Den4ikAI/ruBert-tiny-questions-classifier
Модель классифицирует вопрос на два класса:
- Точный вопрос (exact_question) - вопрос требующий четкого, фактологичного ответа.
- Неточный вопрос (innacurate_question) - вопрос допускающий рассуждения, философские вопросы.
Использование
import torch
from transformers import AutoTokenizer, AutoModelForSequenceClassification
class RuBertQuestionClassifier:
def __init__(self):
self.device = torch.device("cpu")
self.tokenizer = AutoTokenizer.from_pretrained("Den4ikAI/ruBert-tiny-questions-classifier")
self.model = AutoModelForSequenceClassification.from_pretrained("Den4ikAI/ruBert-tiny-questions-classifier").to(
self.device).eval()
self.classes = ['exact_question', 'inaccurate_question']
def get_question_type(self, text):
text = text.lower().replace(',', '').replace('?', '')
inputs = self.tokenizer(text, max_length=512, add_special_tokens=False, truncation=True,
return_tensors='pt').to(self.device)
with torch.no_grad():
logits = self.model(**inputs).logits
probas = list(torch.sigmoid(logits)[0].cpu().detach().numpy())
out = self.classes[probas.index(max(probas))]
self.logger.debug('Mode: {}'.format(out))
return out
classifier = RuBertQuestionClassifier()
print(classifier.get_question_type('когда я найду своего принца'))
print(classifier.get_question_type('что такое цианид'))
Citation
@MISC{Den4ikAI/ruBert-tiny-questions-classifier,
author = {Denis Petrov},
title = {Russian question type classifier model},
url = {https://huggingface.co/Den4ikAI/ruBert-tiny-questions-classifier},
year = 2023
}
- Downloads last month
- 20
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.