quiz bu lëkkale ak tegtal · Diggante Niveau

Kubernetes ngir ML

Jëfandikool plasement pod Kubernetes ML, jumtukaayi aparey GPU, ak muñ taint, boole ci li tax tàggat yaram soxla kàttan bu méngoo ak node bu nekk.

Yooni njiit yu jëm ciKubernetes ngir ML
Laaj 1 de 8

Lan mooy liggéey bi njëkk ci Kubernetes scheduler ngir ML?

Xoolaatal laaj bu nekk8 laaj

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

  1. Lan mooy liggéey bi njëkk ci Kubernetes scheduler ngir ML?
    • Bind kodu tàggat ngir gëstukati done
    • Xool ban node ci cluster bi mooy doxal pod bu nekk
    • Etiketu done tàggat yaram ci saasi
    • Tann arsitektir reso neuronal
  2. naka la ko Kubernetes di def ba GPU yi mëna am ci pod bi?
    • Jaaraleko ci ay plugin yuy siiwal ay jumtukaay yu melni nvidia.com/gpu
    • Soo bataaxal jaaykatu GPU bi ngir am lisence
    • Mënu jëfandikoo GPU ci Kubernetes
    • Sooy soppi core CPU yi ci GPU yi ci runtime
  3. With a NoSchedule taint on a GPU node, what does a matching pod toleration do?
    • Forces the pod onto that node regardless of capacity
    • Changes the model’s optimizer
    • Allows the pod to be considered for that node without guaranteeing placement
    • Encrypts every GPU tensor
  4. Ban jumtukaay mooy yokk kàttanu ML yu melni operatëri tàggat yuñ séddale ci kaw Kubernetes?
    • Word
    • Apache ngir joxe xëti web yu fippu
    • makro këyitu xayma
    • Kubeflow
  5. Lan moo tax Kubernetes méngoo bu baax ak liggéey yu bari ci ML?
    • Dafay jàppale liggéey yi ci jumtukaay yu amul fayda ba noppi bàyyi aparey bi su liggéey yi jeexee
    • Dafay denc GPU yépp ngir benn jëfandikukat
    • Dafay tere bépp liggéey wàññeeku
    • Benn conteneur kese lay def ci cluster bu nekk
  6. Lan mooy 'yeesal ab poñ dellu ginaaw ci nul' ci Kubernetes?
    • efaase model bi ba fàww
    • Def njub ci 0 pursaa
    • Doxal zero kopi sudee trafik amul, ba noppi wëlbatiku ci laaj bi ci topp
    • Wàññi parametru model bi ba amul dara
  7. One ML pod requests four GPUs, while two nodes each have only two available. Can ordinary scheduling combine those nodes for that one pod?
    • Yes; every pod automatically spans all cluster nodes
    • Yes; the scheduler converts CPU cores into the missing GPUs
    • Only if the pod name contains the word distributed
    • No; the pod must fit on one eligible node
  8. An inference service scales to zero while idle. Which user-facing delay should its team test when traffic resumes?
    • The training-set labeling speed
    • Cold-start delay while a serving replica becomes ready
    • The file size of an unrelated notebook
    • The number of discarded vocabulary tokens