USER-bge-m3
Universal Sentence Encoder for Russian (USER) is a sentence-transformer model for extracting embeddings exclusively for Russian language. It maps sentences & paragraphs to a 1024 dimensional dense vector space and can be used for tasks like clustering or semantic search.
This model is initialized from TatonkaHF/bge-m3_en_ru
which is shrinked version of baai/bge-m3
model and trained to work mainly with the Russian language. Its quality on other languages was not evaluated.
Usage
Using this model becomes easy when you have sentence-transformers
installed:
pip install -U sentence-transformers
Then you can use the model like this:
from sentence_transformers import SentenceTransformer
input_texts = [
"ΠΠΎΠ³Π΄Π° Π±ΡΠ» ΡΠΏΡΡΠ΅Π½ Π½Π° Π²ΠΎΠ΄Ρ ΠΏΠ΅ΡΠ²ΡΠΉ ΠΌΠΈΠ½ΠΎΠ½ΠΎΡΠ΅Ρ Β«Π‘ΠΏΠΎΠΊΠΎΠΉΠ½ΡΠΉΒ»?",
"ΠΡΡΡ Π»ΠΈ Π½Π΅ΡΡΡ Π² Π£Π΄ΠΌΡΡΡΠΈΠΈ?",
"Π‘ΠΏΠΎΠΊΠΎΠΉΠ½ΡΠΉ (ΡΡΠΌΠΈΠ½Π΅Ρ)\nΠΠ°ΡΠΈΡΠ»Π΅Π½ Π² ΡΠΏΠΈΡΠΊΠΈ ΠΠΠ€ Π‘Π‘Π‘Π 19 Π°Π²Π³ΡΡΡΠ° 1952 Π³ΠΎΠ΄Π°.",
"ΠΠ΅ΡΡΠ΅ΠΏΠΎΠΈΡΠΊΠΎΠ²ΡΠ΅ ΡΠ°Π±ΠΎΡΡ Π² Π£Π΄ΠΌΡΡΡΠΈΠΈ Π±ΡΠ»ΠΈ Π½Π°ΡΠ°ΡΡ ΡΡΠ°Π·Ρ ΠΏΠΎΡΠ»Π΅ ΠΡΠΎΡΠΎΠΉ ΠΌΠΈΡΠΎΠ²ΠΎΠΉ Π²ΠΎΠΉΠ½Ρ Π² 1945 Π³ΠΎΠ΄Ρ ΠΈ ΠΏΡΠΎΠ΄ΠΎΠ»ΠΆΠ°ΡΡΡΡ ΠΏΠΎ ΡΠ΅ΠΉ Π΄Π΅Π½Ρ. ΠΠΎΠ±ΡΡΠ° Π½Π΅ΡΡΠΈ Π½Π°ΡΠ°Π»Π°ΡΡ Π² 1967 Π³ΠΎΠ΄Ρ."
]
model = SentenceTransformer("deepvk/USER-bge-m3")
embeddings = model.encode(input_texts, normalize_embeddings=True)
However, you can use model directly with transformers
import torch.nn.functional as F
from torch import Tensor, inference_mode
from transformers import AutoTokenizer, AutoModel
input_texts = [
"ΠΠΎΠ³Π΄Π° Π±ΡΠ» ΡΠΏΡΡΠ΅Π½ Π½Π° Π²ΠΎΠ΄Ρ ΠΏΠ΅ΡΠ²ΡΠΉ ΠΌΠΈΠ½ΠΎΠ½ΠΎΡΠ΅Ρ Β«Π‘ΠΏΠΎΠΊΠΎΠΉΠ½ΡΠΉΒ»?",
"ΠΡΡΡ Π»ΠΈ Π½Π΅ΡΡΡ Π² Π£Π΄ΠΌΡΡΡΠΈΠΈ?",
"Π‘ΠΏΠΎΠΊΠΎΠΉΠ½ΡΠΉ (ΡΡΠΌΠΈΠ½Π΅Ρ)\nΠΠ°ΡΠΈΡΠ»Π΅Π½ Π² ΡΠΏΠΈΡΠΊΠΈ ΠΠΠ€ Π‘Π‘Π‘Π 19 Π°Π²Π³ΡΡΡΠ° 1952 Π³ΠΎΠ΄Π°.",
"ΠΠ΅ΡΡΠ΅ΠΏΠΎΠΈΡΠΊΠΎΠ²ΡΠ΅ ΡΠ°Π±ΠΎΡΡ Π² Π£Π΄ΠΌΡΡΡΠΈΠΈ Π±ΡΠ»ΠΈ Π½Π°ΡΠ°ΡΡ ΡΡΠ°Π·Ρ ΠΏΠΎΡΠ»Π΅ ΠΡΠΎΡΠΎΠΉ ΠΌΠΈΡΠΎΠ²ΠΎΠΉ Π²ΠΎΠΉΠ½Ρ Π² 1945 Π³ΠΎΠ΄Ρ ΠΈ ΠΏΡΠΎΠ΄ΠΎΠ»ΠΆΠ°ΡΡΡΡ ΠΏΠΎ ΡΠ΅ΠΉ Π΄Π΅Π½Ρ. ΠΠΎΠ±ΡΡΠ° Π½Π΅ΡΡΠΈ Π½Π°ΡΠ°Π»Π°ΡΡ Π² 1967 Π³ΠΎΠ΄Ρ."
