ms-marco-MiniLM-L4-v2
| Property | Value |
|---|---|
| Model Type | Cross-Encoder |
| Author | cross-encoder |
| Performance (NDCG@10) | 73.04 |
| Processing Speed | 2500 docs/sec |
| Model URL | HuggingFace |
What is ms-marco-MiniLM-L4-v2?
ms-marco-MiniLM-L4-v2 is a specialized cross-encoder model designed for passage ranking tasks, particularly optimized for the MS Marco dataset. It represents the second version of the 4-layer MiniLM architecture, striking an excellent balance between performance and efficiency.
Implementation Details
The model implements a cross-encoder architecture specifically trained on the MS Marco Passage Ranking task. It can be easily integrated using either SentenceTransformers or the Transformers library, making it versatile for different implementation needs.
- Built on MiniLM architecture with 4 layers
- Achieves 37.70 MRR@10 on MS Marco Dev set
- Processes 2500 documents per second on V100 GPU
- Supports both query-passage pair scoring and ranking
Core Capabilities
- Efficient passage ranking and retrieval
- Query-passage relevance scoring
- Information retrieval enhancement
- Compatible with major deep learning frameworks
Frequently Asked Questions
Q: What makes this model unique?
This model offers an optimal balance between performance and speed, achieving a strong NDCG@10 score of 73.04 while maintaining relatively high processing speed. It's particularly effective for production environments where both accuracy and efficiency are crucial.
Q: What are the recommended use cases?
The model is ideal for information retrieval systems, particularly when implementing a retrieve & re-rank architecture. It's well-suited for applications requiring passage ranking, document search, and query-passage relevance scoring.
