AI Dynamics

Global AI News Aggregator

About

R-Sparse: Rank-Aware Activation Sparsity for LLM Inference

R-Sparse: Rank-Aware Activation Sparsity for Efficient LLM Inference
Paper: https://
arxiv.org/pdf/2504.19449
v1

Code: https://
github.com/VITA-Group/R-S
parse

→ View original post on X — @jiqizhixin,