee jegewaale ci espace bi. Mooy fundamaasu seetlu semantik, xalaat, dajale, ak seet ci ginaaw assistant IA yu bari.
Luy dugal mbind?
Ab embedding dafay boole mbind ci vecteur numérique bu am guddaay bu takku suko defee lu nuróo luñuy tekki génne ay vecteur yu jegewaale ci espace bi.
Ci barab bu baax, lan moo wara nekk dëgg ci kàddu yi 'kontaan' ak 'bànneex'?
Ndax baat yi dañuy nuru luñuy tekki, seeni vecteur embedding dañu wara toog jegewaale, waaye baat yi amul benn lëkkaloo lu melni 'kontaan' ak 'asphalt' dañu wara sori seen biir.
Lan moo nekkoon ay jafe-jafe yu mag ci jëfandikoo kàddu yu njëkk yu melni Word2Vec ak GloVe?
Baatu static embeddings yi dañuy jox benn vecteur ci baat bu nekk, kon baat bu bari maana bu melni 'bank' dafay am benn representation muy tekki tefesu dex wala institution financière.
Ak naka lay deme bay gëna xéewale ci samp gi ci baat bu fippu?
Embeddings contextuel yu sukkandiko ci transformateur yi dañuy defar vecteur bu aju ci contexte bi, kon 'bank' ci frase finansieer wuute na ak 'bank' ci wetu dex.
Ban nattukaay lañu gëna jëfandikoo ngir méngale ñaari mbind yuñ tënk ngir nuru?
Nuru cosinus mooy natt angle bi am ci digganté vecteur yi, di jàpp nuru ci yoon (li muy tekki) magnitude bumu mëna doon.