Ner Rubert Tiny News

🚀 Ner-rubert-tiny-RuNews

This is a Named Entity Recognition (NER) model designed for Russian news texts.

The model is based on RuBERT-tiny2 and is fine - tuned on the Collection3 dataset. This dataset is rich in news content, allowing the model to effectively identify entities such as Persons, Organizations, and other types of entities in Russian news.

✨ Features

The recognized entity types are as follows:

📦 Installation

No installation steps are provided in the original document, so this section is skipped.

💻 Usage Examples

Basic Usage

from transformers import AutoTokenizer, AutoModelForTokenClassification, pipeline

label2id = {
    'O': 0,
    'B-GEOPOLIT': 1, 'I-GEOPOLIT': 2,
    'B-MEDIA': 3,    'I-MEDIA': 4,
    'B-LOC': 5,      'I-LOC': 6,
    'B-ORG': 7,      'I-ORG': 8,
    'B-PER': 9,      'I-PER': 10
}
id2label = {v: k for k, v in label2id.items()}

model_id = "r1char9/ner-rubert-tiny-RuNews"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForTokenClassification.from_pretrained(
    model_id,
    num_labels=len(label2id),
    id2label=id2label,
    label2id=label2id
)

ner_pipeline = pipeline(
    "ner",
    model=model,
    tokenizer=tokenizer,
    aggregation_strategy="simple"
)

text = (
    "–ì–µ–Ω–µ—Ä–∞–ª—å–Ω—ã–π –¥–∏—Ä–µ–∫—Ç–æ—Ä –°–±–µ—Ä–±–∞–Ω–∫–∞ –ì–µ—Ä–º–∞–Ω –ì—Ä–µ—Ñ –Ω–∞ –∫–æ–Ω—Ñ–µ—Ä–µ–Ω—Ü–∏–∏ –≤ –ú–æ—Å–∫–≤–µ –∑–∞—è–≤–∏–ª, "
    "—á—Ç–æ —Å–æ—Ç—Ä—É–¥–Ω–∏—á–µ—Å—Ç–≤–æ —Å –Ø–Ω–¥–µ–∫—Å–æ–º –≤ –æ–±–ª–∞—Å—Ç–∏ –∏—Å–∫—É—Å—Å—Ç–≤–µ–Ω–Ω–æ–≥–æ –∏–Ω—Ç–µ–ª–ª–µ–∫—Ç–∞ –≤—ã—Ö–æ–¥–∏—Ç –Ω–∞ –Ω–æ–≤—ã–π —É—Ä–æ–≤–µ–Ω—å. "
    "–û–Ω —Ç–∞–∫–∂–µ –æ—Ç–º–µ—Ç–∏–ª, —á—Ç–æ –ø—Ä–∞–≤–∏—Ç–µ–ª—å—Å—Ç–≤–æ –†–æ—Å—Å–∏–π—Å–∫–æ–π –§–µ–¥–µ—Ä–∞—Ü–∏–∏ –ø–æ–¥–¥–µ—Ä–∂–∏–≤–∞–µ—Ç —Ä–∞–∑–≤–∏—Ç–∏–µ —Ü–∏—Ñ—Ä–æ–≤–æ–π —ç–∫–æ–Ω–æ–º–∏–∫–∏, "
    "–æ—Å–æ–±–µ–Ω–Ω–æ –≤ —Ä–∞–º–∫–∞—Ö –ï–≤—Ä–∞–∑–∏–π—Å–∫–æ–≥–æ —ç–∫–æ–Ω–æ–º–∏—á–µ—Å–∫–æ–≥–æ —Å–æ—é–∑–∞."
)

results = ner_pipeline(text)

for entity in results:
    print(entity)

# {'entity_group': 'ORG', 'score': 0.951569, 'word': '–°–±–µ—Ä–±–∞–Ω–∫–∞', 'start': 21, 'end': 30}
# {'entity_group': 'PER', 'score': 0.9922959, 'word': '–ì–µ—Ä–º–∞–Ω –ì—Ä–µ—Ñ', 'start': 31, 'end': 42}
# {'entity_group': 'LOC', 'score': 0.60198957, 'word': '–ú–æ—Å–∫–≤–µ', 'start': 60, 'end': 66}
# {'entity_group': 'GEOPOLIT', 'score': 0.9631994, 'word': '–†–æ—Å—Å–∏–π—Å–∫–æ–π –§–µ–¥–µ—Ä–∞—Ü–∏–∏', 'start': 203, 'end': 223}
# {'entity_group': 'ORG', 'score': 0.85091865, 'word': '–ï–≤—Ä–∞–∑–∏–π—Å–∫–æ–≥–æ —ç–∫–æ–Ω–æ–º–∏—á–µ—Å–∫–æ–≥–æ —Å–æ—é–∑–∞.', 'start': 284, 'end': 318}

Advanced Usage

No advanced usage examples are provided in the original document, so this part is not added.

📚 Documentation

Metrics

🔧 Technical Details

No technical details are provided in the original document, so this section is skipped.

📄 License

This project is licensed under the MIT license.