InstructPix2Pix Umarnin Gyara
InstructPix2Pix yana ba ku damar shirya hoto ta hanyar buga umarni a sarari kamar 'sanya shi hunturu' ko 'juya cat ya zama kare', babu abin rufe fuska ko kayan aikin zaɓi da ake buƙata.
Dubawa
It taught a diffusion model to follow editing instructions directly.
Zurfafa nutsewa
InstructPix2Pix (Brooks et al., 2023) samfurin watsawa ne mai kyau wanda aka gyara don ɗaukar hoton shigarwa da umarnin rubutu da fitar da hoton da aka gyara a cikin wucewar gaba ɗaya. Dabararsa mai wayo ita ce bayanan horo: marubutan sun yi amfani da GPT-3 don samar da nau'i-nau'i-nau'i-nau'i-nau'i-nau'i na gaba-da-bayan, sannan suka yi amfani da Saurin-zuwa-Bayyana tare da Stable Diffusion don haɗa matching kafin/bayan nau'ikan hoto. Wannan ya ba su babban adadin bayanan (hoton asali, umarni, hoton da aka gyara) sau uku don horarwa, duk ba tare da alamar hannu ba. Saboda umarnin yana bayyana canji maimakon cikakken yanayin, ƙirar tana adana sassan hoton da ba a ambata ba. Yana amfani da ma'aunin jagora guda biyu, ɗaya don yadda yake bin umarnin kuma ɗayan don yadda aminci yake manne da ainihin hoton, yana barin masu amfani suyi musayar ƙarfin gyarawa akan aminci.
Fahimtar Fasaha
Sharuɗɗan ƙirar akan duka hoton tushe da koyarwa, yin amfani da jagorar kyauta tare da gatari biyu. Ɗayan ma'auni yana ɗaukar umarnin rubutu, ɗayan yana ɗaukar hoton shigarwa. Ɗaukaka sikelin hoto yana kiyaye ƙarin ainihin asali, yayin da haɓaka ma'aunin rubutu yana sa gyara ya zama mai ƙarfi. Wannan jagorar dual shine abin da ke ba da damar koyarwa guda ɗaya ta dogara da canza al'amari ɗaya yayin barin sauran hoton a iya ganewa.
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 InstructPix2Pix Umarnin Editan
Daidaita tushen umarni yana zama tsohuwar dubawa don kayan aikin hoto, yanzu ana gasa cikin manyan ƙa'idodi da magada kamar MagicBrush da masu gyara masu juyawa da yawa. Yi tsammanin mafi kyawun adana dalla-dalla, ingantaccen kulawar umarnin sararin samaniya kamar 'matsar da fitilar hagu', da ƙari maras kyau zuwa bidiyo, inda umarni ɗaya ke gyara cikakken shirin. Haɗa waɗannan samfuran tare da wakilan harshe zai iya ba ku damar kwatanta cikakken zaman gyara ta tattaunawa.
Aiwatar da Gaskiyar Duniya
Masu rubutun ra'ayin yanar gizo suna nau'in 'ƙara foliage' don sake fasalin hoton yanayin bazara don matsayi na yanayi.
Mai siyar da kasuwancin e-commerce yana ba da umarni 'canza launin rigar zuwa sojojin ruwa' don samar da bambance-bambancen launi na samfur daga harbi ɗaya.
Wani malami yana gyara hoton tarihi tare da 'mai canza launin wannan' don sanya hoton tarihin baƙar fata da fari ya bayyana ga darasi.
Wani mahaliccin meme ya ba da umarnin 'sanya tabarau akan kare' ba tare da rufe fuskar kare da hannu ba.
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 InstructPix2Pix Instruction Editing 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
DragGAN Interactive Editan
Tambayoyin da ake yawan yi
What is InstructPix2Pix Instruction Editing?
InstructPix2Pix yana ba ku damar shirya hoto ta hanyar buga umarni a sarari kamar 'sanya shi hunturu' ko 'juya cat ya zama kare', babu abin rufe fuska ko kayan aikin zaɓi da ake buƙata. Ya koyar da samfurin yadawa don bin umarnin gyara kai tsaye.
Ta yaya za ku gaya wa InstructPix2Pix abin da za a canza?
Kuna kawai rubuta umarni kamar 'sa shi hunturu'; ba a buƙatar abin rufe fuska ko zaɓin yanki.
Ta yaya aka ƙirƙiri bayanan horon InstructPix2Pix?
Marubutan sun haɗa nau'i-nau'i-nau'i-nau'i-nau'i na GPT-3 tare da haɗakar hoto da gaggawa don gina sau uku ta atomatik.
Menene InstructPix2Pix ke sarrafa ma'aunin jagora guda biyu?
Ma'auni ɗaya yana sarrafa yadda ƙarfi da gyara ya bi umarnin; ɗayan yana mulkin nawa ne aka adana hoton tushen.
Me yasa samfurin yakan bar sassan hoton da ba a ambata ba?
Umarni yana nufin takamaiman canji, don haka ƙirar tana adana yankuna da umarnin bai faɗi ba.
Fassara nawa nawa InstructPix2Pix ke buƙatar samar da gyara?
Samfurin yaɗuwar yanayi guda ɗaya ne wanda ke ɗaukar hoto da koyarwa tare da fitar da gyara.