Hugging Face's logo

spacy
/
ja_core_news_md

Token Classification
spaCy
Japanese
Eval Results
Model card Files Community
ja_core_news_md / README.md
osanseviero's picture
osanseviero HF staff
Update spaCy pipeline
6f3e94b
preview code
|
raw history blame
No virus
3.83 kB
metadata
tags:
  - spacy
  - token-classification
language:
  - ja
license: cc-by-sa-4.0
model-index:
  - name: ja_core_news_md
    results:
      - task:
          name: NER
          type: token-classification
        metrics:
          - name: NER Precision
            type: precision
            value: 0.737704918
          - name: NER Recall
            type: recall
            value: 0.679245283
          - name: NER F Score
            type: f_score
            value: 0.7072691552
      - task:
          name: POS
          type: token-classification
        metrics:
          - name: POS Accuracy
            type: accuracy
            value: 0.9715755942
      - task:
          name: SENTER
          type: token-classification
        metrics:
          - name: SENTER Precision
            type: precision
            value: 0.9862475442
          - name: SENTER Recall
            type: recall
            value: 0.9901380671
          - name: SENTER F Score
            type: f_score
            value: 0.9881889764
      - task:
          name: UNLABELED_DEPENDENCIES
          type: token-classification
        metrics:
          - name: Unlabeled Dependencies Accuracy
            type: accuracy
            value: 0.9224188392
      - task:
          name: LABELED_DEPENDENCIES
          type: token-classification
        metrics:
          - name: Labeled Dependencies Accuracy
            type: accuracy
            value: 0.9224188392

Details: https://spacy.io/models/ja#ja_core_news_md

Japanese pipeline optimized for CPU. Components: tok2vec, morphologizer, parser, senter, ner, attribute_ruler.

Feature Description
Name ja_core_news_md
Version 3.2.0
spaCy >=3.2.0,<3.3.0
Default Pipeline tok2vec, morphologizer, parser, attribute_ruler, ner
Components tok2vec, morphologizer, parser, senter, attribute_ruler, ner
Vectors 480443 keys, 20000 unique vectors (300 dimensions)
Sources UD Japanese GSD v2.8 (Omura, Mai; Miyao, Yusuke; Kanayama, Hiroshi; Matsuda, Hiroshi; Wakasa, Aya; Yamashita, Kayo; Asahara, Masayuki; Tanaka, Takaaki; Murawaki, Yugo; Matsumoto, Yuji; Mori, Shinsuke; Uematsu, Sumire; McDonald, Ryan; Nivre, Joakim; Zeman, Daniel)
UD Japanese GSD v2.8 NER (Megagon Labs Tokyo)
chiVe: Japanese Word Embedding with Sudachi & NWJC (chive-1.1-mc90-500k) (Works Applications)
License CC BY-SA 4.0
Author Explosion

Label Scheme

View label scheme (66 labels for 4 components)
Component Labels
morphologizer POS=NOUN, POS=ADP, POS=VERB, POS=SCONJ, POS=AUX, POS=PUNCT, POS=PART, POS=DET, POS=NUM, POS=ADV, POS=PRON, POS=ADJ, POS=PROPN, POS=CCONJ, POS=SYM, POS=NOUN|Polarity=Neg, POS=AUX|Polarity=Neg, POS=INTJ, POS=SCONJ|Polarity=Neg
parser ROOT, acl, advcl, advmod, amod, aux, case, cc, ccomp, compound, cop, csubj, dep, det, dislocated, fixed, mark, nmod, nsubj, nummod, obj, obl, punct
senter I, S
ner CARDINAL, DATE, EVENT, FAC, GPE, LANGUAGE, LAW, LOC, MONEY, MOVEMENT, NORP, ORDINAL, ORG, PERCENT, PERSON, PET_NAME, PHONE, PRODUCT, QUANTITY, TIME, TITLE_AFFIX, WORK_OF_ART

Accuracy

Type Score
TOKEN_ACC 99.69
TOKEN_P 97.65
TOKEN_R 97.90
TOKEN_F 97.77
POS_ACC 97.13
MORPH_ACC 0.40
MORPH_MICRO_P 34.01
MORPH_MICRO_R 98.04
MORPH_MICRO_F 50.51
SENTS_P 98.62
SENTS_R 99.01
SENTS_F 98.82
DEP_UAS 92.24
DEP_LAS 90.75
TAG_ACC 97.16
LEMMA_ACC 96.59
ENTS_P 73.77
ENTS_R 67.92
ENTS_F 70.73