arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2408.05752cs.CV
收藏

RTF-Q:基于免重训练量化的高效无监督域自适应

RTF-Q: Efficient Unsupervised Domain Adaptation with Retraining-free Quantization

Shenzhen International Graduate School Tsinghua University(清华大学深圳国际研究生院)

Nanyang Du, Chen Tang, Yuxiao Jiang, Yuan Meng, Zhi Wang

AI总结:

针对边缘设备无监督域自适应训练成本高、存在精度冗余的问题,提出RTF-Q免重训练量化方法,通过权重共享、多比特宽联合训练等实现多架构适配,在三个基准上精度达SOTA水平且大幅降低内存与计算成本。

摘要:

在资源受限的边缘设备上执行无监督域自适应极具挑战性。现有研究通常采用架构优化(例如设计可瘦身网络),但需要高昂的训练成本,且未考虑参数和激活存在的大量精度冗余。为解决这些局限,我们提出基于免重训练量化(ReTraining-Free Quantization,RTF-Q)的高效无监督域自适应方法。该方法采用计算成本各异的低精度量化架构,可适配具有动态计算预算的设备。我们巧妙配置子网维度,利用权重共享在单组权重内优化多种架构,从而可直接使用开源仓库中的预训练模型。此外,我们引入了多比特宽联合训练与SandwichQ规则,二者均能有效处理子网间的多种量化比特宽。实验结果表明,我们的网络在三个基准上取得了与当前最优方法相当的精度,同时显著降低了内存与计算成本。

英文摘要:

Performing unsupervised domain adaptation on resource-constrained edge devices is challenging. Existing research typically adopts architecture optimization (e.g., designing slimmable networks) but requires expensive training costs. Moreover, it does not consider the considerable precision redundancy of parameters and activations. To address these limitations, we propose efficient unsupervised domain adaptation with ReTraining-Free Quantization (RTF-Q). Our approach uses low-precision quantization architectures with varying computational costs, adapting to devices with dynamic computation budgets. We subtly configure subnet dimensions and leverage weight-sharing to optimize multiple architectures within a single set of weights, enabling the use of pre-trained models from open-source repositories. Additionally, we introduce multi-bitwidth joint training and the SandwichQ rule, both of which are effective in handling multiple quantization bit-widths across subnets. Experimental results demonstrate that our network achieves competitive accuracy with state-of-the-art methods across three benchmarks while significantly reducing memory and computational costs.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