HACScale : hardware-aware compound scaling for resource-efficient DNNs

• المؤلفون الرئيسيون: Kong, Hao, Liu, Di, Luo, Xiangzhong, Liu, Weichen, Subramaniam, Ravi
• مؤلفون آخرون: School of Computer Science and Engineering
• التنسيق: Conference or Workshop Item
• اللغة: English
• منشور في: 2022
• الموضوعات: Engineering::Computer science and engineering; Deep Learning; Design Automation
• الوصول للمادة أونلاين: https://hdl.handle.net/10356/155808

Model scaling is an effective way to improve the accuracy of deep neural networks (DNNs) by increasing the model capacity. However, existing approaches seldom consider the underlying hardware, causing inefficient utilization of hardware resources and consequently high inference latency. In this pape...