pytorch_joint-word-segmentation-and-pos-tagging's Introduction

JointPS

A re-implementation of A Simple and Effective Neural Model for Joint Word Segmentation and POS Tagging based on PyTorch.

The C++ Code [zhangmeishan/NNTranJSTagger].

PyTorch-0.3.1 Code release on here. [PyTorch-0.3.1]

Requirement

pip3 install -r requirements.txt
Python  == 3.6  
PyTorch == 1.0.1

Usage

modify the config file, detail see the Config directory
Train:
(1) sh run_train_p.sh
(2) python -u main.py --config ./Config/config.cfg --device cuda:0--train -p 
    [device: "cpu", "cuda:0", "cuda:1", ......]

Config

optimizer: Adam
lr: 0.001
dropout: 0.25
embed_char_dim: 200
embed_bichar_dim: 200
rnn_dim: 200
rnn_hidden_dim: 200
pos_dim: 100
oov: avg 
Refer to config.cfg file for more details.

final log in [final_log]

Network Structure

Performance

	CTB5	CTB6	CTB7	PKU	NCC
Model	SEG POS	SEG POS	SEG POS	SEG POS	SEG POS
Our Model (No External Embeddings)	97.69 94.16	95.37 90.83	95.32 90.25	95.22 92.62	93.97 89.47
Our Model (Basic Embeddings)	97.93 94.44	95.78 91.79	95.77 91.12	95.82 93.42	94.52 89.82
Our Model (Word-context Embeddings)	98.50 94.95	96.36 92.51	96.25 91.87	96.35 94.14	95.30 90.42

Cite

@Article{zhang2018jointposseg,  
  author    = {Zhang, Meishan and Yu, Nan and Fu, Guohong},  
  title     = {A Simple and Effective Neural Model for Joint Word Segmentation and POS Tagging},  
  journal   = {IEEE/ACM Transactions on Audio, Speech and Language Processing (TASLP)},  
  year      = {2018},  
  volume    = {26},  
  number    = {9},
  pages     = {1528--1538},
  publisher = {IEEE Press},
}

Question

if you have any question, you can open a issue or email [email protected]、[email protected]、bamtercelboo@{gmail.com, 163.com}.
if you have any good suggestions, you can PR or email me.

Recommend Projects