24689

Neural Network Libraries: A Deep Learning Framework Designed from Engineers’ Perspectives

Akio Hayakawa, Masato Ishii, Yoshiyuki Kobayashi, Akira Nakamura, Takuya Narihira, Yukio Obuchi, Andrew Shin, Takuya Yashima, Kazuki Yoshiyama
Sony Corporation
arXiv:2102.06725 [cs.LG], (12 Feb 2021)

@misc{hayakawa2021neural,

   title={Neural Network Libraries: A Deep Learning Framework Designed from Engineers’ Perspectives},

   author={Akio Hayakawa and Masato Ishii and Yoshiyuki Kobayashi and Akira Nakamura and Takuya Narihira and Yukio Obuchi and Andrew Shin and Takuya Yashima and Kazuki Yoshiyama},

   year={2021},

   eprint={2102.06725},

   archivePrefix={arXiv},

   primaryClass={cs.LG}

}

While there exist a plethora of deep learning tools and frameworks, the fast-growing complexity of the field brings new demands and challenges, such as more flexible network design, speedy computation on distributed setting, and compatibility between different tools. In this paper, we introduce Neural Network Libraries, a deep learning framework designed from engineer’s perspective, with emphasis on usability and compatibility as its core design principles. We elaborate on each of our design principles and its merits, and validate our attempts via experiments.
No votes yet.
Please wait...

* * *

* * *

HGPU group © 2010-2021 hgpu.org

All rights belong to the respective authors

Contact us: