Parameter Quantization

Parameter quantization aims to reduce the memory footprint and computational cost of large neural networks by representing their parameters with fewer bits, thereby improving efficiency without significantly sacrificing accuracy. Current research focuses on developing novel quantization techniques for various architectures, including large language models (LLMs) and convolutional neural networks (CNNs), often employing adaptive strategies like column-level quantization or information retention methods to mitigate accuracy loss, particularly at very low bit-widths (e.g., 2-bits). This area is crucial for deploying large models on resource-constrained devices and accelerating training in federated learning settings, impacting both the scalability of AI and its accessibility across diverse hardware platforms.

Papers

October 1, 2024

STanH : Parametric Quantization for Variable Rate Learned Image Compression
Alberto Presta, Enzo Tartaglione, Attilio Fiandrotti, Marco Grangetto
Image Compression Encoder Decoder Quantization Learning Variable Rate Parameter Quantization

May 27, 2024

CLAQ: Pushing the Limits of Low-Bit Post-Training Quantization for LLMs
Haoyu Wang, Bei Liu, Hang Shao, Bo Xiao, Ke Zeng, Guanglu Wan, Yanmin Qian
Large Language Model Continuum Limit Quantization Operator Low Bit LLM Quantization Weight Only Quantization Class Centroid Parameter Quantization

February 8, 2024

Accurate LoRA-Finetuning Quantization of LLMs via Information Retention
Haotong Qin, Xudong Ma, Xingyu Zheng, Xiaoyang Li, Yang Zhang, Shouda Liu, Jie Luo, Xianglong Liu, Michele Magno
Medical LLM Integer Quantization Small LLM Information Retention Parameter Quantization

July 25, 2023

January 20, 2023

Optimized learned entropy coding parameters for practical neural-based image and video compression
Amir Said, Reza Pourreza, Hoang Le
Many Parameter Video Compression Efficient Optimization Entropy Coding Entropy Solution Uniform Quantization Based Image Quantization Challenge Parameter Quantization

January 15, 2023

RedBit: An End-to-End Flexible Framework for Evaluating the Accuracy of Quantized CNNs
André Santos, João Dinis Ferreira, Onur Mutlu, Gabriel Falcao
Convolutional Neural Network ImageNet Dataset Different Quantization State of the Art Quantization Parameter Quantization

May 6, 2022

Online Model Compression for Federated Learning with Large Models
Tien-Ju Yang, Yonghui Xiao, Giovanni Motta, Françoise Beaufays, Rajiv Mathews, Mingqing Chen
Speech Recognition Model Compression Large Model Large Deep Learning Model Compression Method Parameter Quantization