Dynamic Quantization Range Control for Analog-in-Memory Neural Networks Acceleration
العنوان: | Dynamic Quantization Range Control for Analog-in-Memory Neural Networks Acceleration |
---|---|
المؤلفون: | Nathan Laubeuf, Jonas Doevenspeck, Ioannis A. Papistas, Michele Caselli, Stefan Cosemans, Peter Vrancx, Debjyoti Bhattacharjee, Arindam Mallik, Peter Debacker, Diederik Verkest, Francky Catthoor, Rudy Lauwereins |
المصدر: | ACM Transactions on Design Automation of Electronic Systems. 27:1-21 |
بيانات النشر: | Association for Computing Machinery (ACM), 2022. |
سنة النشر: | 2022 |
مصطلحات موضوعية: | Technology, Science & Technology, Computer Science, quantization, Electrical and Electronic Engineering, Computer Science, Hardware & Architecture, Computer Science, Software Engineering, in-memory-computing, Computer Graphics and Computer-Aided Design, Neural networks, COMPUTING SRAM MACRO, Computer Science Applications |
الوصف: | Analog in Memory Computing (AiMC) based neural network acceleration is a promising solution to increase the energy efficiency of deep neural networks deployment. However, the quantization requirements of these analog systems are not compatible with state-of-the-art neural network quantization techniques. Indeed, while the quantization of the weights and activations is considered by modern deep neural network quantization techniques, AiMC accelerators also impose the quantization of each Matrix Vector Multiplication (MVM) result. In most demonstrated AiMC implementations, the quantization range of MVM results is considered a fixed parameter of the accelerator. This work demonstrates that dynamic control over this quantization range is possible but also desirable for analog neural networks acceleration. An AiMC compatible quantization flow coupled with a hardware aware quantization range driving technique is introduced to fully exploit these dynamic ranges. Using CIFAR-10 and ImageNet as benchmarks, the proposed solution results in networks that are both more accurate and more robust to the inherent vulnerability of analog circuits than fixed quantization range based approaches. |
تدمد: | 1557-7309 1084-4309 |
URL الوصول: | https://explore.openaire.eu/search/publication?articleId=doi_dedup___::3a50bed920fdc64780ebd87cc1a37351 https://doi.org/10.1145/3498328 |
حقوق: | OPEN |
رقم الأكسشن: | edsair.doi.dedup.....3a50bed920fdc64780ebd87cc1a37351 |
قاعدة البيانات: | OpenAIRE |
تدمد: | 15577309 10844309 |
---|