Rubertconv Toxic Clf

🚀 RuBERTConv Toxic Classifier

A text classification model for detecting toxicity in Russian text, based on the rubert-base-cased-conversational model.

🚀 Quick Start

The RuBERTConv Toxic Classifier is a text classification model designed to identify toxic content in Russian text. It is built upon the rubert-base-cased-conversational model.

You can try it out using the provided Colab link: link

from transformers import pipeline

model_name = "IlyaGusev/rubertconv_toxic_clf"
pipe = pipeline("text-classification", model=model_name, tokenizer=model_name, framework="pt") 

text = "Ты придурок из интернета"
pipe([text])

✨ Features

• Based on the rubert-base-cased-conversational model, which provides a solid foundation for text classification tasks in Russian.
• Capable of detecting toxic content in various forms of Russian text.

📦 Installation

No specific installation steps are provided in the original README.

💻 Usage Examples

Basic Usage

from transformers import pipeline

model_name = "IlyaGusev/rubertconv_toxic_clf"
pipe = pipeline("text-classification", model=model_name, tokenizer=model_name, framework="pt") 

text = "Ты придурок из интернета"
pipe([text])

📚 Documentation

Intended uses & limitations

This model is intended for text classification tasks related to detecting toxic content in Russian text.

Training data

The model was trained on the following datasets:

• 2ch
• Odnoklassniki
• Toloka Persona Chat Rus
• Koziev's Conversations with toxic words vocabulary

The following data augmentations were applied during training:

• ё -> е
• Remove or add "?" or "!"
• Fix CAPS
• Concatenate toxic and non-toxic texts
• Concatenate two non-toxic texts
• Add toxic words from vocabulary
• Add typos
• Mask toxic words with "*", "@", "$"

📄 License

This project is licensed under the Apache-2.0 license.