A general tensor-structured compression scheme for efficient LLMs

2 pointsposted 4 hours ago
by wslh
(alphaxiv.org)

No comments yet