How to use deepvk/USER2-small with sentence-transformers:
from sentence_transformers import SentenceTransformer model = SentenceTransformer("deepvk/USER2-small") sentences = [ "Это счастливый человек", "Это счастливая собака", "Это очень счастливый человек", "Сегодня солнечный день" ] embeddings = model.encode(sentences) similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [4, 4]
This model has an advantage of shorter tokenization for Russian, but "search_query: " adds whole 6 tokens that neglect all gain for short search queries. It would be great to shorten it or introduce a reserved token or so.Thanks!
· Sign up or log in to comment
