Skip to content

findatalab/llm-city

Repository files navigation

llm-city

Virtual model of a city in which residents are represented using large language models (LLM).

Baseline модели для анализа тональности русскоязычных отзывов
(negative / neutral / positive).

Dataset: RuReviews
https://github.com/sismetanin/rureviews

Требования

Python 3.11+

Установленные библиотеки из requirements.txt.

Модели

  • TF-IDF + Logistic Regression
  • TF-IDF + Linear SVM
  • RuBERT (DeepPavlov/rubert-base-cased)

Результаты

TF-IDF + Linear SVM: accuracy ≈ 0.70,
TF-IDF + Logistic Regression: accuracy ≈ 0.73,
RuBERT: accuracy ≈ 0.78.

pip install -r requirements.txt

About

Virtual model of a city in which residents are represented using large language models (LLM).

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors