DLSpec: A deep learning task exchange specification

Abdul Dakkak; Cheng Li; Jinjun Xiong; Wen Mei Hwu

DLSpec: A deep learning task exchange specification

Abdul Dakkak
, Cheng Li
, Jinjun Xiong
, Wen Mei Hwu

Department of Computer Science and Engineering

University of Illinois at Urbana-Champaign

Research output: Contribution to conference › Paper › peer-review

1 Scopus citations

Abstract

Deep Learning (DL) innovations are being introduced at a rapid pace. However, the current lack of standard specification of DL tasks makes sharing, running, reproducing, and comparing these innovations difficult. To address this problem, we propose DLSpec, a model-, dataset-, software-, and hardware-agnostic DL specification that captures the different aspects of DL tasks. DLSpec has been tested by specifying and running hundreds of DL tasks.

Original language	English
State	Published - 2020
Event	2020 USENIX Conference on Operational Machine Learning, OpML 2020 - Virtual, Online Duration: Jul 28 2020 → Aug 7 2020

Conference

Conference	2020 USENIX Conference on Operational Machine Learning, OpML 2020
City	Virtual, Online
Period	07/28/20 → 08/7/20

Cite this

@conference{066e4b10539e4dd490b32c289c56eef4,

title = "DLSpec: A deep learning task exchange specification",

abstract = "Deep Learning (DL) innovations are being introduced at a rapid pace. However, the current lack of standard specification of DL tasks makes sharing, running, reproducing, and comparing these innovations difficult. To address this problem, we propose DLSpec, a model-, dataset-, software-, and hardware-agnostic DL specification that captures the different aspects of DL tasks. DLSpec has been tested by specifying and running hundreds of DL tasks.",

author = "Abdul Dakkak and Cheng Li and Jinjun Xiong and Hwu, \{Wen Mei\}",

note = "Publisher Copyright: {\textcopyright} OpML 2020 - 2020 USENIX Conference on Operational Machine Learning. All rights reserved.; 2020 USENIX Conference on Operational Machine Learning, OpML 2020 ; Conference date: 28-07-2020 Through 07-08-2020",

year = "2020",

language = "English",

}

TY - CONF

T1 - DLSpec

T2 - 2020 USENIX Conference on Operational Machine Learning, OpML 2020

AU - Dakkak, Abdul

AU - Li, Cheng

AU - Xiong, Jinjun

AU - Hwu, Wen Mei

PY - 2020

Y1 - 2020

N2 - Deep Learning (DL) innovations are being introduced at a rapid pace. However, the current lack of standard specification of DL tasks makes sharing, running, reproducing, and comparing these innovations difficult. To address this problem, we propose DLSpec, a model-, dataset-, software-, and hardware-agnostic DL specification that captures the different aspects of DL tasks. DLSpec has been tested by specifying and running hundreds of DL tasks.

AB - Deep Learning (DL) innovations are being introduced at a rapid pace. However, the current lack of standard specification of DL tasks makes sharing, running, reproducing, and comparing these innovations difficult. To address this problem, we propose DLSpec, a model-, dataset-, software-, and hardware-agnostic DL specification that captures the different aspects of DL tasks. DLSpec has been tested by specifying and running hundreds of DL tasks.

UR - https://www.scopus.com/pages/publications/85091946925

M3 - Paper

AN - SCOPUS:85091946925

Y2 - 28 July 2020 through 7 August 2020

ER -

DLSpec: A deep learning task exchange specification

Abstract

Conference

Other files and links

Fingerprint

Cite this