quiz bu lëkkale ak tegtal · Diggante Niveau

Laajteb tuning bu gaaw

Xoolaat laajte yu woyof yi ci lalu duggal bi, tabb model bu gelé bi, ak njariñ yi aju ci eskaal bi ñu xamle ngir gaaw ci ajustement bi.

Yooni njiit yu jëm ciTuning bu gaaw
Laaj 1 de 8

Luy 'soft prompt' ci tuning bu gaaw?

Xoolaatal laaj bu nekk8 laaj

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

  1. Luy 'soft prompt' ci tuning bu gaaw?
    • Jàng nañu ay vecteur yuy wéy di boole te soxla wuñu ñu boole leen ci kàddu yu dëggu yi
    • Tegtal buñ bind ci kàddu yu rafet
    • ab kopi distiye bu gëna ndaw ci xeetu model bi
    • ab ensemble de logits de sortie congélé
  2. Fan la tuning bu gaaw bi di yokk ay vecteur yuñ mëna tàggat?
    • Ci bepp lalu transformatër
    • ci layer biy dugal rek
    • Ci couche de sortie bi kese
    • Ci biir etaa optimisatër
  3. Ban njeexte bu yéeme la këyit bi njëkk fësal ci wàllu eskaalu model bi?
    • Tuning bu gaaw bi dafay gëna bon lu model yi di màgg
    • Escale amul benn njeexital ci tuning bu gaaw bi
    • Lu model yi di gëna rëy, tuning bu gaaw bi dafay jegesi kalite tuning bu mat sëkk
    • Modèle yu gëna mag mënatuñu leen gaaw defar
  4. Kan moo dugal tuning bu gaaw?
    • Xel mu xóot ci 2016
    • Anthropic in 2023
    • Microsoft in 2019
    • Gëstukat yi ci Google (Lester, Al-Rfou, Constant) ci atum 2021
  5. naka lay def ba mëna def ay liggéey yu bari ci benn model?
    • Liggéey bu nekk dafay jëfandikoo ab prompt bu ndaw bu yomb ci jamono ji model bi di des
    • Dafay tàggataat model bi yépp ngir laaj bu nekk
    • Dafay efaase liggéey yiñ jëfandikoowul ci poid yi
    • Dafay laaj GPU bu wuute ci liggéey bu nekk
  6. How does Lester and colleagues’ paper describe prompt tuning in relation to prefix tuning?
    • As a replacement for frozen language models
    • As a simpler related method that the paper compares with prefix tuning
    • As a tokenizer with no trainable parameters
    • As a method that updates every model weight
  7. Which model family did Lester and colleagues use to study prompt tuning across scales?
    • WaveNet
    • AlphaFold2
    • T5
    • àllu isolation
  8. How are soft prompts optimized in prompt tuning?
    • By changing all model weights for each task
    • By replacing embeddings with hard-coded words only
    • By training a new tokenizer for every example
    • By backpropagation through a frozen language model to update the prompt vectors