HejuruUbuyobozi bukurikira
Skeleton-y-Igitekerezo Iringaniza Decoding
Ururimi AI
Amabwiriza ya AI
MaskGIT generates images by predicting many tokens at once and selecting high-confidence predictions for subsequent refinement, replacing left-to-right generation with parallel iterations.
Umva iki gitaboAI-byavuzwe · ~2 min
Confidence is a model score used for selection, not a guarantee that a token is correct.
MaskGIT (Masked Generative Image Transformer), from Google in 2022, rethinks how token-based image models decode. Earlier transformers such as VQGAN generate tokens autoregressively, one at a time in raster order. MaskGIT instead trains with a masked modeling objective like BERT: random subsets of image tokens are hidden and the model learns to predict them simultaneously using bidirectional attention. At generation time it starts from a masked grid and decodes over a configured number of refinement iterations; the original paper reports examples using roughly 8 to 12 steps, but the count depends on the setup rather than defining the algorithm. Each step predicts masked tokens, keeps selected high-confidence predictions, and re-masks other positions. Confidence ranks candidates for the next iteration; it does not prove a prediction is correct.
Gukora ishusho yuzuye muburyo bugera kuri 8 kugeza kuri 12 aho kuba amagana ya autoregressive token
Gushushanya agace gafunze ifoto wongeye guhanura gusa ibimenyetso byihishe hamwe nibidukikije
Urwego-rwerekana ishusho synthesis kuri ImageNet kurwego rwiza rwo guhatana hamwe na moderi zitinda cyane
Gukora nkumugongo wa decoding ya sisitemu yinyandiko-shusho nka __AIU_PROTECTED_11 __ 'MUSE ikeneye ibisekuruza byihuse
MaskGIT ibangikanye na decoding decoding yahumekeye imashini itanga amashanyarazi adasanzwe, harimo MUSE kumyandiko-ku-shusho hamwe nuburyo bwihishe kuri videwo. Igishushanyo, guhanura ibimenyetso mu buryo bubangikanye no gutunganya intambwe nkeya, bicara hagati ya GAN imwe imwe no gukwirakwiza intambwe nyinshi, bitanga ubuziranenge bwihuse-bwihuse. Tegereza ibimenyetso byerekana ibimenyetso byerekana ibimenyetso kugirango ukomeze kugaragara muri moteri yihuta ya moteri nyinshi hamwe na sisitemu yo guhindura aho gushushanya no kuzuza ibintu ari ibintu bisanzwe.
Ibaruwa y'amakuru y'ubuntu
Inkuru eshatu zizewe za AI buri gitondo cy'iminsi y'akazi, zanditse mu rurimi rworoshye. Ubuntu iteka ryose, nta kwamamaza.
Imeri imwe buri munsi y'akazi. Kwiyimura mu buryo bumwe. Ntitwigera tugurisha cyangwa ngo dusangire aderesi yawe.
Isuzume
Ibisubizo ako kanya kuri buri kibazo, n'icyemezo gishobora gusangirwa gifite ID yagenzurwa umaze gutsinda isomo.
Fasha uburezi bwa AI bw'ubuntu. AI Understanding ni umuryango utegamiye kuri leta wa 501(c)(3) — nta kwamamaza, nta kwishyura, iteka ryose. Tanga impano
MaskGIT generates images by predicting many tokens at once and selecting high-confidence predictions for subsequent refinement, replacing left-to-right generation with parallel iterations. Confidence is a model score used for selection, not a guarantee that a token is correct.
MaskGIT ihanura ibimenyetso byose bipfundikijwe icyarimwe hamwe no kwitondera ibyerekezo byombi, bitandukanye na VQGAN itinda ibumoso-iburyo-autodegressive decoding.
MaskGIT ihisha uduce duto duto twikimenyetso kandi yiga kubihanura, intego yo kwerekana icyitegererezo isa na BERT.
At each iteration, MaskGIT retains the predictions with the highest model confidence and re-masks other positions. Confidence is a ranking signal for selecting tokens, not proof that each selected token is correct.
The guide cites roughly 8 to 12 steps as an example from the original paper’s image-generation setups; the count is configuration-dependent rather than a universal requirement.
Gahunda ya cosine igaragaza ibimenyetso bike hakiri kare na nyuma, kuringaniza ubuziranenge n'umuvuduko ukabije.
Komeza wige
Abandi bayobozi batoranijwe kuriyi ngingo
HejuruUbuyobozi bukurikira
Skeleton-y-Igitekerezo Iringaniza Decoding
Ururimi AI