Kediska Celceliska Miisaanka Stochastic

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

1. Waa maxay celceliska celceliska miisaanka Stochastic?
   • Miisaanka tusaalaha ayaa laga soo qaatay dhibco badan goor dambe ee tababarka
   • Saadaasha dhowr nooc oo si gaar ah loo tababaray
   • Gradients ayaa lagu xisaabiyay dhammaan dufcadaha yaryar
   • Heerarka waxbarashada ee la isticmaalay xilli kasta
2. Waa maxay sababta SWA ay u janjeerto inay horumariso guud?
   • Waxay si sax ah u xifdisanaysaa habka tababarka
   • Waxa ay helaysaa minima ka ballaadhan, faashaxsan oo aad ugu adag xogta aan la arkin
   • Waxay kordhisaa tirada qiyaasaha ee qaabka
   • Waxay meesha ka saaraysaa baahida loo qabo go'an ansixinta
3. Waa maxay tillaabada dheeraadka ah ee ay SWA uga baahan tahay shabakadaha isticmaala dufcadda caadiga ah?
   • Dib-u-tababarka shabakada xoqan
   • Laba-laabista cabbirka dufcadda
   • Baadh dheeraad ah oo hore loo sii marayo si dib loogu xisaabiyo tirakoobka socodsiinta BN
   • Ka saarida dhammaan lakabyada ka-tagidda
4. Waa maxay dabeecadda heerka-wax-barashada ee sida caadiga ah loo isticmaalo inta lagu jiro marxaladda celceliska SWA?
   • Heerkii waxbarashadu wuu xumaaday oo noqday eber
   • Heerka waxbarasho ee mid ka mid ah xasilloonida
   • Wax barasho haba yaraatee ma jiro tan iyo markii tababarka la qaboojiyay
   • Joogto ah ama meerto ah, heerka waxbarasho ee xad sare
5. Sidee buu qiimaha wax-ku-darka SWA u barbar dhigaya isku-dhafka dhaqameed ee moodooyinka N?
   • Waxay ku kacaysaa qiyaas ahaan hal nooc, maadaama natiijadu tahay hal miisaan
   • Waa N jeer ka qaalisan
   • Waxay la mid tahay socodsiinta moodooyinka N
   • Looma isticmaali karo fikrad ahaan
6. Why does the batch-normalization adjustment belong after SWA weight averaging?
   • The averaging step deletes every convolution
   • Its running statistics must be recomputed for the averaged weights
   • It converts batch normalization into layer normalization
   • It restores the original optimizer learning rate
7. An SWA running mean combines n previous samples with a new weight vector w_i. Which update gives equal weight to all n+1 samples?
   • w_SWA + n · w_i
   • (w_SWA + w_i) divided by n
   • (n · w_SWA + w_i) divided by (n + 1)
   • n · (w_SWA − w_i)
8. Why is SWA commonly started during the later part of training?
   • It requires the model to remain at random initialization
   • It can only average models with different architectures
   • It replaces the training data with validation examples
   • It averages nearby useful checkpoints from a trajectory after the model has made substantial progress