quiz bu lëkkale ak tegtal · Diggante Niveau
Decay poids ak quiz buy yamale L2
Test L2 daan, réduction ci zero, SGD méngoo, AdamW decoupling, scaling adaptif, ak dooley yàqu ci tàggat yaram.
Laaj 1 de 8
Lan la L2 regularisation yokk ci fonction perte bi?
Xoolaatal laaj bu nekk8 laaj
Sooy def quiz bi, tontu yi ak leeral yi dañuy feeñ.
- Lan la L2 regularisation yokk ci fonction perte bi?
- Limu couche yi ci reso bi
- Daan ci ni ñuy jàngee
- Limu misaali tàggat yi
- Daan bu méngoo ak limu poid kaare yi
- Lan mooy jafe-jafe bi gëna mag bi yaram mëna moytu?
- Done yu yeex
- Overfitting, fu model bi di jottali bruit bi ci done yiñ tàggat
- ay jafe-jafe ci aparey bi
- Etiket yu jaarul yoon
- Ci ban wàll la diisaay biy dakkal di puus diisaayu model bi?
- Ci wàllu infini
- Ci diisaay bi gëna mag
- Ci wàllu zero
- Ci wàllu valeur yu bari
- Lu tax ñu dugal AdamW?
- Ngir dindi yàqu-yàqu ci diisaay bi ak yeesali gradient biy méngoo, suko defee yàqu-yàqu bi doxalee nimu waree
- Ngir dindi degrade bi lëmm
- ngir yeexal tàggat yaram
- Ngir dindi soxla taxawaayu jàng
- Ngir wàccinu gradient stochastic bu leer, naka la L2 di doxalee ak diisaay biy wàññeeku?
- Dañuy wuute saa yu nekk
- ci wàllu math dañuy nuru
- Diisaay bi dafay dindi degrade bi
- L2 du dox ludul ci SGD ak momentum
- Lan la koeffisien lambda di saytu ci yamale L2?
- Dayo lote bi
- limu jamono yi
- Fonksioŋu aktivasioŋ bi
- Dooley penalti ci poid yu rëy
- Why can adding an L2 term to Adam’s loss differ from applying decoupled weight decay?
- Adaptive per-parameter gradient scaling can distort the coupled penalty
- Adam cannot train with any regularization
- The two methods use different training labels
- Decoupled decay removes the data loss
- In a simple decoupled update, what changes the fractional shrinkage of a weight?
- Only the number of classes
- Only the model’s input dimension
- The learning rate and decay coefficient together
- Only the validation-set size