nkjp/nkjp-ner
Viewer • Updated • 19.8k • 214 • 1
How to use CodeNLP/pdn2_v08_nkjp_large with Transformers:
# Use a pipeline as a high-level helper
from transformers import pipeline
pipe = pipeline("feature-extraction", model="CodeNLP/pdn2_v08_nkjp_large") # Load model directly
from transformers import AutoTokenizer, AutoModel
tokenizer = AutoTokenizer.from_pretrained("CodeNLP/pdn2_v08_nkjp_large")
model = AutoModel.from_pretrained("CodeNLP/pdn2_v08_nkjp_large")A transformer-based model for named entity recognition for Polish. The model was trained on the NKJP corpus to recognize 12 categories of NEs.
conda create -n pdn2_demo python=3.8
conda activate pdn2_demo
https://pypi.clarin-pl.eu/packages/poldeepner2-0.8.2-py3-none-any.whl
git clone https://huggingface.co/CodeNLP/pdn2_v08_nkjp_large
import poldeepner2
ner = poldeepner2.load("pdn2_v08_nkjp_large", device="cpu")
text = "Od 15 września 2023 roku premierem Łotwy jest Evika Siliņa"
for an in ner.process_text(text):
print(f"[{an.start}:{an.end}] {an.text} ({an.label})")
Expected output:
[3:24] 15 września 2023 roku (date)
[35:40] Łotwy (placeName_country)
[46:58] Evika Siliņa (persName)
[46:51] Evika (persName_forename)
[52:58] Siliņa (persName_surname)
Michał Marcińczuk marcinczuk@gmail.com