Ikibazo gihuza ibibazo · Hagati Urwego

Ikibazo cyo Kumenyekanisha Ubumenyi

Koresha ikwirakwizwa rya mwarimu, intego zoroheje, hamwe nisuzuma ryabanyeshuri muri disillation, harimo n'impamvu kwigana bidashobora kwemeza urwego rwabarimu.

Inzira zijyanye no kuyoboraGukwirakwiza ubumenyi
Ikibazo 1 Bya 8

Mu gusibanganya ubumenyi, ni ikihe cyitegererezo cyabanyeshuri batojwe gukora?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. Mu gusibanganya ubumenyi, ni ikihe cyitegererezo cyabanyeshuri batojwe gukora?
    • Gutoza icyitegererezo gito cyabanyeshuri kwigana icyitegererezo kinini cyabarimu
    • Gukora icyitegererezo kinini kandi cyuzuye
    • Kuranga amakuru mu buryo bwikora
    • Kugirango uhishe uburemere bwikitegererezo
  2. Bisobanura iki 'ubumenyi bwijimye' bwa mwarimu?
    • Imyitwarire mibi yihishe murugero
    • Igice cyibanga cya dosiye yicyitegererezo
    • Ibisobanuro muburyo bushoboka mwarimu agenera amasomo atariyo
    • Amakuru umwarimu ntabwo yigeze ahugurwa
  3. Ni uruhe ruhare ubushyuhe (T) bugira mu gusibanganya?
    • Igenzura umuvuduko wimyitozo yicyitegererezo gusa
    • Yoroshya softmax kuburyo buto hagati y-ibyiciro bisa biga
    • Ishiraho umubare wibyiciro mubanyeshuri
    • Ipima ubushyuhe bwa GPU
  4. Igihombo cya classique ya classique ihuza ibice bibiri?
    • Uburyo bubiri butandukanye bwo guhemba
    • Gereranya ikosa rya kwaduka no gutakaza hinge
    • Igihano cya KL kirwanya politiki yerekana nigikorwa cyagaciro
    • Ijambo KL-gutandukana rihuye na mwarimu woroheje wongeyeho cross-entropy kuri label yukuri
  5. Niyihe ngero yicyitegererezo cyururimi?
    • ResNet-50
    • Ijambo2Vec
    • Yamagata
    • AlexNet
  6. Ni mu buhe buryo ibintu bishingiye ku gutandukanya ibintu bitandukanye nuburyo busanzwe busohoka?
    • Ihuza intera ihishe-igereranya, ntabwo ibisubizo byanyuma
    • Irengagiza rwose mwarimu
    • Ikoresha gusa ibirango bikomeye
    • Itoza umwarimu aho kuba umunyeshuri
  7. A teacher assigns similar small probabilities to wolf and fox when predicting dog. What can soft-target distillation transfer beyond the top label?
    • A guarantee that either alternative is the correct label
    • The teacher’s complete original training corpus
    • The teacher’s exact number of hidden layers
    • Relative relationships among the teacher’s alternative class scores
  8. A distilled student is smaller than its teacher. Which evaluation is needed before deployment?
    • Assume its accuracy must equal the teacher’s
    • Measure its task accuracy and runtime on the intended system
    • Assume fewer parameters guarantee lower latency on every device
    • Only compare the models’ names