Rubert Tiny2 Russian Sentiment

🚀 RuBERT-tiny2 for Russian Sentiment Classification

This project presents a fine - tuned RuBERT-tiny2 model designed for the sentiment classification of short Russian texts. It addresses a multi - class classification task with well - defined labels.

🚀 Quick Start

💻 Usage Examples

Basic Usage

from transformers import pipeline
model = pipeline(model="seara/rubert-tiny2-russian-sentiment")
model("Привет, ты мне нравишься!")
# [{'label': 'positive', 'score': 0.9398769736289978}]

📚 Documentation

Dataset

This model was trained on the union of the following datasets:

• Kaggle Russian News Dataset
• Linis Crowd 2015
• Linis Crowd 2016
• RuReviews
• RuSentiment

An overview of the training data can be found on S. Smetanin Github repository.

Download links for all Russian sentiment datasets collected by Smetanin can be found in this repository.

Training

Training was conducted in this project with the following parameters:

tokenizer.max_length: 512
batch_size: 64
optimizer: adam
lr: 0.00001
weight_decay: 0
epochs: 5

The Train/validation/test splits are 80%/10%/10%.

Eval results (on test split)

📄 License

This project is licensed under the MIT license.

✨ Features

• Model Type: Fine - tuned RuBERT - tiny2 for Russian sentiment classification.
• Task: Multi - class classification for sentiment analysis of short Russian texts.
• Labels:

0: neutral
1: positive
2: negative

• Label Mapping:

neutral: нейтральный
positive: позитивный
negative: негативный