B

BGE M3 Ko

Developed by dragonkue
A Korean-English bilingual sentence embedding model optimized based on BAAI/bge-m3, supporting semantic text similarity, information retrieval, and other tasks
Downloads 29.78k
Release Time : 9/17/2024

Model Overview

This is a model trained on the sentence-transformers framework, specifically optimized for Korean and English. It maps sentences and paragraphs into a 1024-dimensional dense vector space, suitable for tasks such as semantic text similarity, semantic search, paraphrase mining, text classification, and clustering.

Model Features

Korean Optimization
Specially trained and optimized for Korean based on the standard BGE-M3
Long Text Support
Supports sequences up to 8192 tokens, suitable for processing longer texts
High-Performance Retrieval
Outstanding performance in Korean embedding benchmarks, achieving a Top-1 F1 score of 0.7456
Multiple Similarity Calculations
Supports both cosine similarity and dot product similarity calculations

Model Capabilities

Semantic text similarity calculation
Information retrieval
Text feature extraction
Text classification
Text clustering
Paraphrase mining

Use Cases

Information Retrieval
Korean Document Retrieval
Retrieve the most relevant documents from a Korean document library based on query statements
Achieved an F1 score of 0.7456 in Top-1 retrieval
Text Similarity
Similar Question Matching
Identify semantically similar questions with different expressions
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase