paper-with-me

홈 › Papers

Two Sparse Matrices are Better than One: Sparsifying Neural Networks with Double Sparse Factorization

2024-09-27 · Vladimír Boža, Vladimír Macko

Neural networks are often challenging to work with due to their large size and complexity. To address this, various methods aim to reduce model size by sparsifying or decomposing weight matrices, such as magnitude pruning and low-rank or block-diagonal factorization. In this work, we present Double Sparse Factorization (DSF), where we factorize each weight matrix into two sparse matrices. Although solving this problem exactly is computationally infeasible, we propose an efficient heuristic based on alternating minimization via ADMM that achieves state-of-the-art results, enabling unprecedented sparsification of neural networks. For instance, in a one-shot pruning setting, our method can reduce the size of the LLaMA2-13B model by 50% while maintaining better performance than the dense LLaMA2-7B model. We also compare favorably with Optimal Brain Compression, the state-of-the-art layer-wise pruning approach for convolutional neural networks. Furthermore, accuracy improvements of our method persist even after further model fine-tuning. Code available at: https://github.com/usamec/double_sparse.

📄 PDF Abstract BibTeX arXiv:2409.18850

Code (1)

usamec/double_sparse 공식 구현 pytorch

Methods 이 논문이 사용한 방법론

Pruning 설명 없음
ADMM The alternating direction method of multipliers (ADMM) is an algorithm that solves convex optimization problems by breaking them into smaller pieces, each of which are…

Similar Papers 제목 키워드 기반

Pixelated Butterfly: Simple and Efficient Sparse training for Neural Network Models

2021-11-30 · ICLR 2022 4 · Tri Dao, Beidi Chen, Kaizhao Liang, Jiaming Yang 외

Overparameterized neural networks generalize well but are expensive to train. Ideally, one would like to reduce their computational cost while retaining their generalization benefits. Sparse model training is a simple an…

Language ModelingLanguage Modelling

Learning Explicitly Conditioned Sparsifying Transforms

2024-03-05 · Andrei Pătraşcu, Cristian Rusu, Paul Irofti

Sparsifying transforms became in the last decades widely known tools for finding structured sparse representations of signals in certain transform domains. Despite the popularity of classical transforms such as DCT and W…

RIP sensing matrices construction for sparsifying dictionaries with application to MRI imaging

2024-07-30 · Jinn Ho, Wen-Liang Hwang, Andreas Heinecke

Practical applications of compressed sensing often restrict the choice of its two main ingredients. They may (i) prescribe using particular redundant dictionaries for certain classes of signals to become sparsely represe…

compressed sensing

RadiX-Net: Structured Sparse Matrices for Deep Neural Networks

2019-04-30 · Ryan A. Robinett, Jeremy Kepner

The sizes of deep neural networks (DNNs) are rapidly outgrowing the capacity of hardware to store and train them. Research over the past few decades has explored the prospect of sparsifying DNNs before, during, and after…

SEA: Sparse Linear Attention with Estimated Attention Mask

2023-10-03 · Heejun Lee, Jina Kim, Jeffrey Willette, Sung Ju Hwang

The transformer architecture has driven breakthroughs in recent years on tasks which require modeling pairwise relationships between sequential elements, as is the case in natural language understanding. However, long se…

Knowledge DistillationLanguage ModelingLanguage ModellingNatural Language Understanding