metadata
tags:
- spacy
- token-classification
language:
- en
license: mit
model-index:
- name: en_core_med7_lg
results:
- task:
name: NER
type: token-classification
metrics:
- name: NER Precision
type: precision
value: 0.8778156997
- name: NER Recall
type: recall
value: 0.8840918466
- name: NER F Score
type: f_score
value: 0.8809425949
| Feature | Description |
|---|---|
| Name | en_core_med7_lg |
| Version | 3.1.3.1 |
| spaCy | >=3.1.4,<3.2.0 |
| Default Pipeline | tok2vec, ner |
| Components | tok2vec, ner |
| Vectors | 684830 keys, 684830 unique vectors (300 dimensions) |
| Sources | n/a |
| License | MIT |
| Author | Andrey Kormilitzin |
Label Scheme
View label scheme (7 labels for 1 components)
| Component | Labels |
|---|---|
ner |
DOSAGE, DRUG, DURATION, FORM, FREQUENCY, ROUTE, STRENGTH |
Accuracy
| Type | Score |
|---|---|
ENTS_F |
88.09 |
ENTS_P |
87.78 |
ENTS_R |
88.41 |
TOK2VEC_LOSS |
115648.09 |
NER_LOSS |
279069.77 |