Reranker Models — Prediction Guard

This page provides information on the Reranker models that are available in the Prediction Guard API. These models are designed for semantically ranking text, and are used in the /rerank endpoint.

Models

Model Name	Type	Use Case	Context Length	More Info
bge-reranker-v2-m3	Reranker	Used for semantically ranking queries	512	link

Model Descriptions

bge-reranker-v2-m3

BAAI/bge-reranker-v2-m3 is a lightweight, multilingual reranker model designed for efficient and accurate text retrieval tasks.

Type: Reranker
Use Case: Used for Semantically Ranking Documents

https://huggingface.co/BAAI/bge-reranker-v2-m3

Unlike embedding models, rerankers take a query and document (or passage) as input and directly output a similarity score. The output relevance score can be converted to a float value in the range [0,1] using a sigmoid function.

Key Features: • Multilingual Support: Excels across multiple languages with strong cross-lingual capabilities. • Efficiency: Lightweight design ensures fast inference and easy deployment. • Versatility: Supports a range of use cases and scenarios.

This model offers an excellent balance between performance and deployment efficiency, making it a powerful choice for a wide range of text retrieval scenarios.