|
νκ΅μΈ μ΄λ¦ μΈμ λͺ¨λΈ |
|
|
|
kor-bert fine-tuning λͺ¨λΈ |
|
|
|
μμ£Ό μμ°λ νκΈμ΄λ¦ κΈ°μ€μΌλ‘ |
|
μμ±κΈ°λ₯Ό λ§λ€μ΄μ, 16λ§κ°μ νκΈ μ΄λ¦μ μμ± ν νμ΅ν λͺ¨λΈμ
λλ€. |
|
|
|
ex) μλ
νμΈμ. μμ€μμ
λλ€. -> μλ
νμΈμ. ***μ
λλ€. |
|
|
|
|
|
```python |
|
from transformers import BertTokenizerFast, BertForTokenClassification |
|
from transformers import pipeline |
|
|
|
model_name = 'joon09/kor-naver-ner-name' |
|
tokenizer = BertTokenizerFast.from_pretrained(model_name) |
|
model = BertForTokenClassification.from_pretrained(model_name) |
|
nlp = pipeline("ner", model=model, tokenizer=tokenizer) |
|
|
|
ner('μλ
νμΈμ. μμ€μμ
λλ€.',grouped_entities=True,aggregation_strategy='average') |
|
|
|
[{'entity_group': 'PER', |
|
'score': 0.99999785, |
|
'word': 'μ', |
|
'start': 7, |
|
'end': 8}, |
|
{'entity_group': 'PER', |
|
'score': 0.82035744, |
|
'word': '##μ€μ', |
|
'start': 8, |
|
'end': 10}] |
|
``` |
|
|
|
|