ance msmarco passage

Introduction

ANCE-MSMARCO-PASSAGE is an information retrieval model developed by Castorini. It is designed for effective and reproducible first-stage retrieval in a multi-stage ranking architecture. The model is based on the RoBERTa architecture and primarily handles English language tasks.

Architecture

The model leverages Pyserini, a toolkit designed for reproducible information retrieval research using both sparse and dense representations. It is a part of a multi-stage ranking system and aims to improve the first-stage retrieval process.

Training

The training details, including data and procedures, are not fully provided. The associated research paper mentions the use of the MS MARCO passage corpus, which comprises 8.8 million passages. The default ranking model employed is BM25, a bag-of-words ranking method.

Guide: Running Locally

1. Install Required Libraries: Ensure you have the transformers library installed.

   pip install transformers
   

2. Load the Model: Use the following Python code to load the tokenizer and model.

   from transformers import AutoTokenizer, AnceEncoder
   
   tokenizer = AutoTokenizer.from_pretrained("castorini/ance-msmarco-passage")
   model = AnceEncoder.from_pretrained("castorini/ance-msmarco-passage")
   

3. Cloud GPUs: For enhanced performance, consider using cloud-based GPUs like those offered by AWS, Google Cloud, or Azure.

License

The specific license details for ANCE-MSMARCO-PASSAGE are not provided in the documentation. Users should verify the license terms before using the model in their applications.