Machine learning

[NEEDS TRANSLATION]

Dugotrajni transformeri poput Longformera (Beltagy, Peters & Cohan, 2020) i BigBirda (Zaheer et al., 2020) zamjenjuju standardnu O(n²) pažnju transformera rijetkim (sparse) obrascima pažnje koji linearno, O(n), skaliraju s duljinom sekvence. To omogućuje jednom modelu da obrati pažnju na tisuće tokena — cijele dokumente, pravne tekstove ili genomske sekvence — koji ne bi stali u konvencionalni transformer.

Otvorite u MethodMindUskoroVideoUskoroDownload slides

Pročitajte cijelu metodu

Samo za članove

Prijavite se besplatnim računom kako biste pročitali ovaj odjeljak.

Prijavite se

Method map

The neighbourhood of related methods — select a node to explore.

Izvori

  1. Beltagy, I., Peters, M. E. & Cohan, A. (2020). Longformer: The Long-Document Transformer. arXiv. link
  2. Zaheer, M. et al. (2020). Big Bird: Transformers for Longer Sequences. NeurIPS. link

Kako citirati ovu stranicu

ScholarGate. (2026, June 1). Long-Sequence Transformers with Sparse Attention (Longformer / BigBird). ScholarGate. https://scholargate.app/hr/deep-learning/longformer-bigbird

Which method?

Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.

Compare side by side

Citirana u

ScholarGateLongformer / BigBird (Long-Sequence Transformers with Sparse Attention (Longformer / BigBird)). Preuzeto 2026-06-15 s https://scholargate.app/hr/deep-learning/longformer-bigbird · Skup podataka: https://doi.org/10.5281/zenodo.20539026