Maswali yanayounganishwa na mwongozo · Kati Kiwango

Maswali ya Kupanga Kiwango cha Kujifunza

Hukagua hali ya joto, kosini na ratiba za hatua, anzisha upya tabia, kutokuwa na utulivu wa mapema na kichocheo cha kawaida cha mafunzo ya uvuguvugu-kisha-uozo.

Njia za mwongozo zinazohusianaKupanga Kiwango cha Kujifunza
Swali 1 ya 8

Madhumuni ya awamu ya 'joto' katika ratiba ya kiwango cha kujifunza ni nini?

Hakiki kila swali8 maswali

Majibu na maelezo huonekana unapojibu maswali.

  1. Madhumuni ya awamu ya 'joto' katika ratiba ya kiwango cha kujifunza ni nini?
    • Kuanza na kasi ndogo na kuongeza kasi, kuepuka kuyumba kutokana na viwango vya mapema vya kelele
    • Ili kufanya usahihi wa mwisho kuwa chini
    • Ili kuruka makundi machache ya kwanza ya data
    • Ili kufungia uzani wote mwanzoni
  2. Je, ni ratiba gani inayofuata kwa upole mkunjo wa nusu-kosini kuelekea kiwango cha chini zaidi?
    • Ratiba ya mara kwa mara
    • Cosine annealing
    • Ratiba bila mpangilio
    • Kuongezeka kwa mstari
  3. Nini kitatokea ikiwa kiwango cha kujifunza kimewekwa juu sana?
    • Mafunzo huungana mara moja kwa usahihi kamili
    • Mfano huo hufanya mazoezi polepole lakini hufaulu kila wakati
    • Hasara inaweza kutofautiana au kuzunguka badala ya kupungua
    • Hakuna mabadiliko; kiwango hakina athari
  4. Je, uozo wa hatua unafanya nini kwa kiwango cha kujifunza?
    • Huongeza kwa mstari kila kundi
    • Huiweka sawa kabisa
    • Inaweka nasibu kila hatua
    • Huidondosha kwa kipengele maalum katika nyakati zilizobainishwa awali
  5. What does a warm restart do to a learning-rate schedule?
    • It raises the learning rate again to begin another schedule cycle
    • It resets the model parameters to random values
    • It permanently fixes the learning rate at zero
    • It deletes the optimizer’s moment estimates
  6. In Goyal et al.’s large-minibatch ImageNet setup, what did their warmup scheme address?
    • It removed the need for a validation set
    • It eased early optimization difficulties after the learning-rate scale increased with batch size
    • It changed the model architecture after each epoch
    • It converted the learning-rate curve into a cosine wave
  7. In the cosine formula, what does T denote?
    • The number of model layers
    • Ukubwa wa kundi
    • The schedule horizon or total number of steps used for the decay
    • The optimizer’s momentum coefficient
  8. If the cosine schedule uses lr_min greater than zero, where does the rate approach at the end of the horizon?
    • Exactly zero regardless of lr_min
    • The original warmup starting rate
    • An unbounded value
    • The specified minimum learning rate