Kedis ku xidhan hagaha · Dhexdhexaad Heerka

Kediska Models-Luuqada CLIP iyo Aragtida

Dib u eeg metelaadyada la wadaago ee CLIP iyo isbarbardhigga lammaane, oo wata sumadaha eber-ka-soo-baxa, qoysaska codeeyaha asalka ah, tirada dhibcaha-dhibcaha, iyo unCLIP.

Wadooyinka haga ee la xidhiidhaQaababka Luuqadda Clip Iyo Aragtida
Su'aal 1 ee 8

Waa maxay fikradda ugu muhiimsan ee ka dambeysa CLIP?

Horudhac su'aal kasta8 su'aalo

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

  1. Waa maxay fikradda ugu muhiimsan ee ka dambeysa CLIP?
    • Ku cadaadi sawirada faylal yaryar
    • Gelida sawirada iyo qoraalka meel isku mid ah oo la wadaago
    • Samaynta moodooyinka 3D ee sawirada
    • U turjumaya hadalka qoraal
  2. Waa maxay habka tababarka ee CLIP adeegsato?
    • Xoojinta barashada jawaab celinta aadanaha
    • Gacanta ku calaamadinta shay kasta oo sawir kasta ah
    • Barashada is-khilaafsan ee soo jiidata lammaane qoraal-sawir ah oo is-bar-bar socda
    • Ka koobiyeynta miisaanka qaabka hadalka
  3. What does zero-shot classification mean in the original CLIP transfer setup?
    • Using text-defined candidate categories without new task-specific supervised classifier fitting
    • Using a model that never received any pretraining data
    • Running without any computation or electricity
    • A guarantee that every category is classified correctly
  4. Sidee buu CLIP u cabbiraa in sawirka iyo qoraalku is-waafaqayaan?
    • Marka la barbardhigo cabbirka faylka
    • Adigoo tirinaya pixels midab kasta
    • Adigoo hubinaya shaambada wakhtiga sawirka
    • Marka la eego isku ekaanshaha u dhexeeya vectoryadooda
  5. Qiyaastii xog intee le'eg ayaa CLIP lagu tababaray?
    • Dhowr kun oo sawir gacmeed lagu calaamadeeyay
    • Qoraal kaliya, sawiro malaha
    • Ku dhawaad 400 milyan oo sawir-qoraalka lammaane ee shabakada
    • Matxafka kaliya ee la soo ururiyey
  6. How does the cited unCLIP image-generation paper use CLIP representations?
    • The original standalone CLIP encoder directly outputs every image pixel
    • A prior supplies CLIP image embeddings to a generative decoder
    • It removes all image representations and uses only filenames
    • It uses CLIP only as a network-encryption protocol
  7. Which image-encoder families were evaluated in the original CLIP paper?
    • Only recurrent text encoders
    • Only a mandatory linear pixel average
    • Only a raw audio codec
    • Modified ResNets and Vision Transformers
  8. A CLIP batch contains 32 paired images and texts. How many entries are in its full cross-modal score matrix?
    • 1024
    • 32
    • 64
    • 992