I-VISual AI GUIDE

I-MaskGIT Parallel Token Decoding

Uhlolojikelele

I-MaskGIT Parallel Token Decoding ingeyokugeleza kokusebenza kombono wekhompyutha okuhumusha noma okukhiqiza imidiya ebonakalayo ukuze ihlaziywe, isebenze, futhi isungule.

I-Deep Dive

I-MaskGIT (I-Masked Generative Image Transformer), evela ku-Google ngo-2022, icabanga kabusha ukuthi amamodeli ezithombe asuselwa kumathokheni aqoka kanjani. Ama-transformer angaphambilini afana ne-VQGAN akhiqize amathokheni ngokuzenzakalelayo, eyodwa ngesikhathi ngokulandelana kwe-raster, ehamba kancane futhi engeyona eyemvelo ezithombeni ze-2D. I-MaskGIT esikhundleni salokho iqeqesha ngomgomo wokumodela ofihliwe njenge-BERT: amasethi angaphansi angahleliwe amathokheni esithombe afihliwe futhi imodeli ifunda ukuwabikezela wonke kanyekanye isebenzisa ukunaka okukabili. Ngesikhathi sokukhiqiza iqala kugridi efihlwe ngokugcwele futhi iqoshwe ngenani elinqunyiwe lokuphindaphinda (ngokuvamile 8 kuya ku-12). Isinyathelo ngasinye ibikezela yonke ithokheni efihliwe, igcina izibikezelo zokwethembeka okuphezulu, futhi imaski kabusha okusele emzuliswaneni olandelayo. Lokhu kukhiqiza izithombe zekhwalithi ephezulu cishe ngokulandelana kwezinyathelo ezimbalwa kunokukhipha amakhodi okuzenzakalelayo.

I-Technical Insight

Ingxenye ebalulekile uhlelo lokufihla ubuso olusekelwe ukuzethemba. Ishejuli ye-cosine inquma ukuthi mangaki amathokheni azovezwa ukuphindaphinda ngakunye, aqale kancane futhi asheshise. Ngenxa yokuthi ukunaka kuqondiswa kabili, yonke ithokheni ibona sonke isithombe esingaphelele, ngakho ukwenza izibikezelo ezizethembayo kuqala kuvumela izinyathelo ezizayo esimweni esiqinile, njengokuxazulula izingxenye ezilula zendida ngaphambi kwalezo ezingacacile.

I-Mastering MaskGIT Parallel Token Decoding

I-MaskGIT ikhiqiza izithombe ngokubikezela amathokheni amaningi ngesikhathi esisodwa futhi igcwalise ezithembeke kakhulu kuqala, ishintshe isizukulwane esihamba kancane sisuka kwesokunxele siye kwesokudla ngezinyathelo ezimbalwa ezisheshayo ezihambisanayo. I-MaskGIT Parallel Token Decoding ingeyokugeleza kokusebenza kombono wekhompyutha okuhumusha noma okukhiqiza imidiya ebonakalayo ukuze ihlaziywe, isebenze, futhi isungule. Ukuze wakhe ukuqonda okujulile, phatha i-MaskGIT Parallel Token Decoding njengemodeli yokusebenza, hhayi isici esisodwa: chaza imiphumela oyifunayo, ucacise ukucabanga, futhi uhlukanise lokho uhlelo olungakwenza ngokwethembeka kulokho okusadinga ukwahlulela kochwepheshe.

Empeleni, amaqembu aqinile asebenzisa ukunemba kwebhalansi ye-MaskGIT Parallel Token Decoding namaqiniso okusebenza njengekhwalithi yedatha, ukuhluka kokukhanya, nokuvumelana kwamalebula. Babhala imibandela yempumelelo ecacile, ukuhlola okuqhathaniswa nedatha engokoqobo nokugeleza komsebenzi, futhi baphindaphinde ngokusekelwe kumaphethini okuhluleka aqashiwe esikhundleni sokuwina kwebhentshimakhi yesikhathi esisodwa. Yilapho ukuqonda kwethiyori kuguquka kube amandla ahlala njalo kuwo wonke umkhiqizo, inqubomgomo, kanye nokusebenza.

I-Visual AI ingakwazi ukuhlola, ukutholwa, nokumaka imisebenzi esikalini. Ngesikhathi esifanayo, amalungelo ezithombe kanye nemvume kungaba ubungozi bomthetho uma ukutholakala kungacacile. Indlela eqine kakhulu iwukuhlanganisa isivinini sokuhlola nesiyalo sokuphatha: qhuba abashayeli bezindiza, bamba ubufakazi, ushicilele amalogi ezinqumo, futhi ubuyekeze izivikelo ngokuqhubekayo njengoba imodeli yokuziphatha, okulindelwe ngabasebenzisi, kanye nezimfuneko zokulawula zishintsha.

