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 2302.08469cs.LGcs.ET
收藏

使用基于存内计算的加速器进行大规模和多样化深度学习推理工作负载的硬件感知训练

Hardware-aware training for large-scale and diverse deep learning inference workloads using in-memory computing-based accelerators

IBM Research(IBM研究院)
Malte J. Rasch, Charles Mackin, Manuel Le Gallo, An Chen, Andrea Fasoli, Frederic Odermatt, Ning Li, S. R. Nandakumar, Pritish Narayanan, Hsinyu Tsai, Geoffrey … 展开作者

Malte J. Rasch, Charles Mackin, Manuel Le Gallo, An Chen, Andrea Fasoli, Frederic Odermatt, Ning Li, S. R. Nandakumar, Pritish Narayanan, Hsinyu Tsai, Geoffrey W. Burr, Abu Sebastian, Vijay Narayanan

摘要:

模拟存内计算(AIMC)--一种很有前途的深度学习工作负载节能加速方法--计算矩阵向量乘法(MVMs),但由于非理想性通常是非确定性或非线性的,因此只能近似计算。与常规浮点(FP)实现相比,这可能不利地影响可实现的深度神经网络(DNN)推理准确性。虽然之前已经建议重新训练以提高鲁棒性,但之前的工作仅探索了少数DNN拓扑,使用不同且过于简化的AIMC硬件模型。在这里,我们使用硬件感知(HWA)训练来系统地检查AIMC在多个DNN拓扑中的多个常见人工智能(AI)工作负载的准确性,并研究对广泛的非理想性的敏感性和鲁棒性。通过引入一个新的和高度现实的AIMC交叉模型,我们显着改善早期的再培训方法。我们发现,许多不同拓扑结构的大规模DNN,包括卷积神经网络(CNN),递归神经网络(RNN)和Transformers,实际上可以成功地重新训练,以显示AIMC的等精度。我们的研究结果进一步表明,向输入或输出添加噪声的AIMC非理想性,而不是权重,对DNN精度的影响最大,并且RNN对所有非理想性都特别鲁棒。

英文摘要:

Analog in-memory computing (AIMC) -- a promising approach for energy-efficient acceleration of deep learning workloads -- computes matrix-vector multiplications (MVMs) but only approximately, due to nonidealities that often are non-deterministic or nonlinear. This can adversely impact the achievable deep neural network (DNN) inference accuracy as compared to a conventional floating point (FP) implementation. While retraining has previously been suggested to improve robustness, prior work has explored only a few DNN topologies, using disparate and overly simplified AIMC hardware models. Here, we use hardware-aware (HWA) training to systematically examine the accuracy of AIMC for multiple common artificial intelligence (AI) workloads across multiple DNN topologies, and investigate sensitivity and robustness to a broad set of nonidealities. By introducing a new and highly realistic AIMC crossbar-model, we improve significantly on earlier retraining approaches. We show that many large-scale DNNs of various topologies, including convolutional neural networks (CNNs), recurrent neural networks (RNNs), and transformers, can in fact be successfully retrained to show iso-accuracy on AIMC. Our results further suggest that AIMC nonidealities that add noise to the inputs or outputs, not the weights, have the largest impact on DNN accuracy, and that RNNs are particularly robust to all nonidealities.

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。

↑