Maswali yanayounganishwa na mwongozo · Kati Kiwango

Maswali ya Miundo ya BERT na Kisimbaji

Kagua muktadha wa BERT, mafunzo ya awali yaliyofichwa, usimbaji, urekebishaji, uwasilishaji, vibadala bora, shabaha za sentensi inayofuata na upachikaji wa ingizo.

Njia za mwongozo zinazohusianaMiundo ya Bert na Encoder
Swali 1 ya 8

Which contextual access does bidirectional attention permit in the original BERT encoder input?

Hakiki kila swali8 maswali

Majibu na maelezo huonekana unapojibu maswali.

  1. Which contextual access does bidirectional attention permit in the original BERT encoder input?
    • Information from both left and right within the allowed input
    • Only the previous token in every layer
    • Only a second unrelated language
    • A guarantee of two correct answers to every question
  2. Which original BERT pretraining objective predicts selected token identities using surrounding context?
    • Only next-token prediction under a causal mask
    • Only human-written class labels for complete documents
    • Only a fixed image reconstruction objective
    • Muundo wa lugha uliofichwa
  3. Which transformer component provides the original BERT representation architecture?
    • Only the autoregressive decoder
    • The encoder
    • A recurrent-only sequence network
    • A convolution-only image decoder
  4. What sequence of training stages is illustrated by BERT pretraining followed by downstream fine-tuning?
    • Always train a completely new network from random parameters for every example
    • Apply the task head before creating any learned representation
    • Learn a corpus-based starting representation, then adapt it to a defined task
    • Guarantee that every target task uses no labeled data
  5. What output can the original BERT encoder provide to classification or span-prediction heads?
    • Only a long autoregressive continuation in every task
    • Mawimbi ghafi ya sauti
    • Pixel-level image output
    • Contextual text representations
  6. Which comparison accurately distinguishes the parameter-related designs of DistilBERT and ALBERT?
    • DistilBERT uses distillation; ALBERT uses factorized embeddings and cross-layer sharing
    • Both are guarantees of identical speed and accuracy on every task
    • Both replace all text input with image pixels
    • Neither retains any learned encoder representation
  7. Which label distinction is used in the original BERT next-sentence prediction task?
    • A positive versus negative product review
    • A token's verb tense versus noun number
    • An actual next sentence versus a randomly sampled alternative
    • A source sentence's font size versus text color
  8. Which three representations are added in the original BERT input embedding?
    • Only a single corpus-frequency number
    • Token, position, and segment embeddings
    • Decoder probabilities, image pixels, and audio pitch
    • Only one fixed vector shared by every input position