"FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness."

Tri Dao et al. (2022)

Details and statistics

DOI: 10.48550/ARXIV.2205.14135

access: open

type: Informal or Other Publication

metadata version: 2023-08-28

a service of  Schloss Dagstuhl - Leibniz Center for Informatics