quiz bu lëkkale ak tegtal · Diggante Niveau

Jàngalekat biy forse ci quiz model yu toppalante

Saytu ni prefix dëgg-suuf di tàggatee ab dekodeer bu toppalante, ni mask causal di jàppale tàggat paralel, ak lu tax inference wuute.

Yooni njiit yu jëm ciJàngalekat bi di forse ci model yu toppalante
Laaj 1 de 8

Bu jàngalekat bi di forse, lan mooy dugal ci jéego bu nekk ci dekodaas bi?

Xoolaatal laaj bu nekk8 laaj

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

  1. Bu jàngalekat bi di forse, lan mooy dugal ci jéego bu nekk ci dekodaas bi?
    • Ground-truth token bi njëkk
    • Xalaatal bu njëkk bu model bi
    • Jeton bu bari
    • Jeton biy mujjee génne
  2. During teacher-forced training, what target is paired with decoder input y_(t−1)?
    • The previous ground-truth token
    • The next target token y_t
    • A token sampled from the validation set
    • The maximum-probability output from an earlier training run
  3. Jafe-jafe bi jàngalekat bi di forse moo waral examen train bi, ñu xamee ko ci:
    • Degrade yuy réer
    • Mode daanu
    • Exposition bias
    • Fuite etiket
  4. Ci decodeur Transformer, ban mecanisme mooy may jàngalekat bi di forse tàggat yaram mu dox ci paralel ci position yi?
    • Dàddu
    • Normalisasioŋ ci lots
    • Seetug limyéer
    • Masku bàyyi xel ci sabab (xool kanam)
  5. Ci jamonoy inference, lu tax mënu ñu jëfandikoo forcing jàngalekat?
    • Amul benn dëgg-dëgg ëlëg token ngir dundal
    • Fonction perte bi amul
    • GPU bi yeexna lool
    • Masku bàyyi xel yi deñuy tëju
  6. Ban perte lañuy gëna jëfandikoo ci jéego bu nekk ci diiru tàggat bi jàngalekat bi di forse?
    • Jumtukaay kaare moyen ci logits yu ñor
    • Entropie croisée ci digganté séddale biñu séentu ak token bi ci topp dëgg
    • Perte de charnière sur l'intelligation
    • cosin nuróo
  7. The target sequence is “red, blue, green.” Under teacher forcing, which token is supplied when training the prediction of “green”?
    • The model’s prediction for “green”
    • The future token after “green”
    • The ground-truth previous token “blue”
    • A randomly sampled vocabulary item
  8. At decoder position t, which tokens must a causal attention mask block?
    • The source sequence
    • All earlier target tokens
    • The current position’s own input
    • Target tokens at positions later than t