Ikibazo gihuza ibibazo · Hagati Urwego

GLIGEN Ikibazo Cyibisekuru

Shakisha uburyo GLIGEN yongerera umwanya umwanya wo gukwirakwiza inyandiko-ku-shusho mugihe ugumana ubumenyi muburyo bwateganijwe bwahagaritswe.

Inzira zijyanye no kuyoboraGligen Igisekuru
Ikibazo 1 Bya 8

Ni ikihe kibazo GLIGEN ikemura mbere na mbere?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. Ni ikihe kibazo GLIGEN ikemura mbere na mbere?
    • Kubura kugenzura neza aho ibintu bigaragara
    • Buhoro buhoro ibisekuruza
    • Kudashobora gukora amashusho yamabara
    • Kubura amajwi muri videwo yakozwe
  2. Nigute GLIGEN ibika ubumenyi bwikitegererezo cyateganijwe?
    • Igarura uburemere bwose guhera
    • Ihagarika uburemere bwumwimerere kandi ikongeramo imyitozo mishya yatojwe yo kwiyitaho
    • Isiba inyandiko ya kodegisi
    • Ikoresha GAN aho gukwirakwiza
  3. Ni ubuhe buryo busanzwe bwo kwinjiza GLIGEN yemera?
    • Amanota yumuziki
    • Urutonde rwa hyperparameter
    • Agasanduku gahujwe gahujwe ninyandiko cyangwa ishusho yikintu
    • Urupapuro rwerekana agaciro ka pigiseli
  4. Kuki irembo muri GLIGEN rishya ryitondewe ryatangijwe kuri zeru?
    • Gukora amashusho ibara ryambere
    • Guhagarika inyandiko
    • Kugabanya imiterere yishusho
    • Imyitwarire yibanze rero ibikwa hakiri kare mumahugurwa kandi kugenzura byongerwa buhoro buhoro
  5. Niki 'gufungura-isi' ibisekuruza byashingiweho bisobanura muri GLIGEN?
    • Irashobora gushiraho ibitekerezo bisobanuwe uko bishakiye hamwe nuburyo bitagaragara mugihe cyamahugurwa yo hasi
    • Ikora gusa mubyerekanwe murugo
    • Irasaba icyitegererezo cyisi ya 3D
    • Itanga gusa ahantu nyaburanga
  6. Nihehe GLIGEN nshya yugarijwe no kwiyitaho yinjijwe muri U-Net?
    • Gusa kubisohoka VAE
    • Hagati yo kwiyitaho kwariho no guhuza ibitekerezo
    • Imbere muri kodegisi gusa
    • Mbere yo gutegura urusaku
  7. What does a GLIGEN grounding token combine?
    • A denoising timestep with a VAE decoder output
    • A color histogram with the final image resolution
    • A text or image embedding with spatial information such as a bounding box
    • A training dataset name with a classifier label
  8. Besides bounding boxes, which inputs can GLIGEN use to ground objects?
    • Only unstructured audio waveforms
    • A crop-size label without any object information
    • A list of model checkpoint filenames
    • Keypoints or reference images