InoteveraGaidhi rinotevera
Custom Diffusion Multi-Concept Tuning
AI Yekutarisa
Gwara reVisual AI
SDXL uye Cascaded Diffusion
SDXL ndeye Stability AI's high-resolution text-to-image modhi iyo inobatanidza ine simba base jenareta ine inonatsa, ukuwo cascaded diffusion inosunga mamodheru akawanda ekuvaka mifananidzo kubva pasi kusvika pakakwirira resolution.
Teerera kugwaro iriAI-yakarondedzerwa · ~2 min
Pa peji ino2 min verenga
Pfupiso
Pamwe chete vanotsanangura nzira yemazuva ano yakavhurika-sosi mifananidzo jenareta yakarova photorealistic mhando.
Kudzika Kwakadzika
SDXL (Stable Diffusion XL) ingangoita 3.5-bhiriyoni-parameter diffusion modhi iyo yekuzvarwa inogadzira 1024x1024 mifananidzo, kusvetuka kukuru pamusoro peiyo 512x512 yekutanga Stable Diffusion. Inoshandisa maencoder maviri emavara (OpenCLIP ViT-bigG uye CLIP ViT-L) kuti anzwisise nekukasira, pamwe nekukura uye mamiriro echirimwa kuitira kuti modhi izive gadziriso nekumisikidza kwakanangwa. SDXL ngarava sepaipi-matanho maviri: modhi yepasi inogadzira iyo yakadzikama mufananidzo, ipapo inosarudzika yekunatsa modhi inowedzera yakanyatsojeka mumatanho ekupedzisira ekuita denoising. Cascaded diffusion ndiyo pfungwa yakafara kuseri kweizvi: pane kuti modhi imwe iite zvese, unosunga modhi diki inogadzira mufananidzo wakaderera une super-resolution diffusion modhi inokwirisa, imwe neimwe yakadzidziswa padanho rayo. Google's Imagen yakaita mukurumbira.
Kuitwa muNyika yeChokwadi
Kugadzira 1024x1024 kushambadzira uye pfungwa art zvakananga kubva kune zvinyorwa zvinokurudzira pasina yakaparadzana upscaler.
Kushandisa iyo SDXL base-plus-refiner pombi yekuwedzera crisp ruzivo kuzviso uye maumbirwo mune zvigadzirwa mockups.
Inomhanya SDXL Turbo yepedyo-pakarepo mufananidzo wekutarisa mune inopindirana dhizaini maturusi
Kuvaka tsika yepamusoro-resolution cascade kushandura yakaderera-res sketches kuita yakakwirira-resolution mifananidzo.
Ramangwana reSDXL uye Cascaded Diffusion
Maitiro ari kune mashoma, nhanho dzinokurumidza uye akabatana architecture. Distillation nzira seSDXL Turbo uye Latent Consistency Models yatocheka chizvarwa kusvika kune imwe kusvika mana nhanho. Diffusion transformers (sezviri muStable Diffusion 3 uye FLUX) iri kunyanya kutsiva U-Net musana, uye kuguma-kusvika-kumagumo chizvarwa chepamusoro-chigadziriso chiri kuderedza kuvimba nemakasi akajeka. Tarisira kubatanidzwa kwakasimba kwekunatsiridza, kupa zvinyorwa zviri nani, uye chaiyo-nguva pa-mudziyo mufananidzo synthesis sezvo kushanda zvakanaka kunoramba kuchiwedzera.
Kuzvishandisa
Nei zvichikosha
- Kumhanya uye chiyero. Visual AI inogona kuita otomatiki yekuongorora, yekuona, uye yekumaka mabasa pachiyero.
- Vaka sarudzo. Zvikwata zvekugadzira zvinogona prototype pfungwa nekukurumidza nekudzokororwa kwemaoko mashoma.
- Team uye workflow. Mashandisirwo anogona kushandisa masaini emifananidzo nemavhidhiyo ayo aimbove akaoma kugadzirisa.
Chenjera
- Kodzero dzemifananidzo uye kubvumirwa kunogona kuve njodzi dzepamutemo kana provenance isina kujeka.
- Kuita kwemuenzaniso kunogona kusiyanisa kupenya, huwandu hwevanhu, uye nharaunda.
- Manyepo enhema anogona kusacherechedzwa kunze kwekunge zvikumbaridzo zvekuvimba zvikatariswa.
Matanho anotevera
- Tsanangura maitiro ekugamuchirwa echokwadi, kurangarira, uye mutengo wekukanganisa.
- Edzai nedata rinoenderana nemamiriro chaiwo ekugadzira.
- Wedzera ongororo yemunhu kune yakaderera-kusavimbika kana yakakwirira-inokanganisa kufanotaura.
- Tevera modhi kudonha uye simbisa mushure mekuchinja kwekamera kana dataset.
Ramba Uchiongorora
Tsamba yemahara
Tora iyo yemazuva ese AI pfupiso
Nhatu dzakasimbiswa dzeAI nyaya mangwanani ega ega pakati pevhiki, dzakanyorwa neChirungu chakajeka. Yemahara zvachose, hapana ads.
Imwe email zuva rega rega revhiki. Bvisa kunyoresa kamwe chete. Hatimbotengese kana kugovera kero yako.
Zviedze iwe pachako
Take the SDXL and Cascaded Diffusion quiz
Mhinduro yepakarepo pamhinduro yega yega, uye chitupa chinogovaniswa chine ID inobatika kana wapedza kosi.
Tsigira dzidzo yemahara yeAI. AI Understanding ndeye 501(c)(3) isingabatsiri - hapana ads, hapana paywall, zvachose. Ita chipo
Mibvunzo inowanzo bvunzwa
Chii chinonzi SDXL uye Cascaded Diffusion?
SDXL ndeye Stability AI's high-resolution text-to-image modhi iyo inobatanidza ine simba base jenareta ine inonatsa, ukuwo cascaded diffusion inosunga mamodheru akawanda ekuvaka mifananidzo kubva pasi kusvika pakakwirira resolution. Pamwe chete vanotsanangura maitiro emazuva ano akavhurika-sosi majenareta anorova photorealistic mhando.
Ndeipi gadziriso yemuno iyo SDXL inogadzira mifananidzo pairi, ichienzaniswa neyekutanga Yakagadzikana Diffusion?
SDXL yekuzvarwa inogadzira 1024x1024 mifananidzo, nzvimbo-yakakura-kana-yakakura pane yekutanga Stable Diffusion's 512x512.
Ndeipi basa re SDXL's refiner modhi?
Munatsi ndiyo imwe nyanzvi modhi yakashandiswa kumagumo kwepombi kurodza zvakadzama mushure mekunge modhi yepasi yagadzira mufananidzo uripo.
Mangani emavara encoder anoshandiswa neSDXL?
SDXL inoshandisa maencoder maviri emavara, OpenCLIP ViT-bigG uye CLIP ViT-L, yakasanganiswa kuti iwedzere kunzwisisa nekukurumidza.
Ndeipi pfungwa huru yekupararira kwecascade?
Cascaded diffusion cheni diki base jenareta ine imwe kana anopfuura epamusoro-resolution diffusion modhi, imwe neimwe yakarongedzwa pane yapfuura yakaderera-resolution yakabuda.
Sei SDXL ichidenha munzvimbo yakanyarara kwete yakananga pamapikisheni?
A VAE inodzvanya mapikicha munzvimbo diki yakavanzika, saka iyo diffusion process yakachipa zvakanyanya pane kushanda pane yakazara-resolution mbishi pixels.
Ramba uchidzidza
Mazano anoenderana
Mamwe madhairekitori akasarudzirwa nyaya iyi