M

Mstsb Paraphrase Multilingual Mpnet Base V2

Developed by AIDA-UPM
Fine-tuned version of the paraphrase-multilingual-mpnet-base-v2 model based on sentence-transformers, optimized for semantic text similarity tasks in 15 languages
Downloads 404
Release Time : 3/2/2022

Model Overview

This model maps sentences and paragraphs to a 768-dimensional dense vector space, supporting clustering, semantic search, and similarity measurement tasks for multilingual texts

Model Features

Multilingual Support
Supports semantic similarity calculation for 15 languages, including Arabic, Chinese, English, etc.
High-Quality Fine-tuning
Fine-tuned using the STSb dataset extended to 15 languages, ensuring cross-language performance
Semantic Understanding
Capable of capturing deep sentence semantics, suitable for complex semantic matching scenarios

Model Capabilities

Sentence Embedding Generation
Cross-Language Semantic Search
Text Clustering Analysis
Semantic Similarity Calculation

Use Cases

Information Retrieval
Multilingual Document Search
Building a semantic search engine that supports multiple languages
Improves retrieval accuracy for non-English documents
Content Analysis
Cross-Language Content Deduplication
Identifying similar content expressed in different languages
Reduces content redundancy on multilingual platforms
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase