Hebrew NER
Collection
3 items • Updated
How to use iahlt/ner-baseline-dictabert-he with Transformers:
# Use a pipeline as a high-level helper
from transformers import pipeline
pipe = pipeline("token-classification", model="iahlt/ner-baseline-dictabert-he") # Load model directly
from transformers import AutoTokenizer, AutoModelForTokenClassification
tokenizer = AutoTokenizer.from_pretrained("iahlt/ner-baseline-dictabert-he")
model = AutoModelForTokenClassification.from_pretrained("iahlt/ner-baseline-dictabert-he") Label P R F1 support
----------------------------------------------
ANG 0.76 0.86 0.81 91
DUC 0.46 0.53 0.49 323
EVE 0.55 0.67 0.61 663
FAC 0.52 0.54 0.53 804
GPE 0.88 0.86 0.87 4915
INFORMAL 0.00 0.00 0.00 8
LOC 0.58 0.69 0.63 762
MISC 0.62 0.63 0.63 1548
ORG 0.75 0.79 0.77 5672
PER 0.94 0.96 0.95 7892
TIMEX 0.79 0.84 0.81 2623
TTL 0.68 0.72 0.70 2830
WOA 0.82 0.86 0.84 1102
---------------------------------------------
Mean 0.80 0.83 0.81 Total: 29233
For additional information and the dataset the model was trained on, see https://github.com/IAHLT/hebrew_named_entities_open_dataset.