language: ru tags: - NER - RoBERTa - domain-adaptation - augmentation - job-postings license: mit datasets: - custom: hh.ru metrics: - f1
ruRoBERTa-large (Augmented) for NER on Russian Job Postings
Best-performing model for extracting Skill and Knowledge entities from Russian job texts.
Base model: ai-forever/ruRoBERTa-large
Training: MLM on 9,329 job postings + NER fine-tuning with augmentation (500 samples).
F1: 0.69
Usage
from transformers import AutoTokenizer, AutoModelForTokenClassification, pipeline
model = AutoModelForTokenClassification.from_pretrained("seleznevchamp1/ruroberta-large-NER-augmented")
tok = AutoTokenizer.from_pretrained("seleznevchamp1/ruroberta-large-NER-augmented")
nlp = pipeline("ner", model=model, tokenizer=tok)
print(nlp("Необходимо знание Python и управление проектами"))
- Downloads last month
- 4
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support