paper-with-me

홈 › Papers

Rank and run-time aware compression of NLP Applications

2020-10-06 · EMNLP (sustainlp) 2020 11 · Urmish Thakker, Jesse Beu, Dibakar Gope, Ganesh Dasika, Matthew Mattina

Sequence model based NLP applications can be large. Yet, many applications that benefit from them run on small devices with very limited compute and storage capabilities, while still having run-time constraints. As a result, there is a need for a compression technique that can achieve significant compression without negatively impacting inference run-time and task accuracy. This paper proposes a new compression technique called Hybrid Matrix Factorization that achieves this dual objective. HMF improves low-rank matrix factorization (LMF) techniques by doubling the rank of the matrix using an intelligent hybrid-structure leading to better accuracy than LMF. Further, by preserving dense matrices, it leads to faster inference run-time than pruning or structure matrix based compression technique. We evaluate the impact of this technique on 5 NLP benchmarks across multiple tasks (Translation, Intent Detection, Language Modeling) and show that for similar accuracy values and compression factors, HMF can achieve more than 2.32x faster inference run-time than pruning and 16.77% better accuracy than LMF.

📄 PDF Abstract BibTeX arXiv:2010.03193

Code (0)

등록된 구현이 없습니다.

Tasks

Intent DetectionLanguage ModelingLanguage ModellingTranslation

Methods 이 논문이 사용한 방법론

Pruning 설명 없음

Similar Papers 제목 키워드 기반

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

2026-06-05 · Ernests Lavrinovics, Marco Letizia, Roy Janco, Shai Segal 외 arxiv

We present SigmaScale, a method for learning auxiliary scaling matrices $S$ to aid truncated Singular Value Decomposition (SVD) based Large Language Model (LLM) compression. Instead of deriving scaling matrices analytica…

HALOC: Hardware-Aware Automatic Low-Rank Compression for Compact Neural Networks

2023-01-20 · Jinqi Xiao, Chengming Zhang, Yu Gong, Miao Yin 외

Low-rank compression is an important model compression strategy for obtaining compact neural network models. In general, because the rank values directly determine the model complexity and model accuracy, proper selectio…

GPULow-rank compressionModel Compression

Low Rank Optimization for Efficient Deep Learning: Making A Balance between Compact Architecture and Fast Training

2023-03-22 · Xinwei Ou, Zhangxin Chen, Ce Zhu, Yipeng Liu

Deep neural networks have achieved great success in many data processing applications. However, the high computational complexity and storage cost makes deep learning hard to be used on resource-constrained devices, and …

Model CompressionQuantization

Beyond RAG: Task-Aware KV Cache Compression for Comprehensive Knowledge Reasoning

2025-03-06 · Giulio Corallo, Orion Weller, Fabio Petroni, Paolo Papotti

Incorporating external knowledge in large language models (LLMs) enhances their utility across diverse applications, but existing methods have trade-offs. Retrieval-Augmented Generation (RAG) fetches evidence via similar…

RAGRetrieval-augmented Generation

BALF: Budgeted Activation-Aware Low-Rank Factorization for Fine-Tuning-Free Model Compression

2025-09-29 · David González-Martínez arxiv

Activation-aware low-rank factorization techniques yield strong compression results but are generally confined to linear layers, while existing whitening-based theory typically makes an implicit full-rank assumption on a…

Model Compression