Skip to content

Dataset · BeIR

English Text Retrieval Benchmark (BEIR)

BeIR/msmarco

A subset of MS MARCO in the BEIR benchmark format of millions of English documents and search queries.

What's inside

The dataset is a subset of MS MARCO from the famous BEIR information retrieval benchmark. Inside, there are between 1 and 10 million English-language records, divided into a document corpus (_id, title, text fields) and a set of search queries.

What you can do with it

  • Evaluating the quality of text-retrieval models and embeddings
  • Testing zero-shot classification and re-ranking systems in English
Tasks
Zero shot classification, Text retrieval
Languages
English
Size
1M–10M rows
Downloads this month
5.3K
Likes
16

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("BeIR/msmarco")