Copertina

Terra: Imperative-Symbolic Co-Execution of Imperative Deep Learning Programs

Salvato in:
Pubblicato in:arXiv.org (Jan 23, 2022), p. n/a
Autore principale: Kim, Taebum
Altri autori: Jeong, Eunji, Geon-Woo, Kim, Koo, Yunmo, Kim, Sehoon, Yu, Gyeong-In, Byung-Gon Chun
Pubblicazione:
Cornell University Library, arXiv.org
Soggetti:
Usability
Dynamic control
Deep learning
Performance evaluation
Learning programs
Machine learning
Imperative programming
Decoupling
Artificial neural networks
Accesso online:Citation/Abstract
Full text outside of ProQuest
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!

MARC

LEADER 00000nab a2200000uu 4500
001 2622690773
003 UK-CbPIL
022 |a 2331-8422 
035 |a 2622690773 
045 0 |b d20220123 
100 1 |a Kim, Taebum 
245 1 |a Terra: Imperative-Symbolic Co-Execution of Imperative Deep Learning Programs 
260 |b Cornell University Library, arXiv.org  |c Jan 23, 2022 
513 |a Working Paper 
520 3 |a Imperative programming allows users to implement their deep neural networks (DNNs) easily and has become an essential part of recent deep learning (DL) frameworks. Recently, several systems have been proposed to combine the usability of imperative programming with the optimized performance of symbolic graph execution. Such systems convert imperative Python DL programs to optimized symbolic graphs and execute them. However, they cannot fully support the usability of imperative programming. For example, if an imperative DL program contains a Python feature with no corresponding symbolic representation (e.g., third-party library calls or unsupported dynamic control flows) they fail to execute the program. To overcome this limitation, we propose Terra, an imperative-symbolic co-execution system that can handle any imperative DL programs while achieving the optimized performance of symbolic graph execution. To achieve this, Terra builds a symbolic graph by decoupling DL operations from Python features. Then, Terra conducts the imperative execution to support all Python features, while delegating the decoupled operations to the symbolic execution. We evaluated the performance improvement and coverage of Terra with ten imperative DL programs for several DNN architectures. The results show that Terra can speed up the execution of all ten imperative DL programs, whereas AutoGraph, one of the state-of-the-art systems, fails to execute five of them. 
653 |a Usability 
653 |a Dynamic control 
653 |a Deep learning 
653 |a Performance evaluation 
653 |a Learning programs 
653 |a Machine learning 
653 |a Imperative programming 
653 |a Decoupling 
653 |a Artificial neural networks 
700 1 |a Jeong, Eunji 
700 1 |a Geon-Woo, Kim 
700 1 |a Koo, Yunmo 
700 1 |a Kim, Sehoon 
700 1 |a Yu, Gyeong-In 
700 1 |a Byung-Gon Chun 
773 0 |t arXiv.org  |g (Jan 23, 2022), p. n/a 
786 0 |d ProQuest  |t Engineering Database 
856 4 1 |3 Citation/Abstract  |u https://www.proquest.com/docview/2622690773/abstract/embedded/7BTGNMKEMPT1V9Z2?source=fedsrch 
856 4 0 |3 Full text outside of ProQuest  |u http://arxiv.org/abs/2201.09210