arXiv cs.AI· Reem Abdel-Salam, Ahmed H. Abdel-Gawad, Amr G. Wassal·· 3 天前AI 评分22
VLCQ：用可变长度编码实现深度神经网络训练后量化
VLCQ: Post-training quantization for deep neural networks using variable length coding
AI 导读
VLCQ 是一种放宽定点编码约束的训练后量化方法，通过可变长度编码在降低位宽的同时无需训练数据或模型微调。在图像分类、分割和检测任务上，其逐通道量化在多数模型上超越 FP32 精度和 PWLQ，相比 FP32 最高实现 6x 压缩、相比 PWLQ 最高 1.7x；以压缩优先时相比 FP32 最高达 12.25x 压缩，性能损失低于 4%。
正文 · 原文
Abstract:Quantization plays a crucial role in efficiently deploying deep learning models on resource-constrained devices. Post-training quantization does not require access to the original dataset or retraining the full model. Current methods that achieve high performance close to baseline results typically require INT8 fixed-point integers. However, achieving higher model compression with lower bit-widths often leads to significant performance degradation. In this paper, we propose VLCQ, a post-training quantization method that relaxes the fixed-point encoding constraint, enabling more flexible exploration of quantization techniques through variable-length encoding. This approach achieves competitive or improved performance while reducing bit-widths, without requiring access to training data or model fine-tuning. Extensive experiments on image classification, segmentation, and object detection demonstrate that VLCQ improves model compression and accuracy compared with state-of-the-art post-training quantization approaches. For per-channel quantization, VLCQ surpasses FP32 accuracy and Piece-Wise Linear Quantization (PWLQ) on most evaluated models, achieving up to 6x compression compared with FP32 and up to 1.7x compared with PWLQ. When prioritizing compression with limited accuracy loss, VLCQ achieves up to 12.25x compression compared with FP32 with less than 4% performance degradation. For per-tensor quantization, VLCQ is competitive with Data-Free Quantization (DFQ), while providing greater flexibility to achieve lower bit rates across different tasks and models.
| Comments: | Published in Future Generation Computer Systems, Volume 166 (2025), Article 107654 |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Performance (cs.PF) |
| Cite as: | arXiv:2610.15092 [cs.CV] |
| (or arXiv:2610.15092v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.15092 arXiv-issued DOI via DataCite (pending registration) |
|
| Journal reference: | Reem Abdel-Salam, Ahmed H. Abdel-Gawad, and Amr G. Wassal. VLCQ: Post-training quantization for deep neural networks using variable length coding. Future Generation Computer Systems, 166, 107654, 2025 |
| Related DOI: | https://doi.org/10.1016/j.future.2024.107654
DOI(s) linking to related resources |
Submission history
From: Reem Abdel-Salam [view email]
[v1]
Sat, 10 Oct 2026 18:02:45 UTC (2,866 KB)
来源：arXiv cs.AI · arxiv.org