P

Polish Reranker Base Ranknet

Developed by sdadas
Polish text ranking model trained with RankNet loss function, suitable for information retrieval tasks
Downloads 332
Release Time : 2/3/2024

Model Overview

This is a Polish text ranking model trained using the RankNet loss function, primarily designed to improve query-document relevance ranking in information retrieval systems.

Model Features

RankNet training method
Uses RankNet loss function based on relative ranking of query-document pairs rather than processing each document independently
Large-scale training data
Training set contains 1.4 million queries and 10 million documents covering multiple domains
Knowledge distillation
Utilizes knowledge distillation training with large MT5-XXL teacher model

Model Capabilities

Query-document relevance scoring
Search results re-ranking
Multi-document relevance comparison

Use Cases

Information retrieval systems
Search engine results optimization
Re-rank documents returned by search engines to improve ranking of relevant documents
QA systems
Select the most relevant answer from candidate responses
Medical field
Medical QA ranking
Rank relevance of answers in medical QA systems
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase