Kedis ku xidhan hagaha · Adag Heerka

Kediska Yaraynta Ogaanshaha-Foonaanta

Waxay hubisaa cusboonaysiinta labada-baas ee SAM, radius xaafadeeda, culayska qallafsan iyo isdhaafsiga u dhexeeya simannaanta iyo xisaabinta.

Wadooyinka haga ee la xidhiidhaYaraynta Ogaanshaha Fiican
Su'aal 1 ee 8

Waa maxay nooca ugu yar ee SAM isku dayo inuu helo?

Horudhac su'aal kasta8 su'aalo

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

  1. Waa maxay nooca ugu yar ee SAM isku dayo inuu helo?
    • Ugu yaraan fidsan halkaas oo khasaaruhu uu ahaado mid hooseeya marka loo eego xaafadaha miisaanka
    • Ugu yar ee suurtogalka ah
    • Inta ugu yar ee leh cabbirada ugu badan
    • Wax kasta oo ugu yar ee lagu gaari karo hal tallaabo
  2. Immisa baas oo hore-u-dhac ah ayay u baahan tahay tallaabada SAM ee caadiga ah?
    • Baas badhkii
    • Qiyaastii laba
    • Dhab ahaan hal
    • Toban ama ka badan
  3. Muxuu ka taliyaa radius hyperparameter rho gudaha SAM?
    • Xawaaraha qudhunka ee waxbarashada
    • Tirada waayihii tababarka
    • Baaxadda xaafadda ku hareeraysan miisaanka oo ka ilaalinaya caarada luminta
    • Suurtagalnimada ka-tagidda
  4. Waa maxay ta koowaad ee labada tallaabo ee SAM ee soo noqnoqon kasta?
    • Iska yaree qaabka
    • Celcelis ahaan miisaanka dhammaan talaabooyinkii hore
    • Dib u dajin gobolka wanaajinta
    • Qaad tallaabo kor u qaad cabbir cabbir rho ah oo u jeedda meesha ugu xun ee u dhow
  5. Maxay SAM inta badan u wanaajisaa adkaanta si ay u calaamadiso buuqa?
    • Sababtoo ah minima dabaqa ah ayaa aad ugu nugul xifdinta calaamadaha la musuqmaasuqay
    • Sababtoo ah waxay si toos ah u tirtirtaa calaamadaha buuqa badan
    • Sababtoo ah waxay isticmaashaa ereyo ka sii weyn
    • Sababtoo ah waxay saddex laabmaysaa cabbirka dufcadda
  6. Waa kuwee kuwan kala duwanansho loogu talagalay in lagu dhimo kharashka xisaabinta SAM?
    • ReLU
    • ESAM (SAM hufan)
    • BatchNorm
    • AdaGrad
  7. In SAM’s second pass, which gradient is used to update the original model weights?
    • The gradient of a random label
    • The first pass gradient without evaluating the perturbed point
    • The gradient evaluated at the perturbed nearby weights
    • A gradient from a different model checkpoint
  8. If SAM’s neighborhood radius rho is set to zero, what happens to its perturbation step?
    • The model parameters are all set to zero
    • The loss becomes independent of the weights
    • The optimizer takes an infinite step
    • There is no weight perturbation, so the neighborhood reduces to the current point