paper-with-me

Papers

MixLLM: LLM Quantization with Global Mixed-precision between Output-features and Highly-efficient System Design

2024-12-19 · Zhen Zheng, Xiaonan Song, Chuanjie Liu

Quantization has become one of the most effective methodologies to compress LLMs into smaller size. However, the existing quantization solutions still show limitations of either non-negligible accuracy drop or system inefficiency. In this paper, we make a comprehensive analysis of the general quantization principles on their effect to the triangle of accuracy, memory consumption and system efficiency. We propose MixLLM that explores the new optimization space of mixed-precision quantization between output features based on the insight that different output features matter differently in the model. MixLLM identifies the output features with high salience in the global view rather than within each single layer, effectively assigning the larger bit-width to output features that need it most to achieve good accuracy with low memory consumption. We present the sweet spot of quantization configuration of algorithm-system co-design that leads to high accuracy and system efficiency. To address the system challenge, we design the two-step dequantization to make use of the int8 Tensor Core easily and fast data type conversion to reduce dequantization overhead significantly, and present the software pipeline to overlap the memory access, dequantization and the MatMul to the best. Extensive experiments show that with only 10% more bits, the PPL increasement can be reduced from about 0.5 in SOTA to within 0.2 for Llama 3.1 70B, while on average MMLU-Pro improves by 0.93 over the SOTA of three popular models. In addition to its superior accuracy, MixLLM also achieves state-of-the-art system efficiency.

📄 PDF Abstract BibTeX arXiv:2412.14590

Code (0)

등록된 구현이 없습니다.

Tasks

MMLUQuantization

Methods 이 논문이 사용한 방법론

LLaMA LLaMA is a collection of foundation language models ranging from 7B to 65B parameters. It is based on the transformer architecture with various improvements that were…

Similar Papers 제목 키워드 기반

Patch-wise Mixed-Precision Quantization of Vision Transformer

2023-05-11 · Junrui Xiao, Zhikai Li, Lianwei Yang, Qingyi Gu

As emerging hardware begins to support mixed bit-width arithmetic computation, mixed-precision quantization is widely used to reduce the complexity of neural networks. However, Vision Transformers (ViTs) require complex …

QuantizationSensitivity

WINDQuant: Weight-Informed Neural Decision-Making for Global Mixed-Precision LLM Quantization

2026-05-26 · Phong Nam Huu Nguyen, Khoi M. Le, Cong-Duy T Nguyen, Anh Tuan Luu 외 arxiv

Quantization is an effective approach to reduce the memory footprint and inference cost of large language models (LLMs), yet maintaining performance in the ultra-low-bit regime remains challenging. Existing post-training…

Reinforcement Learning

GEMQ: Global Expert-Level Mixed-Precision Quantization for MoE LLMs

2026-05-21 · Jianing Deng, Song Wang, Dongwei Wang, Zijie Liu 외 arxiv

Mixture-of-Experts Large Language Models (MoE-LLMs) achieve strong performance but incur substantial memory overhead due to massive expert parameters. Mixed-precision quantization mitigates this cost by allocating expert…

AMED: Automatic Mixed-Precision Quantization for Edge Devices

2022-05-30 · Moshe Kimhi, Tal Rozen, Avi Mendelson, Chaim Baskin

Quantized neural networks are well known for reducing the latency, power consumption, and model size without significant harm to the performance. This makes them highly appropriate for systems with limited resources and …

Quantization

SDQ: Stochastic Differentiable Quantization with Mixed Precision

2022-06-09 · Xijie Huang, Zhiqiang Shen, Shichao Li, Zechun Liu 외

In order to deploy deep models in a computationally efficient manner, model quantization approaches have been frequently used. In addition, as new hardware that supports mixed bitwidth arithmetic operations, recent resea…

Knowledge DistillationNeural Architecture SearchQuantization