paper-with-me

홈 › Papers

The Uniqueness of LLaMA3-70B Series with Per-Channel Quantization

2024-08-27 · Minghai Qin

We have observed a distinctive quantization-related behavior in the LLaMA3/3.1-70B models that is absent in both the LLaMA2-70B and LLaMA3/3.1/3.2-1B/3B/8B/405B models. Quantization is a crucial technique for deploying large language models (LLMs) efficiently. The impact of W8A8 post-training quantization on model accuracy, especially on the recently released LLaMA3/3.1 model series, remains contentious. In this paper, we explore three key questions: What makes the LLaMA3-70B model series uniquely vulnerable to quantization? Why is this the case? And how can the issue be addressed? We empirically investigate multiple LLMs featured on an open LLM leaderboard, discovering that the LLaMA3-70B model series have a unique accuracy degradation behavior with W8A8 per-channel post-training quantization. In contrast, other model series such as LLaMA2, LLaMA3/3.1-8B, LLaMA3.2, Qwen, Mixtral, Mistral, Phi-3, and Falcon demonstrate robust performance with W8A8. Contrary to previous assertions attributing degradation to the large dynamic range of activations, our findings indicate that the weight distribution of the LLaMA3-70B is the primary factor behind the vulnerability. By meticulously analyzing the distinct characteristics of weight distributions across Transformer blocks, we propose two solutions that make different tradeoffs in hardware/software overhead. First, we propose a mixed strategy where less than 3\% of the layers employ finer per-group W8A8 quantization granularity. Second, we introduce a bi-smoothing strategy that balances quantization errors between weights and activations while maintaining per-channel quantization throughout. Experimental results demonstrate that both strategies effectively preserve the accuracy of the entire LLaMA3-70B model series under W8A8 quantization, achieving performance on par with their FP16 counterparts.

📄 PDF Abstract BibTeX arXiv:2408.15301

Code (0)

등록된 구현이 없습니다.

Tasks

Quantization

Methods 이 논문이 사용한 방법론

Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Adam 설명 없음
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
Attention 설명 없음
Position-Wise Feed-Forward Layer 설명 없음
Dense Connections Dense Connections, or Fully Connected Connections, are a type of layer in a deep neural network that use a linear operation where every input is connected to every output…
Residual Connection 설명 없음
Multi-Head Attention 설명 없음

Similar Papers 제목 키워드 기반

QLLM: Accurate and Efficient Low-Bitwidth Quantization for Large Language Models

2023-10-12 · Jing Liu, Ruihao Gong, Xiuying Wei, Zhiwei Dong 외

Large Language Models (LLMs) excel in NLP, but their demands hinder their widespread deployment. While Quantization-Aware Training (QAT) offers a solution, its extensive training costs make Post-Training Quantization (PT…

GPUQuantization

KVQuant: Towards 10 Million Context Length LLM Inference with KV Cache Quantization

2024-01-31 · Coleman Hooper, Sehoon Kim, Hiva Mohammadzadeh, Michael W. Mahoney 외

LLMs are seeing growing use for applications which require large context windows, and with these large context windows KV cache activations surface as the dominant contributor to memory consumption during inference. Quan…

GPUQuantization

VPTQ: Extreme Low-bit Vector Post-Training Quantization for Large Language Models

2024-09-25 · Yifei Liu, Jicheng Wen, Yang Wang, Shengyu Ye 외

Scaling model size significantly challenges the deployment and inference of Large Language Models (LLMs). Due to the redundancy in LLM weights, recent research has focused on pushing weight-only quantization to extremely…

Quantization

Outlier Suppression+: Accurate quantization of large language models by equivalent and optimal shifting and scaling

2023-04-18 · Xiuying Wei, Yunchen Zhang, Yuhang Li, Xiangguo Zhang 외

Post-training quantization~(PTQ) of transformer language models faces significant challenges due to the existence of detrimental outliers in activations. We observe that these outliers are concentrated in specific channe…

Quantization

Trainable Smooth-Rotation Transforms with Learned Channel Scales for LLM Quantization

2026-06-07 · Patrik Czakó, Gábor Kertész, Sándor Szénási arxiv

Post-training quantization (PTQ) is one of the most practical ways to reduce the serving cost of Large Language Models (LLMs), but activation quantization remains difficult because outlier-dominated channels lead to larg…