I-Strategic Impact

I-Visual AI ingakwazi ukuhlola, ukutholwa, nokumaka imisebenzi esikalini.

I-Visual AI ingakwazi ukuhlola, ukutholwa, nokumaka imisebenzi esikalini. Ekusetshenzisweni kwekhwalithi ephezulu, lokhu kuhunyushwa emithethweni yokusebenza elinganisekayo, imingcele yobunikazi, nemikhuba yokubuyekeza ephindelelayo ukuze amaqembu akwazi ukukala ukuzethemba esikhundleni sokukala ukungaqondakali.

Amathimba aqanjiwe angakwazi ukulinganisa imiqondo ngokushesha ngezibuyekezo ezimbalwa ezenziwa mathupha.

Amathimba aqanjiwe angakwazi ukulinganisa imiqondo ngokushesha ngezibuyekezo ezimbalwa ezenziwa mathupha. Ekusetshenzisweni kwekhwalithi ephezulu, lokhu kuhunyushwa emithethweni yokusebenza elinganisekayo, imingcele yobunikazi, nemikhuba yokubuyekeza ephindelelayo ukuze amaqembu akwazi ukukala ukuzethemba esikhundleni sokukala ukungaqondakali.

Imisebenzi ingasebenzisa amasiginali wesithombe nawevidiyo obekunzima ukuwenza ngaphambilini.

Imisebenzi ingasebenzisa amasiginali wesithombe nawevidiyo obekunzima ukuwenza ngaphambilini. Ekusetshenzisweni kwekhwalithi ephezulu, lokhu kuhunyushwa emithethweni yokusebenza elinganisekayo, imingcele yobunikazi, nemikhuba yokubuyekeza ephindelelayo ukuze amaqembu akwazi ukukala ukuzethemba esikhundleni sokukala ukungaqondakali.

Ikusasa le-MaskGIT Parallel Token Decoding

Ukwakhiwa kwekhodi okuphindaphindayo okuhambisanayo kwe-MaskGIT kugqugquzele igagasi lamajeneretha angewona ama-autoregressive, okuhlanganisa i-MUSE yombhalo uye esithombeni nezindlela ezifihlekile zevidiyo. Iphethini, ebikezela amathokheni ngokuhambisana futhi icwengisiswa ezinyathelweni ezimbalwa, ihlala phakathi kwe-GAN yeshothi eyodwa kanye nokusabalalisa kwezinyathelo eziningi, ihlinzeka ngokuhwebelana kwekhwalithi-isivinini esiguqulekayo. Lindela ukuqoshwa kwethokheni efihlekile ukuze kuqhubeke kuvela kumajeneretha e-multimodal asheshayo kanye nezinhlelo zokuhlela lapho ukugcwaliswa okungaphakathi nokwemibandela kufaneleka ngokwemvelo.

Ukuqaliswa Komhlaba Wangempela

Ukukhiqiza isithombe esigcwele cishe ngezinyathelo ezihambisanayo eziyi-8 kuye kweziyi-12 esikhundleni samakhulu ezibikezelo zethokheni ezizenzakalelayo

Ukupenda indawo efihliwe yesithombe ngokuphinda ubikezele amathokheni afihliwe anomongo osizungezile

Ukuhlanganiswa kwesithombe esinemibandela yekilasi ku-ImageNet ngekhwalithi yokuncintisana namamodeli anensa kakhulu

Isebenza njengomgogodla wokukhipha amakhodi kumasistimu okuguqula umbhalo ube yisithombe njenge-Google's MUSE edinga ukwenziwa ngokushesha

Amaphethini Okusebenzisa

I-MaskGIT Parallel Token Decoding iyasebenza

Ukukhiqiza isithombe esigcwele cishe ngezinyathelo ezi-8 kuye kweziyi-12 ezihambisanayo esikhundleni samakhulu ezibikezelo zethokheni ezizenzakalelayo.

Ukukhiqiza isithombe esigcwele ngezinyathelo ezihambisanayo ezingaba ngu-8 kuya kweziyi-12 esikhundleni samakhulu ezibikezelo zamathokheni ezizenzakalelayo Amaqembu ngokuvamile athola imiphumela engcono lapho echaza izinga eliphezulu ngaphambili, egcina indlela yokukhuphuka komuntu yamakesi asemaphethelweni, futhi alandelele kokubili izinzuzo zokukhiqiza nezindleko zamaphutha ngokuhamba kwesikhathi.

