Quiz ngir mëna ànd ak domen

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

1. Ban jafe-jafe lañu njëkka sos adaptaasioŋu domen ngir saafara?
   • Wàññi limu parametru model yi
   • Gaawaayu bariwaayu matrix ci kaw GPU yi
   • Daanu performance sudee done test yi wuute nañu ak done tàggat
   • Defar etiketu tàggat synthetik ci saasi
2. Ci biir reso domaine-adversaire, lan la couche de réversation de gradient di def?
   • Dafay push extracteur bi ngir ñu mëna tàqale source ak target
   • Dafay efaase domen biy xaaj ci jamonoy inference
   • Dafay yokk ñaari yoon limu jàng ngir gëna gaaw ci tàggat yaram
   • Dafay soppi génnug regression ci etiketu xaaj
3. Lan moo tax adaptaasioŋu domen bu 'ñu dul saytu' am njariñ?
   • Du soxla benn model tàggat yaram
   • Ci done nataal kese lay dox
   • Dina dindi soxla domen source
   • Soxlawul ay done yuñ etikete ci domen biñ bëgga jëfandikoo
4. Ban senaario mooy misaalu coppite domen?
   • Benn model buñu tàggat ba noppi natt ko ci benn done
   • Benn tumër buñu tàggat ci benn scanner te daa lajj ci beneen scanner
   • Benn xeetu daw bu gëna gaaw ginaaw kantite
   • Reseau bu am ay layer yu bari yeneen reso yi
5. Lan moy pexem adaptaasioŋ domen bu 'amul source' biñ defar ngir moytu?
   • Xool done yiñ jëlee ci tàggat yaram ci diiru ànd
   • Jëfandikoo bépp arsitektu reso neuronal
   • Defar bépp xam-xam bu jëm ci génn
   • Dafay daw ci GPU yu bees yi
6. A team starts adaptation from a broadly pretrained model. What should it still measure?
   • Only the parameter count
   • Only source training accuracy
   • Performance on the actual target data and task
   • Whether target evaluation can be skipped
7. Which DANN head supplies the objective to distinguish source examples from target examples?
   • The task label vocabulary
   • The domain classifier
   • The optimizer checkpoint file
   • A fixed output embedding dimension
8. Does gradient reversal mean changing a positive class label into a negative class label?
   • Yes; every task label is negated
   • Yes; all target samples are relabeled as source
   • No; it freezes every feature weight
   • No; it reverses the domain gradient into the feature extractor