Amabwiriza ya AI

Tekereza Umwandiko-Kuri-Ishusho

Imagen ni __AIU_PROTECTED_11 __ 'sisitemu-y-sisitemu sisitemu ihindura ibisobanuro byanditse mubishusho bifotora.

Umva iki gitaboAI-byavuzwe · ~2 min

Kuriyi page2 min soma
  1. Incamake
  2. Kwibira cyane
  3. Gushyira mu bikorwa Isi
  4. Kazoza ka Imagen Umwandiko-Kuri-Ishusho
  5. Kubishyira mubikorwa
  6. Komeza Ubushakashatsi
  7. Ibibazo bikunze kubazwa

Incamake

Its headline finding was that a large frozen language model, not a bigger image network, was the biggest driver of quality.

Kwibira cyane

Byatangajwe na Google Ubushakashatsi mu 2022, Imagen yerekanye ko gusobanukirwa byimazeyo ibintu byihuse kimwe no kubishushanya neza. Mu mwanya wa kodegisi yuburyo bwa CLIP, Imagen ikoresha kodegisi nini yabigenewe (T5-XXL) ikomeza gukonjeshwa, hanyuma igaburira izo ndimi zikungahaye muburyo bwo gukwirakwiza. Itanga ishusho ntoya ya 64x64 kandi ikoresha ibyiciro bibiri bya super-resolution ikwirakwizwa kugirango izamuke kugeza 1024x1024. Iri tsinda kandi ryashyizeho 'dinamike ntarengwa' kugira ngo amabara ahamye mu buyobozi buhanitse, maze yubaka DrawBench, igipimo cyerekana amayeri yerekana ibizamini byo kubara, umubano w’ahantu, hamwe no guhuza bidasanzwe. Nyuma yaho, Imagen 2 na Imagen 3, ibisobanuro birambuye, gutanga inyandiko, no kwizerwa byihuse, none imbaraga __AIU_PROTECTED_11 __ 'ibikoresho byamashusho.

Gushyira mu bikorwa Isi

Gukora amashusho yamamaza yamashusho avuye muri make yanditse nta gufotora

Gukora ibishushanyo mbonera byo kuvuga inkuru cyangwa ibitabo by'abana uhereye ku nteruro isobanura

Gukora ibicuruzwa byashinyaguriwe nuburyo butandukanye kuri e-ubucuruzi kurutonde

Kwiyumvisha ibitekerezo bya siyansi cyangwa uburezi, nkibishushanyo byumuhanzi byasobanuwe mururimi rworoshye

Kazoza ka Imagen Umwandiko-Kuri-Ishusho

Imirongo ya Imagen igenda yerekeza kumyandiko nziza itanga imbere mumashusho, kwihuta gukurikira kubintu bigoye, hamwe no gutoranya byihuse. Tegereza guhuza byimbitse nurugero rwururimi bityo sisitemu 'impamvu' kubyerekeye icyifuzo mbere yo gushushanya, wongeyeho ibimenyetso bikomeye byamazi nka SynthID kugirango bigaragare. Nkuko bihuza ibicuruzwa bya __AIU_PROTECTED_11 __ nibidukikije bya Gemini, intumbero ihinduka kubisekuru byizewe, umutekano, bigenzurwa aho kuba udushya twinshi.

Kubishyira mubikorwa

Impamvu ari ngombwa

  • Umuvuduko n'igipimo. AI igaragara irashobora gukora igenzura, gutahura, no gutondekanya imirimo kurwego.
  • Kubaka amahitamo. Amakipe arema arashobora prototype ibitekerezo byihuse hamwe nintoki nkeya.
  • Itsinda hamwe nakazi. Ibikorwa birashobora gukoresha amashusho nibimenyetso bya videwo byari bigoye gutunganya.

Witondere

  • Uburenganzira bwishusho hamwe no kwemererwa birashobora guhinduka ibyago byemewe n'amategeko niba ibimenyetso bidasobanutse.
  • Imikorere yicyitegererezo irashobora gutandukana kumurika, demografiya, nibidukikije.
  • Ibyiza byibinyoma birashobora kutamenyekana keretse niba ibyiringiro byateganijwe bikurikiranwa.

Intambwe ikurikira

  1. Sobanura ibipimo byo kwemererwa kugiciro, kwibutsa, nibiciro byamakosa.
  2. Gerageza hamwe namakuru ajyanye nuburyo nyabwo bwo gukora.
  3. Ongeraho isubiramo ryabantu kubwizere buke cyangwa guhanura cyane.
  4. Kurikirana icyitegererezo cya drift hanyuma uhindurwe nyuma ya kamera cyangwa dataset ihinduka.

Komeza Ubushakashatsi

Ibaruwa y'amakuru y'ubuntu

Ibuka amakuru ya buri munsi ya AI

Inkuru eshatu zizewe za AI buri gitondo cy'iminsi y'akazi, zanditse mu rurimi rworoshye. Ubuntu iteka ryose, nta kwamamaza.

Imeri imwe buri munsi y'akazi. Kwiyimura mu buryo bumwe. Ntitwigera tugurisha cyangwa ngo dusangire aderesi yawe.

Isuzume

Take the Imagen Text-to-Image quiz

Ibisubizo ako kanya kuri buri kibazo, n'icyemezo gishobora gusangirwa gifite ID yagenzurwa umaze gutsinda isomo.

Tangira ikibazo

Fasha uburezi bwa AI bw'ubuntu. AI Understanding ni umuryango utegamiye kuri leta wa 501(c)(3) — nta kwamamaza, nta kwishyura, iteka ryose. Tanga impano

Ibibazo bikunze kubazwa

What is Imagen Text-to-Image?

Imagen ni __AIU_PROTECTED_11 __ 'sisitemu-y-sisitemu sisitemu ihindura ibisobanuro byanditse mubishusho bifotora. Umutwe wacyo wasangaga ni uko urugero runini rwururimi rwakonjeshejwe, atari urusobe runini rwamashusho, rwari umushoferi munini wubwiza.

In the original Imagen study's scaling comparison, which change produced larger image-quality and text-alignment gains?

The reported comparisons favored scaling the text encoder; this is a study result, not a universal rule for every architecture.

Niyihe kodegisi ya Imagen yishingikiriza?

Imagen ikoresha kodegisi nini ya T5-XXL gusa, ikomeza gukonjeshwa mugihe cy'amahugurwa.

Nigute Imagen igera kumurongo wo hejuru?

Itanga ishusho ya 64x64 hanyuma ikazamuka ikoresheje moderi ya super-resolution ikwirakwizwa kugeza 1024x1024.

What problem did original Imagen's dynamic thresholding address under strong guidance?

The method clips and rescales predicted pixels to limit saturation under large guidance weights.

DrawBench ni iki?

DrawBench ni igipimo Google yatangijwe na Imagen kugirango ibare ibarwa, umubano utandukanye, hamwe nibidasanzwe.