paper-with-me

홈 › Papers

DANCE: Differentiable Accelerator/Network Co-Exploration

2020-09-14 · Kanghyun Choi, Deokki Hong, Hojae Yoon, Joonsang Yu, Youngsok Kim, Jinho Lee

To cope with the ever-increasing computational demand of the DNN execution, recent neural architecture search (NAS) algorithms consider hardware cost metrics into account, such as GPU latency. To further pursue a fast, efficient execution, DNN-specialized hardware accelerators are being designed for multiple purposes, which far-exceeds the efficiency of the GPUs. However, those hardware-related metrics have been proven to exhibit non-linear relationships with the network architectures. Therefore it became a chicken-and-egg problem to optimize the network against the accelerator, or to optimize the accelerator against the network. In such circumstances, this work presents DANCE, a differentiable approach towards the co-exploration of the hardware accelerator and network architecture design. At the heart of DANCE is a differentiable evaluator network. By modeling the hardware evaluation software with a neural network, the relation between the accelerator architecture and the hardware metrics becomes differentiable, allowing the search to be performed with backpropagation. Compared to the naive existing approaches, our method performs co-exploration in a significantly shorter time, while achieving superior accuracy and hardware cost metrics.

📄 PDF Abstract BibTeX arXiv:2009.06237

Code (0)

등록된 구현이 없습니다.

Tasks

GPUNeural Architecture Search

Methods 이 논문이 사용한 방법론

DANCE Domain Adaptive Neighborhood Clustering via Entropy Optimization (DANCE) is a self-supervised clustering method that harnesses the cluster structure of the target domain using…

Similar Papers 제목 키워드 기반

Enabling Hard Constraints in Differentiable Neural Network and Accelerator Co-Exploration

2023-01-23 · Deokki Hong, Kanghyun Choi, Hye Yoon Lee, Joonsang Yu 외

Co-exploration of an optimal neural architecture and its hardware accelerator is an approach of rising interest which addresses the computational cost problem, especially in low-profile systems. The large co-exploration …

Neural Architecture Search

ConCoDE: Hard-constrained Differentiable Co-Exploration Method for Neural Architectures and Hardware Accelerators

2021-09-29 · Deokki Hong, Kanghyun Choi, Hey Yoon Lee, Joonsang Yu 외

While DNNs achieve over-human performances in a number of areas, it is often accompanied by the skyrocketing computational costs. Co-exploration of an optimal neural architecture and its hardware accelerator is an appro…

Neural Architecture Search

DNA: Differentiable Network-Accelerator Co-Search

2020-10-28 · Yongan Zhang, Yonggan Fu, Weiwen Jiang, Chaojian Li 외

Powerful yet complex deep neural networks (DNNs) have fueled a booming demand for efficient DNN solutions to bring DNN-powered intelligence into numerous applications. Jointly optimizing the networks and their accelerato…

Triple-Search: Differentiable Joint-Search of Networks, Precision, and Accelerators

2021-01-01 · Yonggan Fu, Yongan Zhang, Haoran You, Yingyan Lin

The record-breaking performance and prohibitive complexity of deep neural networks (DNNs) have ignited a substantial need for customized DNN accelerators which have the potential to boost DNN acceleration efficiency by o…

DOSA: Differentiable Model-Based One-Loop Search for DNN Accelerators

2025-09-12 · Charles Hong, Qijing Huang, Grace Dinh, Mahesh Subedar 외 arxiv

In the hardware design space exploration process, it is critical to optimize both hardware parameters and algorithm-to-hardware mappings. Previous work has largely approached this simultaneous optimization problem by sep…