B

Bi Electra Ms Marco German Uncased

Developed by svalabs
Dual-encoder fine-tuned from German ELECTRA model for German passage retrieval tasks
Downloads 17
Release Time : 3/2/2022

Model Overview

This model is a dual-encoder based on the ELECTRA architecture, specifically optimized for German passage retrieval tasks, capable of efficient semantic search and passage matching.

Model Features

German optimization
Specifically optimized for German text, suitable for German semantic search tasks
Efficient retrieval
Dual-encoder architecture provides efficient passage retrieval capabilities
Superior performance
Outperforms traditional BM25 methods on the GermanDPR test set

Model Capabilities

German semantic search
Passage retrieval
Text similarity calculation

Use Cases

Information retrieval
Document retrieval system
Build a German document retrieval system to quickly find relevant documents
Achieves Recall@10 of 0.8205 on the GermanDPR test set
Question answering system
Serves as the retrieval component of a QA system to find relevant passages to questions
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase