low-rank-decomposition

#low-rank-decomposition

SigmaScale: LLM Compression with SVD-based Low-Rank Decomposition and Learned Scaling Matrices

arXiv cs.CL ↗ · 6d ago Cached

Introduces SigmaScale, a method that learns auxiliary scaling matrices for SVD-based LLM compression, showing competitive performance on Llama 3.1 8B and Qwen3-8B benchmarks.

0 favorites 0 likes

low-rank-decomposition

SigmaScale: LLM Compression with SVD-based Low-Rank Decomposition and Learned Scaling Matrices

Submit Feedback