TSB:用于在 NVCIM 加速器上高效部署 DNN 的小型共享块
TSB: Tiny Shared Block for Efficient DNN Deployment on NVCIM Accelerators
AI总结:
本文提出 TSB 方法,通过在 DNN 架构中集成小型共享 1x1 卷积块来稳定特征处理,有效缓解 NVCIM 加速器的器件变化影响,显著提升推理精度并降低训练与映射开销。
摘要:
使用非易失性存储器(NVM)器件的存内计算(CIM)加速器为低能耗和低延迟的深度神经网络(DNN)推理执行提供了极具前景的解决方案。然而,在实际部署中,处理受非易失性存内计算(NVCIM)加速器内部器件固有的变化影响的大量模型权重参数这一挑战,往往成为其障碍。该问题增加了训练开销以及将权重映射到器件状态所需的时间和能量,并降低了推理精度,从而显著抵消了其优势。为了缓解这些挑战,我们提出了“Tiny Shared Block (TSB)”方法,该方法将一个小型共享的 1x1 卷积块集成到 DNN 架构中。该块旨在稳定整个网络的特征处理,有效降低器件变化的影响。大量实验结果表明,与最先进的基线解决方案相比,TSB 实现了超过 20 倍的推理精度差距改善,超过 5 倍的训练加速,降低了权重到器件的映射成本,并且在编程期间仅需对不到 0.4% 的原始权重进行写入验证。我们的方法为在 NVCIM 加速器上部署鲁棒的 DNN 模型提供了实用且高效的解决方案,是对节能 AI 硬件领域的一项宝贵贡献。
英文摘要:
Compute-in-memory (CIM) accelerators using non-volatile memory (NVM) devices offer promising solutions for energy-efficient and low-latency Deep Neural Network (DNN) inference execution. However, practical deployment is often hindered by the challenge of dealing with the massive amount of model weight parameters impacted by the inherent device variations within non-volatile computing-in-memory (NVCIM) accelerators. This issue significantly offsets their advantages by increasing training overhead, the time and energy needed for mapping weights to device states, and diminishing inference accuracy. To mitigate these challenges, we propose the "Tiny Shared Block (TSB)" method, which integrates a small shared 1x1 convolution block into the DNN architecture. This block is designed to stabilize feature processing across the network, effectively reducing the impact of device variation. Extensive experimental results show that TSB achieves over 20x inference accuracy gap improvement, over 5x training speedup, and weights-to-device mapping cost reduction while requiring less than 0.4% of the original weights to be write-verified during programming, when compared with state-of-the-art baseline solutions. Our approach provides a practical and efficient solution for deploying robust DNN models on NVCIM accelerators, making it a valuable contribution to the field of energy-efficient AI hardware.
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。