Salvato in:
Dettagli Bibliografici
Autori principali: Lee, Changwoo, Kwon, Soo Min, Qu, Qing, Kim, Hun-Seok
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:https://arxiv.org/abs/2410.21262
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916460665765888
author Lee, Changwoo
Kwon, Soo Min
Qu, Qing
Kim, Hun-Seok
author_facet Lee, Changwoo
Kwon, Soo Min
Qu, Qing
Kim, Hun-Seok
contents Large-scale foundation models have demonstrated exceptional performance in language and vision tasks. However, the numerous dense matrix-vector operations involved in these large networks pose significant computational challenges during inference. To address these challenges, we introduce the Block-Level Adaptive STructured (BLAST) matrix, designed to learn and leverage efficient structures prevalent in the weight matrices of linear layers within deep learning models. Compared to existing structured matrices, the BLAST matrix offers substantial flexibility, as it can represent various types of structures that are either learned from data or computed from pre-existing weight matrices. We demonstrate the efficiency of using the BLAST matrix for compressing both language and vision tasks, showing that (i) for medium-sized models such as ViT and GPT-2, training with BLAST weights boosts performance while reducing complexity by 70% and 40%, respectively; and (ii) for large foundation models such as Llama-7B and DiT-XL, the BLAST matrix achieves a 2x compression while exhibiting the lowest performance degradation among all tested structured matrices. Our code is available at https://github.com/changwoolee/BLAST.
format Preprint
id arxiv_https___arxiv_org_abs_2410_21262
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle BLAST: Block-Level Adaptive Structured Matrices for Efficient Deep Neural Network Inference
Lee, Changwoo
Kwon, Soo Min
Qu, Qing
Kim, Hun-Seok
Machine Learning
Artificial Intelligence
Large-scale foundation models have demonstrated exceptional performance in language and vision tasks. However, the numerous dense matrix-vector operations involved in these large networks pose significant computational challenges during inference. To address these challenges, we introduce the Block-Level Adaptive STructured (BLAST) matrix, designed to learn and leverage efficient structures prevalent in the weight matrices of linear layers within deep learning models. Compared to existing structured matrices, the BLAST matrix offers substantial flexibility, as it can represent various types of structures that are either learned from data or computed from pre-existing weight matrices. We demonstrate the efficiency of using the BLAST matrix for compressing both language and vision tasks, showing that (i) for medium-sized models such as ViT and GPT-2, training with BLAST weights boosts performance while reducing complexity by 70% and 40%, respectively; and (ii) for large foundation models such as Llama-7B and DiT-XL, the BLAST matrix achieves a 2x compression while exhibiting the lowest performance degradation among all tested structured matrices. Our code is available at https://github.com/changwoolee/BLAST.
title BLAST: Block-Level Adaptive Structured Matrices for Efficient Deep Neural Network Inference
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2410.21262