arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2603.26595cs.LGhep-ex
收藏

PQuantML:用于端到端硬件感知模型压缩的工具

PQuantML: A Tool for End-to-End Hardware-aware Model Compression

European Center for Nuclear Research (CERN)(欧洲核子研究中心);Purdue University(普渡大学);California Institute of Technology(加州理工学院);University of Padova(帕多瓦大学);Physikalisches Institut, Heidelberg University(海德堡大学物理研究所);SLAC National Accelerator Laboratory(SLAC国家加速器实验室);Institute of Physics Belgrade(贝尔格莱德物理研究所)
Roope Niemi, Anastasiia Petrovych, Arghya Ranjan Das, Enrico Lupi, Chang Sun, Dimitrios Danopoulos, Marlon Joshua Helbing, Mia Liu, Sebastian Dittmeier, Michael… 展开作者

Roope Niemi, Anastasiia Petrovych, Arghya Ranjan Das, Enrico Lupi, Chang Sun, Dimitrios Danopoulos, Marlon Joshua Helbing, Mia Liu, Sebastian Dittmeier, Michael Kagan, Vladimir Loncar, Maurizio Pierini

AI总结:

PQuantML通过提供统一接口实现剪枝与量化,支持端到端工作流中的模型压缩,通过多种剪枝方法和固定点量化实现参数和位宽显著减少,同时保持精度。

摘要:

PQuantML是一个新的开源、面向硬件的神经网络模型压缩库,专门针对端到端工作流。受部署高性能模型到有严格延迟约束环境的需求驱动,PQuantML通过提供统一接口简化压缩模型的训练,支持联合或单独应用剪枝和量化。该库实现了多种具有不同粒度的剪枝方法,以及支持高粒度量化的固定点量化。我们评估了PQuantML在代表性任务如喷注子结构分类(即喷注标记)上的性能,该任务与实时LHC数据处理相关。使用各种剪枝方法结合固定点量化,PQuantML实现了参数和位宽的显著减少,同时保持准确性。所得到的压缩效果进一步与现有工具如QKeras和HGQ进行了比较。

英文摘要:

PQuantML is a new open-source, hardware-aware neural network model compression library tailored to end-to-end workflows. Motivated by the need to deploy performant models to environments with strict latency constraints, PQuantML simplifies training of compressed models by providing a unified interface to apply pruning and quantization, either jointly or individually. The library implements multiple pruning methods with different granularities, as well as fixed-point quantization with support for High-Granularity Quantization. We evaluate PQuantML on representative tasks such as the jet substructure classification, so-called jet tagging, an on-edge problem related to real-time LHC data processing. Using various pruning methods with fixed-point quantization, PQuantML achieves substantial parameter and bit-width reductions while maintaining accuracy. The resulting compression is further compared against existing tools, such as QKeras and HGQ.

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。

↑