from Hacker News

Griffin: Mixing Gated Linear Recurrences with Local Attention for Efficient LMs

by panabee on 3/15/24, 7:44 PM with 0 comments