Design Considerations for Efficient Deep Neural Networks on Processing-in-Memory Accelerators
This work addresses efficiency challenges for AI hardware developers and researchers, but it appears incremental as it focuses on design considerations rather than introducing a new method.
The paper tackles the problem of optimizing deep neural networks for efficient and accurate operation on processing-in-memory accelerators, using experiments on state-of-the-art networks with the ImageNet dataset to highlight key design considerations.
This paper describes various design considerations for deep neural networks that enable them to operate efficiently and accurately on processing-in-memory accelerators. We highlight important properties of these accelerators and the resulting design considerations using experiments conducted on various state-of-the-art deep neural networks with the large-scale ImageNet dataset.