de_MRP_NER / README.md
osanseviero's picture
Update spaCy pipeline
5a6da7d
|
raw
history blame
1.15 kB
metadata
tags:
  - spacy
  - token-classification
language:
  - de
license: cc-by
model-index:
  - name: de_MRP_NER
    results:
      - task:
          name: NER
          type: token-classification
        metrics:
          - name: NER Precision
            type: precision
            value: 0.905255631
          - name: NER Recall
            type: recall
            value: 0.8568527919
          - name: NER F Score
            type: f_score
            value: 0.8803894298

NER Model for 'Ministerratsprotokolle'

Feature Description
Name de_MRP_NER
Version 0.0.0
spaCy >=3.1.0,<3.2.0
Default Pipeline tok2vec, ner
Components tok2vec, ner
Vectors 0 keys, 0 unique vectors (0 dimensions)
Sources n/a
License cc-by
Author Peter Andorfer

Label Scheme

View label scheme (4 labels for 1 components)
Component Labels
ner GPE, LOC, ORG, PER

Accuracy

Type Score
ENTS_F 88.04
ENTS_P 90.53
ENTS_R 85.69
TOK2VEC_LOSS 40077.56
NER_LOSS 77727.57