Pesets

Description

Pesets is a Named Entity Recognition (NER) model based on the BERT architecture, designed to extract and classify named entities from unstructured text. It is specifically trained to work with Russian-language texts and datasets, enabling it to process and understand both the context and the semantic structure of the data in Russian. Pesets is optimized for NER, providing accurate and efficient solutions for information extraction tasks in a Russian-language environment.

👨‍💻 Examples of usage

from transformers import pipeline

nlp = pipeline("token-classification", model="MadShift/Pesets", aggregation_strategy="simple")

text = "Сколько бюджета израсходовано в ДО «Север» на Иванова И.?"
results = nlp(text)

for entity in results:
    print(f"Сущность: {entity['word']} | Тип: {entity['entity_group']} | Точность: {entity['score']:.4f}")