Transformers Scale to Long Sequences With Linear Complexity Via Nyström-Based Self-Attention Approximation Transformers Scale to Long Sequences With Linear Complexity Via Nyström-Based Self-Attention Approximation Transformers Scale to Long Sequences With Linear Complexity Via Nyström-Based Self-Attention Approximation Synced

Researchers from the University of Wisconsin-Madison, UC Berkeley, Google Brain and American Family Insurance propose Nyströmformer, an adaption of the Nystrom method that approximates standard sel...