N3LDG: 一种轻量级自然语言处理深度学习库

Qiansheng Wang; Nan Yu; Meishan Zhang; Zijia Han; Guohong Fu

doi:10.13209/j.0479-8023.2018.065

N3LDG: 一种轻量级自然语言处理深度学习库

Translated title of the contribution: N3LDG: A Lightweight Neural Network Library for Natural Language Processing

Qiansheng Wang
, Nan Yu
, Meishan Zhang
, Zijia Han
, Guohong Fu^*

^*Corresponding author for this work

Heilongjiang University

Research output: Contribution to journal › Article › peer-review

1 Scopus citations

Abstract

The authors propose a neural network library N3LDG for natural language processing. N3LDG sup-ports constructing computation graphs dynamically, and organizing executions into batches automatically. Experi-ments show that N3LDG can efficiently construct and execute computation graphs when training CNN, Bi-LSTM, and Tree-LSTM. When using CPU to train above models, the training speed of N3LDG is better than that of PyTorch. When using GPU to train CNN and Tree-LSTM, N3LDG is better than PyTorch.

Translated title of the contribution	N3LDG: A Lightweight Neural Network Library for Natural Language Processing
Original language	Chinese (Traditional)
Pages (from-to)	113-119
Number of pages	7
Journal	Beijing Daxue Xuebao (Ziran Kexue Ban)/Acta Scientiarum Naturalium Universitatis Pekinensis
Volume	55
Issue number	1
DOIs	https://doi.org/10.13209/j.0479-8023.2018.065
State	Published - 20 Jan 2019
Externally published	Yes

Access to Document

10.13209/j.0479-8023.2018.065

Cite this

@article{fb6c129351ed449989cbf41310d20985,

title = "N3LDG: 一种轻量级自然语言处理深度学习库",

abstract = "The authors propose a neural network library N3LDG for natural language processing. N3LDG sup-ports constructing computation graphs dynamically, and organizing executions into batches automatically. Experi-ments show that N3LDG can efficiently construct and execute computation graphs when training CNN, Bi-LSTM, and Tree-LSTM. When using CPU to train above models, the training speed of N3LDG is better than that of PyTorch. When using GPU to train CNN and Tree-LSTM, N3LDG is better than PyTorch.",

keywords = "CUDA, Deep learning library, Lightweight, NLP",

author = "Qiansheng Wang and Nan Yu and Meishan Zhang and Zijia Han and Guohong Fu",

note = "Publisher Copyright: {\textcopyright} 2019 Peking University.",

year = "2019",

month = jan,

day = "20",

doi = "10.13209/j.0479-8023.2018.065",

language = "繁体中文",

volume = "55",

pages = "113--119",

journal = "Beijing Daxue Xuebao (Ziran Kexue Ban)/Acta Scientiarum Naturalium Universitatis Pekinensis",

issn = "0479-8023",

publisher = "Peking University",

number = "1",

}

TY - JOUR

T1 - N3LDG

T2 - 一种轻量级自然语言处理深度学习库

AU - Wang, Qiansheng

AU - Yu, Nan

AU - Zhang, Meishan

AU - Han, Zijia

AU - Fu, Guohong

PY - 2019/1/20

Y1 - 2019/1/20

N2 - The authors propose a neural network library N3LDG for natural language processing. N3LDG sup-ports constructing computation graphs dynamically, and organizing executions into batches automatically. Experi-ments show that N3LDG can efficiently construct and execute computation graphs when training CNN, Bi-LSTM, and Tree-LSTM. When using CPU to train above models, the training speed of N3LDG is better than that of PyTorch. When using GPU to train CNN and Tree-LSTM, N3LDG is better than PyTorch.

AB - The authors propose a neural network library N3LDG for natural language processing. N3LDG sup-ports constructing computation graphs dynamically, and organizing executions into batches automatically. Experi-ments show that N3LDG can efficiently construct and execute computation graphs when training CNN, Bi-LSTM, and Tree-LSTM. When using CPU to train above models, the training speed of N3LDG is better than that of PyTorch. When using GPU to train CNN and Tree-LSTM, N3LDG is better than PyTorch.

KW - CUDA

KW - Deep learning library

KW - Lightweight

KW - NLP

UR - https://www.scopus.com/pages/publications/85062974611

U2 - 10.13209/j.0479-8023.2018.065

DO - 10.13209/j.0479-8023.2018.065

M3 - 文章

AN - SCOPUS:85062974611

SN - 0479-8023

VL - 55

SP - 113

EP - 119

JO - Beijing Daxue Xuebao (Ziran Kexue Ban)/Acta Scientiarum Naturalium Universitatis Pekinensis

JF - Beijing Daxue Xuebao (Ziran Kexue Ban)/Acta Scientiarum Naturalium Universitatis Pekinensis

IS - 1

ER -

N3LDG: 一种轻量级自然语言处理深度学习库

Abstract

Access to Document

Other files and links

Fingerprint

Cite this