]
tokenizer = AutoTokenizer.from_pretrained("deepvk/USER-bge-m3")
model = AutoModel.from_pretrained("deepvk/USER-bge-m3")
model.eval()
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
with torch.no_grad():
model_output = model(**encoded_input)
# Perform pooling. In this case, cls pooling.
sentence_embeddings = model_output[0][:, 0]
# normalize embeddings
sentence_embeddings = torch.nn.functional.normalize(sentence_embeddings, p=2, dim=1)
# [[0.5567, 0.3014],
# [0.1701, 0.7122]]
scores = (sentence_embeddings[:2] @ sentence_embeddings[2:].T)
Also, you can use native FlagEmbedding library for evaluation. Usage is described in bge-m3
model card.
Training Details
We follow the USER-base
model training algorithm, with several changes as we use different backbone.
Initialization: TatonkaHF/bge-m3_en_ru
β shrinked version of baai/bge-m3
to support only Russian and English tokens.
Fine-tuning: Supervised fine-tuning two different models based on data symmetry and then merging via LM-Cocktail
:
Since we split the data, we could additionally apply the AnglE loss to the symmetric model, which enhances performance on symmetric tasks.
Finally, we added the original
bge-m3
model to the two obtained models to prevent catastrophic forgetting, tuning the weights for the merger usingLM-Cocktail
to produce the final model, USER-bge-m3.
Dataset
During model development, we additional collect 2 datasets:
deepvk/ru-HNP
and
deepvk/ru-WANLI
.
Symmetric Dataset | Size | Asymmetric Dataset | Size |
---|---|---|---|
AllNLI | 282 644 | MIRACL | 10 000 |
MedNLI | 3 699 | MLDR | 1 864 |
RCB | 392 | Lenta | 185 972 |
Terra | 1 359 | Mlsum | 51 112 |
Tapaco | 91 240 | Mr-TyDi | 536 600 |
deepvk/ru-WANLI | 35 455 | Panorama | 11 024 |
deepvk/ru-HNP | 500 000 | PravoIsrael | 26 364 |
Xlsum | 124 486 | ||
Fialka-v1 | 130 000 | ||
RussianKeywords | 16 461 | ||
Gazeta | 121 928 | ||
Gsm8k-ru | 7 470 | ||
DSumRu | 27 191 | ||
SummDialogNews | 75 700 |
Total positive pairs: 2,240,961 Total negative pairs: 792,644 (negative pairs from AIINLI, MIRACL, deepvk/ru-WANLI, deepvk/ru-HNP)
For all labeled datasets, we only use its training set for fine-tuning. For datasets Gazeta, Mlsum, Xlsum: pairs (title/text) and (title/summary) are combined and used as asymmetric data.
AllNLI
is an translated to Russian combination of SNLI, MNLI and ANLI.
Experiments
We compare our mode with the basic baai/bge-m3
on the encodechka
benchmark.
In addition, we evaluate model on the russian subset of MTEB
on Classification, Reranking, Multilabel Classification, STS, Retrieval, and PairClassification tasks.
We use validation scripts from the official repositories for each of the tasks.
Results on encodechka:
Model | Mean S | Mean S+W | STS | PI | NLI | SA | TI | IA | IC | ICX | NE1 | NE2 |
---|---|---|---|---|---|---|---|---|---|---|---|---|
baai/bge-m3 |
0.787 | 0.696 | 0.86 | 0.75 | 0.51 | 0.82 | 0.97 | 0.79 | 0.81 | 0.78 | 0.24 | 0.42 |
USER-bge-m3 |
0.799 | 0.709 | 0.87 | 0.76 | 0.58 | 0.82 | 0.97 | 0.79 | 0.81 | 0.78 | 0.28 | 0.43 |
Results on MTEB:
Type | baai/bge-m3 |
USER-bge-m3 |
---|---|---|
Average (30 datasets) | 0.689 | 0.706 |
Classification Average (12 datasets) | 0.571 | 0.594 |
Reranking Average (2 datasets) | 0.698 | 0.688 |
MultilabelClassification (2 datasets) | 0.343 | 0.359 |
STS Average (4 datasets) | 0.735 | 0.753 |
Retrieval Average (6 datasets) | 0.945 | 0.934 |
PairClassification Average (4 datasets) | 0.784 | 0.833 |
Limitations
We did not thoroughly evaluate the model's ability for sparse and multi-vec encoding.
Citations
@misc{deepvk2024user,
title={USER: Universal Sentence Encoder for Russian},
author={Malashenko, Boris and Zemerov, Anton and Spirin, Egor},
url={https://huggingface.co/datasets/deepvk/USER-base},
publisher={Hugging Face}
year={2024},
}
- Downloads last month
- 44,762