quiz bu lëkkale ak tegtal · Diggante Niveau

Reseau Siamese ak quiz perte triplet

Saytu encoder yiñ bokk, distance yiñ dugal, marge triplet yi ak ni misaal yu negatif yu dëgër yi di amee ci siñaalu jàng bi. ndànk.

Yooni njiit yu jëm ciReseau Siamese ak ñàkka am ñatti doom
Laaj 1 de 8

Lan mooy màndarga architecture bu reso Siamese?

Xoolaatal laaj bu nekk8 laaj

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

  1. Lan mooy màndarga architecture bu reso Siamese?
    • Ñaari wala lu ëpp car yu nuróo bokk diisaay
    • Benn car bu amul benn layer bu nëbbu
    • Reseau yu wuute ak poid yu moom seen bopp
    • ab bouclage buy dellusi ci diir bi
  2. Ci triplet perte, lan mooy ñetti dugal yi?
    • Duggal, nëbb, genn
    • Ancre, positif ak négatif
    • Taggat, saytu, saytu
    • Encodeur, decodeur, classificateur
  3. Lan mooy mébetu ñàkka am ñatti doom?
    • Ancre ak negatif dañu wara nuru
    • Bépp embedding dafa wara daanu ci benn poñ
    • Distance ancre-positif warna gëna ndaw distance ancre-négatif ci benn marge
    • Positif ak negatif dañu wara tolloo diggante ancre bi
  4. Lan moo tax 'mining bu dëgër-negatif' am solo ci tàggat triplet perte?
    • Dina dindi bépp lu baax ci tàggat yaram
    • Dafay tax reso bi gëna xóot
    • Dafay wecci metriku distance bi ak benn etiket
    • Negatif yu yomb yi deñuy satisfaire marge bi te duñu defar benn gradient bu am njariñ
  5. Lan moo tax reso Siamese yi am solo ci 'njàng benn yoon'?
    • Mën nañu méngale ab dugal bu bees ak benn royuwaay ba noppi àtte nuru gi, doonte ci ay klaas yuñu gisul
    • Dañu soxla ay junni misaal yuñ etikete ci klaas bu nekk
    • Dañu leen di xaaj ci kategori yuñ tànn bu njëkk
    • Mënuñu yamale lu weesu dàntite tàggat
  6. naka la ko reso Siamese di doxee ndax ñaari dugal dañuy méngoo?
    • Sooy boole seeni valeur pixel yu ñor
    • Ni ñuy méngale diggante seeni vecteur yuñ samp ak benn threshold
    • Ni ñuy xaymaa limu diisaay yiñ jëfandikoo
    • Soo joxee bu nekk etiketu klaas bu bariwul
  7. For triplet loss max(0, d(a,p) − d(a,n) + m), let d(a,p)=0.2, d(a,n)=0.8 and m=0.3. What is the loss?
    • 0.3
    • 0.5
    • 0
    • 1.1
  8. What happens when a negative example is already farther from the anchor than the required margin?
    • The anchor and positive embeddings are discarded
    • The loss forces the negative closer to the anchor
    • The network switches to classification mode
    • The triplet loss becomes zero and that example gives no useful gradient