UniCoS: A Unified Neural and Accelerator Co-Search Framework for CNNs and ViTs

Current algorithm-hardware co-search works often suffer from lengthy training times and inadequate exploration of hardware design spaces, leading to suboptimal performance. This work introduces UniCoS, a unified framework for co-optimizing neural networks and accelerators for CNNs and Vision Transfo...

Celý popis

Uloženo v:

Podrobná bibliografie
Vydáno v:	2025 62nd ACM/IEEE Design Automation Conference (DAC) s. 1 - 6
Hlavní autoři:	Fu, Wei, Lou, Wenqi, Tang, Cheng, Wen, Hongbing, Qin, Yunji, Gong, Lei, Wang, Chao, Zhou, Xuehai
Médium:	Konferenční příspěvek
Jazyk:	angličtina
Vydáno:	IEEE 22.06.2025
Témata:	Accuracy Computer architecture Design automation Hardware acceleration hardware software co-exploration heterogeneous multi-core architecture Navigation Neural networks Software Space exploration Training Transformers zero-shot proxy
On-line přístup:	Získat plný text
Tagy:	Přidat tag Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!

Popis
Shrnutí:	Current algorithm-hardware co-search works often suffer from lengthy training times and inadequate exploration of hardware design spaces, leading to suboptimal performance. This work introduces UniCoS, a unified framework for co-optimizing neural networks and accelerators for CNNs and Vision Transformers (ViTs). By introducing a novel training-free proxy that evaluates accuracy within seconds and a clustering-based algorithm for exploring heterogeneous dataflows, UniCoS efficiently navigates the design spaces of both architectures. Experimental results demonstrate that the solutions generated by UniCoS consistently surpass state-of-the-art (SOTA) methods (e.g., 3.54 \times energy-delay product (EDP) improvement with a 1.76 \% higher accuracy on ImageNet) while requiring notably reduced search time (up to 48 \times, \sim 3 hours). The code is available at https://github.com/mine7777/Unicos.git.
DOI:	10.1109/DAC63849.2025.11133418