Sparse Transformer

GPTKB entity

Statements (23)
Predicate Object
gptkbp:instanceOf gptkb:convolutional_neural_network
gptkbp:appliesTo image generation
language modeling
gptkbp:attentionPattern fixed sparse attention
local attention
strided attention
gptkbp:author gptkb:Ilya_Sutskever
gptkb:Scott_Gray
gptkb:Alec_Radford
gptkb:Rewon_Child
gptkbp:designedFor efficient sequence modeling
gptkbp:developedBy gptkb:OpenAI
gptkbp:enables scaling to longer sequences
https://www.w3.org/2000/01/rdf-schema#label Sparse Transformer
gptkbp:improves Transformer efficiency
gptkbp:introducedIn 2019
gptkbp:publishedIn gptkb:arXiv:1904.10509
gptkbp:reduces computational complexity
gptkbp:relatedTo gptkb:transformation
Attention mechanism
gptkbp:uses sparse attention mechanism
gptkbp:bfsParent gptkb:transformation
gptkbp:bfsLayer 5