ReForm: Static and Dynamic Resource-Aware DNN Reconfiguration Framework for Mobile Device
ReForm: Static and Dynamic Resource-Aware DNN Reconfiguration Framework for Mobile Device
复制标题
DOI：
10.1145/3316781.3324696
复制
发表时间：
2019-06
期刊：
影响因子：
--
通讯作者：
Zirui Xu;Fuxun Yu;Chenchen Liu;Xiang Chen
中科院分区：
文献类型：
--
作者：
Zirui Xu;Fuxun Yu;Chenchen Liu;Xiang Chen
尽管深度神经网络（DNN）技术已广泛应用于各种应用中，但基于DNN的应用对于资源受限的移动设备来说仍然计算量过大。已经提出了许多工作来优化DNN的计算性能，但其中大多数局限于算法视角，忽略了实际部署中的某些计算问题。为了在实践中实现全面的DNN性能提升，预期的DNN优化工作应与特定的硬件和系统约束（即计算能力、能耗、内存占用和推理延迟）紧密配合。因此，在这项工作中，我们提出了ReForm——一个具有资源感知能力的DNN优化框架。通过全面的移动DNN计算分析和创新的模型重构方案（即基于交替方向乘子法的静态模型微调、动态选择性计算），ReForm能够针对各种静态和动态计算资源约束，高效且有效地重构一个预训练的DNN模型以用于实际的移动部署。实验表明，ReForm的优化速度比最先进的资源感知优化方法快约3.5倍。此外，ReForm能够针对具有不同资源约束的不同移动设备有效地重构DNN模型。而且，在静态和动态计算场景中，ReForm在可忽略的精度下降情况下实现了令人满意的计算成本降低（最多减少18%的工作量、16.23%的延迟、48.63%的内存和21.5%的能耗提升）。
Although the Deep Neural Network (DNN) technique has been widely applied in various applications, the DNN-based applications are still too computationally intensive for the resource-constrained mobile devices. Many works have been proposed to optimize the DNN computation performance, but most of them are limited in an algorithmic perspective, ignoring certain computing issues in practical deployment. To achieve the comprehensive DNN performance enhancement in practice, the expected DNN optimization works should closely cooperate with specific hardware and system constraints (i.e. computation capacity, energy cost, memory occupancy, and inference latency). Therefore, in this work, we propose ReForm-a resource-aware DNN optimization framework. Through thorough mobile DNN computing analysis and innovative model reconfiguration schemes (i.e. ADMM based static model fine-tuning, dynamically selective computing), ReForm can efficiently and effectively recon Figure a pre-trained DNN model for practical mobile deployment with regards to various static and dynamic computation resource constraints. Experiments show that ReForm has $\sim 3.5\times$ faster optimization speed than state-of-the-art resource-aware optimization method. Also, ReForm can effective recon Figure a DNN model to different mobile devices with distinct resource constraints. Moreover, ReForm achieves satisfying computation cost reduction with ignorable accuracy drop in both static and dynamic computing scenarios (at most 18% workload, 16.23% latency, 48.63% memory, and 21.5% energy enhancement).