Swin Transformer
Swin Transformer shine mai canza hangen nesa wanda ke aiwatar da hotuna a cikin sauye-sauye, tagogi masu matsayi, yana ba da hankali sosai don auna girman hotuna masu tsayi.
Dubawa
It works as a general-purpose backbone for classification, detection, and segmentation.
Zurfafa nutsewa
Standard Vision Transformers suna lissafta hankali a duk facin hoto, wanda farashin ke girma da girman hoto, cikas ga ayyuka masu yawa kamar ganowa. Microsoft Ne ya gabatar da bincike a cikin 2021, Swin (Shifted WINdows) a maimakon haka ya raba hoton zuwa kananan tagogi marasa rufi kuma yana ƙididdige kulawar kai kawai a cikin kowace taga, yana sa farashi yayi girma daidai da girman hoto. Don barin bayanai su ketare iyakokin taga, madaidaicin yadudduka suna canza grid ɗin taga, don haka faci waɗanda aka raba yanzu suna raba taga. Swin kuma yana gina matsayi: yana farawa da ƙananan faci kuma yana haɗa su a hankali, yana samar da taswirorin fasali masu yawa kamar CNN, wanda ke shiga cikin tsarin ganowa da rarrabuwa.
Fahimtar Fasaha
Ƙwarewar Swin ta fito ne daga mai da hankali kan kai-da-kai na tushen taga (W-MSA): hankali yana iyakance ga ƙayyadaddun windows (misali faci 7x7), don haka ma'auni mai rikitarwa a cikin layi maimakon quadratically tare da adadin faci. Toshe na gaba yana amfani da kulawar taga canzawa (SW-MSA), yana kawar da ɓangaren taga da rabin taga don haka haɗin giciye-taga. Faci-haɗin yadudduka suna haɗa facin maƙwabta a tsakanin matakai, rage ƙudirin sarari da tashoshi biyu don gina dala mai fasali.
Dabarun Tasiri
Gudu da sikelin
Kayayyakin AI na iya sarrafa aiki da bincike, ganowa, da ayyuka masu alama a sikelin.
Gina zaɓuɓɓuka
Ƙungiyoyin ƙirƙira za su iya samar da ra'ayoyi cikin sauri tare da ƙarancin bita da hannu.
Ƙungiya da aikin aiki
Ayyuka na iya amfani da siginar hoto da bidiyo waɗanda a baya suke da wahalar aiwatarwa.
Makomar Swin Transformer
Swin ya nuna cewa masu canji, masu sane da yanki na iya yin hamayya ko doke CNNs a matsayin kashin bayan hangen nesa na duniya, kuma Swin V2 ya tura wannan zuwa ƙirar siga biliyan biliyan da manyan kudurori. Yi tsammanin ci gaba da haɗawa da son zuciya mai jujjuyawa tare da kulawa, bambance-bambancen kulawa mai inganci, da kasusuwan bayan salon Swin suna ciyar da tsarin multimodal da na bidiyo. Kamar yadda ƙirar tushe don girmar hangen nesa, ƙirar ƙira waɗanda ke samar da fasalulluka masu yawa sun kasance masu mahimmanci musamman ga ayyukan tsinkaya.
Aiwatar da Gaskiyar Duniya
Rarraba ImageNet mai inganci a matsayin ingantaccen kashin baya
Gano abu da misalin kashin baya a cikin tsarin kamar Mask R-CNN da Cascade R-CNN
Bangaren ma'ana na al'amuran titi da hotunan tauraron dan adam
Binciken hoto na likita inda babban ƙuduri da cikakkun bayanai masu yawa ke da mahimmanci
Hatsari & Tsare-tsare
Haƙƙoƙin hoto da yarda na iya zama haxarin doka idan ba a fayyace ba.
Ayyukan samfuri na iya bambanta a ko'ina cikin haske, ƙididdiga, da mahalli.
Ƙarya tabbataccen ƙila ba za a iya lura da shi ba sai dai idan an kula da ƙofofin amincewa.
Taswirar Hanya
Ƙayyade ma'auni na karɓa don daidaito, tunowa, da farashi na kuskure.
Gwada tare da bayanan da suka dace da ainihin yanayin samarwa.
Ƙara bita na ɗan adam don ƙarancin amincewa ko tsinkaya mai tasiri.
Bi diddigin ƙirar ƙira kuma sake ingantawa bayan canje-canjen kamara ko saitin bayanai.
Ci gaba da Bincike
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 Swin Transformer 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
Jagora na gaba
Yadawa Transformers
Tambayoyin da ake yawan yi
What is Swin Transformer?
Swin Transformer shine mai canza hangen nesa wanda ke aiwatar da hotuna a cikin sauye-sauye, tagogi masu matsayi, yana ba da hankali sosai don auna girman hotuna masu tsayi. Yana aiki azaman ƙashin baya na gaba ɗaya don rarrabuwa, ganowa, da rarrabuwa.
Menene 'Swin' a cikin Swin Transformer ke tsayawa ga?
Swin yana nufin Windows Shifted, tsarin da ke ba da damar windows hankalin gida ya musanya bayanai a cikin yadudduka.
Me yasa lissafin hankalin kai a cikin tagogin gida yana da fa'ida?
Ƙaddamar da hankali ga ƙayyadaddun windows yana sa farashin lissafi yayi girma daidai da girman hoto, ba kamar haɓakar hankalin duniya ba.
Ta yaya Swin ke barin bayani ya motsa tsakanin tagogi daban?
Matsakaicin yadudduka suna canza grid ɗin taga da rabin taga, don haka faci a baya a cikin tagogi daban-daban na iya halartar juna.
Menene facin-haɗin yadudduka ke yi tsakanin matakan Swin?
Haɗin facin yana haɗa faci na maƙwabta don rage ƙudurin sarari da zurfin tashoshi biyu, yana gina manyan ma'auni.
Me yasa matakan Swin ke da amfani musamman?
Taswirorin sikeli da yawa suna kwaikwayi kasusuwan bayan CNN, don haka Swin yana haɗawa cikin sauƙi tare da tsarin tsinkaya mai yawa kamar Mask R-CNN.