Vietnamese Embedding

🚀 Vietnamese_Embedding

Vietnamese_Embedding is an embedding model fine-tuned from the BGE-M3 model to enhance retrieval capabilities for Vietnamese. It addresses the need for effective Vietnamese text retrieval and provides high - quality embedding representations for Vietnamese language processing.

🚀 Quick Start

from sentence_transformers import SentenceTransformer
import torch

model = SentenceTransformer("AITeamVN/Vietnamese_Embedding")
model.max_seq_length = 2048
sentences_1 = ["Trí tuệ nhân tạo là gì", "Lợi ích của giấc ngủ"]
sentences_2 = ["Trí tuệ nhân tạo là công nghệ giúp máy móc suy nghĩ và học hỏi như con người. Nó hoạt động bằng cách thu thập dữ liệu, nhận diện mẫu và đưa ra quyết định.", 
               "Giấc ngủ giúp cơ thể và não bộ nghỉ ngơi, hồi phục năng lượng và cải thiện trí nhớ. Ngủ đủ giấc giúp tinh thần tỉnh táo và làm việc hiệu quả hơn."]
query_embedding = model.encode(sentences_1)
doc_embeddings = model.encode(sentences_2)
similarity = query_embedding @ doc_embeddings.T
print(similarity)

'''
array([[0.66212064, 0.33066642],
       [0.25866613, 0.5865289 ]], dtype=float32)
'''

✨ Features

• Fine - tuned from the BGE - M3 model to enhance Vietnamese retrieval capabilities.
• Trained on approximately 300,000 triplets of Vietnamese queries, positive documents, and negative documents.
• Trained with a maximum sequence length of 2048.

📦 Installation

The installation can be achieved by installing the sentence-transformers library. You can use the following command:

pip install sentence-transformers

📚 Documentation

Model Details

Evaluation

• Dataset: Entire training dataset of Legal Zalo 2021. Our model was not trained on this dataset.

Vietnamese_Reranker and Vietnamese_Embedding_v2 were trained on 1100000 triplets. Although the score on the legal domain drops a bit on Vietnamese_Embedding_v2, since this phase data is much larger, it is very good for other domains.

You can access 2 models via links: Vietnamese_Embedding_v2, Vietnamese_Reranker

You can reproduce the evaluation result by running the Python code evaluation_model.py (data downloaded from Kaggle).

📄 License

This model is licensed under the Apache 2.0 license.

Contact

Email: nguyennhotrung3004@gmail.com

Developer

Member: Nguyễn Nho Trung, Nguyễn Nhật Quang

Citation

@misc{Vietnamese_Embedding,
  title={Vietnamese_Embedding: Embedding model in Vietnamese language.},
  author={Nguyen Nho Trung, Nguyen Nhat Quang},
  year={2025},
  publisher={Huggingface},
}