AI 中文总结
研究针对PIM架构上CNN微调计算和内存密集问题，提出硬件感知框架ADEPT。该框架考虑训练开销和层敏感度，引入新度量量化权衡，产生依赖平台的微调配置，能减少可训练参数和片外数据访问，微调时预测准确性损失最小。
AI 中文摘要
基于内存处理（PIM）的架构已成为在边缘加速卷积神经网络（CNN）工作负载的一个有前景的解决方案。微调预训练的CNN是部署后提高模型预测准确性的常见需求。然而，微调过程计算和内存密集，会产生大量中间激活，导致频繁的片外内存访问，影响PIM加速器的整体效率。现有微调策略对底层硬件不敏感，平等对待所有层。本文提出一种名为ADEPT的硬件感知框架来加速PIM架构上的CNN微调。与先前方法不同，ADEPT在考虑训练开销和层敏感度的情况下自适应训练模型。具体而言，ADEPT引入一种新颖的度量，量化块基于梯度的敏感度与其特定硬件架构的能量延迟积（EDP）之间的权衡，产生依赖平台的微调配置。总体而言，ADEPT有助于减少微调期间的可训练参数总数和片外数据访问，同时与全参数微调相比，预测准确性损失最小。
英文摘要
Processing-in-memory-based (PIM) architectures have emerged as a promising solution for accelerating Convolutional Neural Network (CNN) workloads at the edge. Fine-tuning pre-trained CNNs is a common requirement to enhance the model predictive accuracy after deployment. However, the fine-tuning process is computational and memory-intensive, generating a significant amount of intermediate activations. This leads to frequent off-chip memory access, affecting the overall efficiency of the PIM accelerator. Existing fine-tuning strategies are agnostic to the underlying hardware, as they treat all layers equally. In this paper, we propose a hardware-aware framework called ADEPT to accelerate CNN fine-tuning on PIM architectures. Unlike prior fine-tuning methods, ADEPT adaptively trains the model considering both the training overhead and layer sensitivity. Specifically, ADEPT introduces a novel metric that quantifies the trade-off between a block's gradient-based sensitivity and its hardware architecture-specific Energy-Delay Product (EDP), producing platform-dependent fine-tuning configurations. Overall, ADEPT helps reduce the total trainable parameters and the off-chip data access during fine-tuning, while incurring minimal loss in predictive accuracy compared to full-parameter fine-tuning.
CommentsAccepted for Publication in IEEE/ACM Embedded Systems Week (ESWEEK-26)
