GLIGEN
GLIGEN (Grounded-Language-to-Image Generation) daf lay may nga xool fi mbir yi di feeñ ci nataal biñ defar, ci joxe model bi ay boîte ak etiket yu nekk ci wetu bind bi.
Résumé
It turns vague text-to-image into precise, layout-controllable synthesis.
Plongeur bu xóot
Xeetu bind-à-nataal yiñ miin dañuy xeex ngir saytu barab bi: laajal 'muus ci cammoy xaj' te barina luñu lay juum ci plasement bi. GLIGEN, biñu dugal ci 2023, dafay saafara jafe-jafe yii ci yokk ay dugal yu am suuf, lu ci melni boyet yuy tënk yuñ boole ak mbind wala entite nataal, poñ yu am solo, wala nataali royuwaay. Lu gëna am solo mooy, dafay gelé diisaayu modelu diffusion biñ njëkka tàggat, ba noppi dugal ay couche self-attention yu bees yuñ mëna tàggat, yuy xëcci jetons yi ci suuf. Loolu dafay tekki ni dafay tabax ci kaw xeetu Stable Diffusion te du yàq xam-xam bimu jàng, ba noppi gating bi dafay tàmbali ci wetu zero suko defee xeetu base bi dafay wéy di am ci tàggat bi. Resultaa bi mooy generation grounded bu àdduna bi ubbeeku: mën nga def mbir yuñ leeral ci barab yuñ la wax, te dafay yamale konsept yi ak layout yi nga gisul ci diiru tàggat grounding.
Gis-gis xarala
GLIGEN dafay màndargaal bépp entite bu nekk ci suuf ni token buy boole bindam wala nataalam ak ay leeral ci barab bi, lu melni ñeenti coordonnées yu benn boîte de bounding buñ kodee ci man-mani Fourier. Yii jetons grounding duggu ci diffusion U-Net frozen jaaraleko ci ay couche self-attention yuñ dugal ci digganté bloc self-attention ak cross-attention yi fi nekk. Benn buntu buñu mëna jàng, buñu tàmbalee ci zero, mooy saytu ni grounding bi di amee ci generation, kon yokk njiit dafay yàqu bu baax, tàggat yaram dafay des ci anam wu dëgër.
njeextalu pexe
Gaawaay ak yaatuwaay
Visual IA mën na otomatise saytu, gis ak etiketu liggéey ci eskaal.
Tabax tànneef
Ekipu kreatif yi mën nañu defar konsept yu gëna gaaw te duñu def lu bari ci loxo.
Ekip ak def liggéey
Liggéeyukaay yi mën nañu jëfandikoo siñaal nataal wala wideo yu jafewoon lool ci liggéey.
Ëlëgu GLIGEN
Generaasioŋ bu am suuf ak layout buñ mëna saytu mingi nekk standard ci jumtukaayi liggéey bi. Xaarandil ni GLIGEN-style spatial conditioning ngir boole ak yeneen pexe yu melni ControlNet ak laaj gox, ak yokk ci wideo ak 3D fu plasement objet ci diir ak espace gëna am solo. Ginaaw model yi dañuy jëfandikoo interface yu topp tegtal yi, seytu jëmmal drag-and-drop ak graphique yu leer ci làkk wi dina tax composition bu jaar yoon mën ci am te doo soxla ay pexe yu gaaw.
Doxal ci àdduna dëgg
Teg benn logo wala produit ci gox bu gëna jub ci piblisite biñ defar, jëfandikoo benn boyet buy tënk
Defar seen yu jaffe di leeral fi personage wala mbir bu nekk wara toog balaa ngay render
Defar ay done di tàggat ngir gis mbir ak barabi boyet dëgg-ci suuf
Defar ab mbir buñu leeral ci gox bu jëfandikukat bi desine ci ab nataal bu fi nekk
Risk yi ak balustrade yi
Yelleefi nataal ak nangu mën na nekk risku yoon sudee fi ñu bawoo leerul.
Performance model bi mën na wuute ci leeraay bi, demographie bi ak environmaa bi.
Njuumteg positive yi mën nañu dem te kenn duko seetlu fileek xool wuñu buntu wóolu sa bopp.
Roadmap ngir samp gi
Mandargal kritërium nangug njub, woowaat ak njëgu njuumte.
Saytu ak done yu méngoo ak anam yi ñuy liggéeyee dëgg.
Yokk jàngat nit ngir xam fu wóorul dara wala am njeexital yu rëy.
Toppal model drift bi nga baaxal ko ginaaw bi kamera bi wala done yi soppeekoo.
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 GLIGEN Grounded Generation 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
Bind-ci-3D
Laaj yi ñuy faral di laaj
What is GLIGEN Grounded Generation?
GLIGEN (Grounded-Language-to-Image Generation) daf lay may nga xool fi mbir yi di feeñ ci nataal biñ defar, ci joxe model bi ay boîte ak etiket yu nekk ci wetu bind bi. Dafay soppi bind-ci-nataal bu leerul nekk synthese bu leer, buñu mëna saytu jëmmal.
Ban jafe-jafe la GLIGEN njëkka saafara?
GLIGEN dafay yokk ay dugal yu melni ay boîte de bounding suko defee nga mëna saytu fu mbir yi di nekk, te bind bu leer du def lu baax.
naka la ko GLIGEN mënee sàmm xam-xam bi ci xeetu diffusion biñ njëkka tàggat?
GLIGEN dafay gelé modelu base bi ba noppi dugal ay couche self-attention yu bees, suko defee xam-xam bi ñu njëkka jàng nekk ci barab bi.
Lan mooy dugal grounding bi GLIGEN di nangu?
GLIGEN dafay jàppale grounding jaaraleko ci ay boyet yu am mbind/nataal, poñ yu am solo ak nataali royuwaay.
Lan moo waral buntu bi ci GLIGEN's layers attention bu bees bi ñu tàmbali ko ci zero?
Gate bu amul benn initialisation dafay tekki ni grounding bi amul benn njeexital ci ndoorte li, di denc model bi njëkk ba noppi di tëye tàggat yaram bi rampe de contrôle bi di yokk.
Luy 'àdduna bu ubbeeku' ci GLIGEN?
GLIGEN dafay yamale lu ëpp limu tàggat ci suuf, def mbir yu bees yuñ leeral ci barab yuñ leeral.