I-MaskGIT Parallel Token Decoding iyasebenza

Ukupenda indawo efihliwe yesithombe ngokuphinda ubikezele amathokheni afihliwe anomongo osizungezile.

Ukupenda indawo efihlekile yesithombe ngokuphinda ubikezele amathokheni afihliwe kuphela anomongo ozungezile Amaqembu ngokuvamile athola imiphumela engcono lapho echaza ikhwalithi ephezulu ngaphambili, egcina indlela yokukhuphuka yabantu yamakesi asemaphethelweni, futhi alandelele kokubili izinzuzo zokukhiqiza nezindleko zamaphutha ngokuhamba kwesikhathi.

I-MaskGIT Parallel Token Decoding iyasebenza

Ukuhlanganiswa kwesithombe esinemibandela yekilasi ku-ImageNet ngekhwalithi yokuncintisana namamodeli anensa kakhulu.

Ukuhlanganiswa kwesithombe esinemibandela yekilasi ku-ImageNet ngokuncintisana kwekhwalithi namamodeli anensa kakhulu Amathimba ngokuvamile athola imiphumela engcono uma echaza izinga eliphezulu ngaphambili, egcina indlela yokukhuphuka yabantu yamakesi asemaphethelweni, futhi alandelele kokubili izinzuzo zokukhiqiza nezindleko zamaphutha ngokuhamba kwesikhathi.

I-MaskGIT Parallel Token Decoding iyasebenza

Isebenza njengomgogodla wokukhipha amakhodi kumasistimu okuguqula umbhalo ube isithombe njenge-MUSE ye-Google edinga ukwenziwa ngokushesha.

Isebenza njengomgogodla wokukhipha amakhodi wezinhlelo zokuguqula umbhalo ube isithombe njenge-MUSE ye-Google edinga Amathimba akhiqiza ngokushesha ngokuvamile athola imiphumela engcono uma echaza imingcele yekhwalithi ngaphambili, egcina indlela yokukhuphuka yomuntu yamakesi asemaphethelweni, futhi alandelele kokubili izinzuzo zokukhiqiza nezindleko zamaphutha ngokuhamba kwesikhathi.

Izingozi & Guardrails

Amalungelo ezithombe kanye nemvume kungaba ubungozi bezomthetho uma ukuvela kungacacile.

Ukusebenza kwemodeli kungahluka kukho konke ukukhanya, izibalo zabantu, kanye nezindawo.

Okuhle okungelona iqiniso kungase kungabonakali ngaphandle uma izinga lokuzethemba liqashelwa.

Ukuqalisa Umhlahlandlela

Chaza indlela yokwamukela yokunemba, ukukhumbula, nezindleko zamaphutha.

Chaza indlela yokwamukela yokunemba, ukukhumbula, nezindleko zamaphutha. Phatha isinyathelo ngasinye njengesango lobufakazi: uma imibandela ingafinyelelwa, misa ukukhishwa, vala igebe, bese unweba ukusetshenziswa.

Hlola ngedatha efana nezimo zangempela zokukhiqiza.

Hlola ngedatha efana nezimo zangempela zokukhiqiza. Phatha isinyathelo ngasinye njengesango lobufakazi: uma imibandela ingafinyelelwa, misa ukukhishwa, vala igebe, bese unweba ukusetshenziswa.

Engeza isibuyekezo somuntu ukuze uthole ukuzethemba okuphansi noma izibikezelo zomthelela omkhulu.

Engeza isibuyekezo somuntu ukuze uthole ukuzethemba okuphansi noma izibikezelo zomthelela omkhulu. Phatha isinyathelo ngasinye njengesango lobufakazi: uma imibandela ingafinyelelwa, misa ukukhishwa, vala igebe, bese unweba ukusetshenziswa.

Landelela ukukhukhuleka kwemodeli bese uqinisekisa kabusha ngemva kwezinguquko zekhamera noma zesethi yedatha.

Landelela ukukhukhuleka kwemodeli bese uqinisekisa kabusha ngemva kwezinguquko zekhamera noma zesethi yedatha. Phatha isinyathelo ngasinye njengesango lobufakazi: uma imibandela ingafinyelelwa, misa ukukhishwa, vala igebe, bese unweba ukusetshenziswa.

Qhubeka Uhlole

I-Computer Vision

Qonda amasistimu ayisisekelo anika amandla i-AI ebonakalayo.

Funda Umhlahlandlela

I-AI Image Generation

Hlola ukugeleza komsebenzi wokudala kanye nokuhwebelana kwamamodeli.

Funda Umhlahlandlela