Ntuziaka Visual AI

Ihe onyonyo ederede-gaa na onyonyo

Imagen bụ sistemu ederede gaa na onyonyo nke Google na-atụgharị nkọwa ederede ka ọ bụrụ eserese foto.

2 nkeji na-agụEmelitere ikpeazụ

Nchịkọta

Its headline finding was that a large frozen language model, not a bigger image network, was the biggest driver of quality.

Ime miri emi

Nchoputa nke Google mara ọkwa na 2022, Imagen gosiri na ịghọta ngwa ngwa ihe dị mkpa dịka ịbịaru ya nke ọma. Kama koodu ntinye ederede ụdị CLIP, Imagen na-eji nnukwu ntinye ederede a zụrụ azụ (T5-XXL) nke edobere oyi, wee na-enye ntinye asụsụ ndị ahụ bara ụba ka ọ bụrụ ụdị mgbasa ozi. Ọ na-ewepụta obere onyonyo 64x64 ma na-eji usoro mgbasa ozi nwere nnukwu mkpebi abụọ iji bulie elu ruo 1024x1024. Ndị otu ahụ webatara 'ọnụ ụzọ dị ike' iji mee ka agba kwụsie ike na ntuzi aka dị elu, wee wuo DrawBench, akara ngosi nke aghụghọ na-akpali agụgụ ule, mmekọrịta gbasara ohere, yana ngwakọta na-adịghị ahụkebe. Ụdị ndị ọzọ, Imagen 2 na Imagen 3, nkọwa dị nkọ, nsụgharị ederede, na ikwesị ntụkwasị obi ozugbo, ma nwee ike ugbu a ngwaọrụ oyiyi Google.

Nghọta nka nka

Nhọrọ pụtara ìhè nke Imagen bụ ibelata ederede ederede karịa ihe na-emepụta ihe oyiyi. T5-XXL, nke a zụrụ naanị na ederede, na-emepụta ntinye nke na-ejide asụsụ dị nro, ndị nchọpụta ahụ chọpụtakwara na ịgbasawanye ya na-eme ka nhazi ihe oyiyi-ederede dịkwuo mma karịa ịgbasa ụdị mgbasa ozi. A na-agbaji ọgbọ: ụdị mgbasa ozi dị ala na-eme onyonyo dị ala, mgbe ahụ ụdị mgbasa ozi nke ukwuu na-eji nke nta nke nta na-ebuli ya elu, yana ụkpụrụ pikselụ na-amachi ike siri ike iji zere nsonaazụ nchacha n'okpuru nduzi siri ike.

Mmetụta atụmatụ

Ọsọ na ọnụ ọgụgụ

Visual AI nwere ike megharịa nyocha, nchọpụta na mkpado ọrụ n'ọtụtụ.

Mee nhọrọ

Otu ndị na-emepụta ihe nwere ike imepụta echiche ngwa ngwa site na ngbanwe akwụkwọ ntuziaka ole na ole.

Team na usoro ọrụ

Ọrụ nwere ike iji onyonyo na akara vidiyo siri ike ịhazi.

Ọdịnihu nke Imagen Ederede gaa na onyonyo

Usoro nke Imagen na-aga n'ihu na-atụgharị ederede ka mma n'ime onyonyo, na-esochi ngwa ngwa maka ihe nkiri dị mgbagwoju anya, yana nlele ngwa ngwa. Na-atụ anya njikọta miri emi na ụdị asụsụ ka sistemụ ahụ ' kpatara' maka arịrịọ tupu eserese, yana akara mmiri siri ike dị ka SynthID maka provenance. Ka ọ na-ejikọta n'ofe ngwaahịa Google yana Gemini, ihe a na-elekwasị anya na-atụgharị na ọgbọ a pụrụ ịdabere na ya, nchekwa, na-achịkwa kama ịbụ ihe ọhụrụ ọhụrụ.

Mmejuputa n'ezie n'ụwa

Na-emepụta ihe nleba anya ahịa nke fotorealistic site na nkenke ederede na-enweghị ịse foto

Ịmepụta ihe atụ echiche maka ịkọ akụkọ ma ọ bụ akwụkwọ ụmụaka sitere na ahịrịokwu nkọwa

Na-emepụta ihe njakịrị ngwaahịa na mgbanwe ọnọdụ maka ndepụta e-azụmahịa

Ilele echiche sayensị ma ọ bụ nkuzi, dị ka nsụgharị onye nka akọwara n'asụsụ dị larịị

Ihe ize ndụ & okporo ụzọ nche

Ikike onyonyo na nkwenye nwere ike bụrụ ihe egwu dị n'iwu ma ọ bụrụ na edoghị anya.

Ọrụ nlereanya nwere ike ịdịgasị iche n'ofe ọkụ, igwe mmadụ, na gburugburu.

Enwere ike ghara ịhụ ihe dị mma ma ọ bụrụ na enyochaghị oke ntụkwasị obi.

Map mmejuputa

1

Kọwaa ụkpụrụ nnabata maka nkenke, icheta, na ụgwọ njehie.

2

Nwalee na data dabara na ọnọdụ mmepụta n'ezie.

3

Tinye nyocha mmadụ maka obere obi ike ma ọ bụ amụma mmetụta dị elu.

4

Sochie ihe nlere anya wee megharịa ka emechara mgbanwe igwefoto ma ọ bụ dataset.

Nọgide na-eme nchọpụta

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 Imagen Text-to-Image quiz

Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.

Malite ajụjụ

Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation

Ntuziaka na-esote

Imagen 2 na Mgbasa Nkwụghachi ụgwọ

Ajụjụ a na-ajụkarị

What is Imagen Text-to-Image?

Imagen bụ sistemu ederede gaa na onyonyo nke Google na-atụgharị nkọwa ederede ka ọ bụrụ eserese foto. Nchọta isi okwu ya bụ na nnukwu ụdị asụsụ oyi kpọnwụrụ, ọ bụghị netwọkụ onyonyo ka ukwuu, bụ ihe kacha akwọ ụgbọ ala.

Gịnị bụ nchọpụta kacha emetụta Imagen?

Imagen gosiri na ibelata nghota asụsụ site na T5-XXL mere ka arụpụta ihe ka mma karịa ịgbasa ụdị mgbasa ozi.

Kedu koodu ntinye ederede Imagen na-adabere na ya?

Imagen na-eji nnukwu ihe ntinye ederede naanị T5-XXL a zụrụ azụ, nke akpọnwụwo n'oge ọzụzụ.

Kedu ka Imagen si enweta mkpebi dị elu?

Ọ na-ewepụta onyonyo 64 × 64 wee bulie elu site na ụdị mgbasa ozi nwere nnukwu mkpebi ruo 1024 × 1024.

Kedu ihe eji eme 'dynamic thresholding' na Imagen?

Ntụnye ọnụ ụzọ na-agbanwe agbanwe na-ejide ụkpụrụ pikselụ ya mere ntụzịaka dị elu anaghị ewepụta onyonyo ekpochapụla.

Kedu ihe bụ DrawBench?

DrawBench bụ akara ngosi Google ewebata ya na Imagen iji nwalee ngụkọ, mmekọrita oghere na mkpali dị ụkọ.