Skip to content
A Japanese NLP Library using spaCy as framework based on Universal Dependencies
Branch: develop
Clone or download

GiNZA NLP Library

GiNZA logo

An Open Source Japanese NLP Library, based on Universal Dependencies


GiNZA NLP Library and GiNZA Japanese Universal Dependencies Models are distributed under The MIT License. You must agree and follow The MIT License to use GiNZA NLP Library and GiNZA Japanese Universal Dependencies Models.


spaCy is the key framework of GiNZA. spaCy LICENSE PAGE

Sudachi and SudachiPy

SudachiPy provides high accuracies for tokenization and pos tagging. Sudachi LICENSE PAGE, SudachiPy LICENSE PAGE

Runtime Environment

This project is developed with Python>=3.6 and pip for it.

(Please see Development Environment section located on bottom too)

Runtime set up

1. Install GiNZA NLP Library with Japanese Universal Dependencies Model

Run following line

$ pip install ""

or download pip install archive from release page and run pip install with it.

$ pip install ginza-2.0.0.tar.gz

2. Execute ginza from command line

Run ginza command from the console, then input some Japanese text. After pressing enter key, you will get the parsed results with conllu format.

$ ginza

Coding example

Following steps shows dependency parsing results with sentence boundary 'EOS'.

import spacy
nlp = spacy.load('ja_ginza')
doc = nlp('依存構造解析の実験を行っています。')
for sent in doc.sents:
    for token in sent:
        print(token.i, token.orth_, token.lemma_, token.pos_, token.tag_, token.dep_, token.head.i)


Please see spaCy API documents.


version 2.0

ginza-2.0.0 (2019-07-08)

  • Add ginza command
    • run ginza from the console
  • Change package structure
    • module package as ginza
    • language model package as ja_ginza
    • spacy.lang.ja is overridden by ginza
  • Remove sudachipy related directories
    • SudachiPy and its dictionary are installed via pip during ginza installation
  • User dictionary available
  • Token extension fields
    • Added
      • token._.bunsetu_bi_label, token._.bunsetu_position_type
    • Remained
      • token._.inf
    • Removed
      • pos_detail (same value is set to token.tag_)

version 1.0

ja_ginza_nopn-1.0.2 (2019-04-07)

  • Set depending token index of root as 0 to meet with conllu format definitions

ja_ginza_nopn-1.0.1 (2019-04-02)

  • Add new Japanese era 'reiwa' to system_core.dic.

ja_ginza_nopn-1.0.0 (2019-04-01)

  • First release version

Development Environment

Development set up

1. Clone from github

$ git clone ''

2. Run python

For normal environment:

$ python develop


To be described

You can’t perform that action at this time.