paper-with-me

홈 › Papers

Outlier-Aware Training for Low-Bit Quantization of Structural Re-Parameterized Networks

2024-02-11 · Muqun Niu, Yuan Ren, Boyu Li, Chenchen Ding

Lightweight design of Convolutional Neural Networks (CNNs) requires co-design efforts in the model architectures and compression techniques. As a novel design paradigm that separates training and inference, a structural re-parameterized (SR) network such as the representative RepVGG revitalizes the simple VGG-like network with a high accuracy comparable to advanced and often more complicated networks. However, the merging process in SR networks introduces outliers into weights, making their distribution distinct from conventional networks and thus heightening difficulties in quantization. To address this, we propose an operator-level improvement for training called Outlier Aware Batch Normalization (OABN). Additionally, to meet the demands of limited bitwidths while upkeeping the inference accuracy, we develop a clustering-based non-uniform quantization framework for Quantization-Aware Training (QAT) named ClusterQAT. Integrating OABN with ClusterQAT, the quantized performance of RepVGG is largely enhanced, particularly when the bitwidth falls below 8.

📄 PDF Abstract BibTeX arXiv:2402.07200

Code (0)

등록된 구현이 없습니다.

Tasks

Quantization

Methods 이 논문이 사용한 방법론

Average Pooling 설명 없음
Batch Normalization 설명 없음
Convolution A convolution is a type of matrix operation, consisting of a kernel, a small matrix of weights, that slides over input data performing element-wise multiplication with the…
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
ReLU How Do I Communicate to Expedia? How Do I Communicate to Expedia? – Call ☎️ +1-(888) 829 (0881) or +1-805-330-4056 or +1-805-330-4056 for Live Support & Special Travel…
Global Average Pooling Global Average Pooling is a pooling operation designed to replace fully connected layers in classical CNNs. The idea is to generate one feature map for each corresponding…
Residual Connection 설명 없음
RepVGG RepVGG is a VGG-style convolutional architecture. It has the following advantages: - The model has a VGG-like plain (a.k.a.…

Similar Papers 제목 키워드 기반

Towards Accurate Post-training Quantization for Reparameterized Models

2024-02-25 · Luoming Zhang, Yefei He, Wen Fei, Zhenyu Lou 외

Model reparameterization is a widely accepted technique for improving inference speed without compromising performance. However, current Post-training Quantization (PTQ) methods often lead to significant accuracy degrada…

Quantization

RoSTE: An Efficient Quantization-Aware Supervised Fine-Tuning Approach for Large Language Models

2025-02-13 · Quan Wei, Chung-Yiu Yau, Hoi-To Wai, Yang 외

Supervised fine-tuning is a standard method for adapting pre-trained large language models (LLMs) to downstream tasks. Quantization has been recently studied as a post-training technique for efficient LLM deployment. To …

Quantization

Colinearity Decay: Training Quantization-Friendly ViTs with Outlier Decay

2026-05-02 · Jin Tong, Guang Liang, Peilin Sun, Jianxin Wu arxiv

Low-bit quantization is a practical route for efficiently deploying vision Transformers, yet activation outliers complicate fully quantized deployment. Existing methods either handle quantization post-training or suppres…

Outlier-Aware Post-Training Quantization for Image Super-Resolution

2025-11-01 · Hailing Wang, jianglin Lu, Yitian Zhang, Yun Fu arxiv

Quantization techniques, including quantization-aware training (QAT) and post-training quantization (PTQ), have become essential for inference acceleration of image super-resolution (SR) networks. Compared to QAT, PTQ ha…

Image Super-Resolution

Astro: Activation-guided Structured Regularization for Outlier-Robust LLM Post-Training Quantization

2026-02-07 · Xi Chen, Ming Li, Junxi Li, Changsheng Li 외 arxiv

Weight-only post-training quantization (PTQ) is crucial for efficient Large Language Model (LLM) deployment but suffers from accuracy degradation caused by weight and activation outliers. Existing mitigation strategies o…