This is the detoxification baseline model trained on the
train
part of "RUSSE 2022: Russian Text Detoxification Based on Parallel Corpora" competition. The source sentences are Russian toxic messages from Odnoklassniki, Pikabu, and Twitter platforms. The base model is
ruT5
provided from Sber.
How to use
from transformers import T5ForConditionalGeneration, AutoTokenizer
base_model_name = 'sberbank-ai/ruT5-base'
model_name = 'SkolkovoInstitute/ruT5-base-detox'
tokenizer = AutoTokenizer.from_pretrained(base_model_name)
model = T5ForConditionalGeneration.from_pretrained(model_name)