quiz bu lëkkale ak tegtal · Diggante Niveau

Adam ak quiz yiy gëna mëna ànd

Dafay saytu jamonoy gradient yu Adam, saafara jafe-jafe yi, eskalaasioŋ biy méngoo, epsilon ak yeesali AdamW biy yàq diisaay bi. ci yeesali parametre buñ miin.

Yooni njiit yu jëm ciAdam ak Optimiseur yuy méngoo
Laaj 1 de 8

Yan ñaari limu Adam di topp ci parametre bu nekk?

Xoolaatal laaj bu nekk8 laaj

Sooy def quiz bi, tontu yi ak leeral yi dañuy feeñ.

  1. Yan ñaari limu Adam di topp ci parametre bu nekk?
    • Moyenne biy daw ci degrade yi ak degrade kaare yi
    • Degrade bi gëna mag ak bi gëna ndaw biñ musa gis
    • Tolluwaayu jàng bi ak dayo lote bi
    • Diisaay bi ak màndargam
  2. Lan moo tax Adam di jëfandikoo njuumte ci xaymaam ci saa si?
    • Ndax gradient yi dañuy nekk negatif ci ndoorte tàggat yaram
    • Ndax moyenne yiy toxu dañu tàmbali ci zero ba noppi dañu jëm ci zero ci ndoorte li
    • Ngir yokk limu jàng ci diir bu gàtt
    • Ngir dindi ay poñ done yu génn ci mbooloo mi
  3. Lan mooy wuutale Adam ak AdamW?
    • AdamW jëfandikoowul benn momentum
    • AdamW dindi na ñaareelu xayma
    • AdamW dafay dindi diisaay bi ci yeesal bi lalu ci degrade
    • AdamW du dox ludul ci reso convolutionnel
  4. Ban cër la tur epsilon bu ndaw bi di def ci sàrtu yeesal bu Adam?
    • Dafay tëral dooley momentum bi
    • Dafay saytu limu jamonoy tàggat yaram
    • Dafay yokk magnitude gradient bi
    • Dafay tere xaaj ak nul ba noppi dakkal yeesal bi
  5. Lu tax ñuy faral di wax ci Adam ni 'mën ànd ak jamono'?
    • Dafay jox parametre bu nekk dayo jéego boppam
    • Dafay soppi architecture reso bi ci diiru tàggat yaram
    • Dafay méngale done yi ak xeetu model bi
    • Dafay weccoo ci saasi diggante xaaj ak régression
  6. Which Adam parameter controls the exponential decay rate of its squared-gradient estimate?
    • Beta1
    • Beta2
    • Epsilon
    • The weight-decay coefficient
  7. A parameter repeatedly receives large gradients. Which Adam quantity helps scale down its effective step?
    • The number of model layers
    • The batch-normalization mean
    • The second-moment estimate of squared gradients
    • The AdamW weight-decay term alone
  8. What happens when beta2 is set closer to one, all else equal?
    • The optimizer stops using gradients
    • The first-moment estimate is deleted
    • Weight decay becomes decoupled automatically
    • The squared-gradient estimate decays more slowly and retains a longer history