Ikibazo gihuza ibibazo · Hagati Urwego

MaskGIT Ikigereranyo cya Token Decoding Ikibazo

Reba MaskGIT ya masike-token imyitozo hamwe na itode igereranya, harimo gahunda yayo yo gusubiramo ibyiringiro hamwe nibice byombi.

Inzira zijyanye no kuyoboraMaskgit Iringaniza Token Decoding
Ikibazo 1 Bya 8

Nigute MaskGIT yerekana ibimenyetso byamashusho bitandukanye na transformateur ya VQGAN?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. Nigute MaskGIT yerekana ibimenyetso byamashusho bitandukanye na transformateur ya VQGAN?
    • Ihanura ibimenyetso muburyo bubangikanye kandi binonosora intambwe nke
    • Ihanura ikimenyetso kimwe icyarimwe murutonde rwa raster
    • Ntabwo ikoresha transformateur
    • Itanga inyandiko gusa
  2. Niyihe ntego yo guhugura MaskGIT itira muburyo bw'ururimi?
    • Ibikurikira-ijambo-guhanura gusa
    • Ibimenyetso byerekana ibimenyetso, nka BERT
    • Itondekanya amashusho
    • Kwiga gushimangira ibitekerezo byabantu
  3. Mugihe cyibisekuru, ni ibihe bimenyetso MaskGIT ikora kuri buri itera?
    • Abizera cyane
    • Ibimenyetso bisanzwe
    • Ibyahanuwe cyane
    • Ibimenyetso byose icyarimwe kumurongo wambere
  4. In the original MaskGIT generation examples, roughly how many refinement iterations were often used?
    • Exactly one iteration for every image and task
    • Hundreds of sequential token predictions
    • A fixed number that cannot be changed by the user
    • About 8 to 12 in those reported configurations
  5. Nihe gahunda igenzura ibimenyetso bingahe MaskGIT ihishura buri ntambwe?
    • Gahunda yo guhisha cosine
    • Umurongo-gusa amategeko ahishura icyarimwe
    • Nta gahunda; ntibisanzwe
    • Ikimenyetso kimwe gihamye kuri buri ntambwe
  6. Kuki kwitondera ibyerekezo byombi bifasha MaskGIT?
    • Ihisha ishusho kurugero
    • Kureka buri kimenyetso kibona ishusho igice cyose kugirango ubone imiterere myiza
    • Ireba gusa ibimenyetso byambere
    • Ikuraho ibikenewe byose
  7. What token grid does MaskGIT begin with when generating an image?
    • A fully decoded image that only needs color correction
    • A left-to-right stream with no masked positions
    • A grid with every image token masked
    • A grid of text embeddings from the prompt only
  8. How can MaskGIT fill a masked region during image inpainting?
    • Discard the observed image and generate an unrelated full image
    • Use only a left-to-right text decoder with no image tokens
    • Replace every visible token with a random codebook entry
    • Keep observed image tokens as context and iteratively predict the masked positions