---
license: apache-2.0
task_categories:
- information-retrieval
- text-retrieval
tags:
- beir
- fever
- information-retrieval
- retrieval
- search
configs:
- config_name: corpus
  data_files:
  - split: train
    path: corpus/train-*
- config_name: queries
  data_files:
  - split: train
    path: queries/train-*
dataset_info:
- config_name: corpus
  features:
  - name: metadata
    dtype: 'null'
  - name: title
    dtype: string
  - name: _id
    dtype: string
  - name: text
    dtype: string
  splits:
  - name: train
    num_bytes: 3095105800
    num_examples: 5416568
  download_size: 1953809592
  dataset_size: 3095105800
- config_name: queries
  features:
  - name: metadata
    dtype: string
  - name: text
    dtype: string
  - name: _id
    dtype: string
  splits:
  - name: train
    num_bytes: 28132663
    num_examples: 123142
  download_size: 11838211
  dataset_size: 28132663
---

# BEIR FEVER Dataset (Migrated)

This is a migrated version of BeIR/fever that is compatible with datasets library 4.0.0+.

## Dataset Description

This dataset contains the fever dataset from the BEIR benchmark, converted from the old script-based format to Parquet format.

## Dataset Structure

### Queries

- **Split 'queries'**: 123,142 examples
  - Features: ['_id', 'text', 'metadata']

- **Total examples**: 123,142


### Corpus

- **Split 'corpus'**: 5,416,568 examples
  - Features: ['_id', 'title', 'text', 'metadata']

- **Total examples**: 5,416,568


## Usage

```python
from datasets import load_dataset

# Load queries (split: queries)
queries = load_dataset("Hyukkyu/beir-fever", "queries", split="queries")

# Load corpus (split: corpus)
corpus = load_dataset("Hyukkyu/beir-fever", "corpus", split="corpus")
```


## Available Splits

### Queries
- `queries`: 123,142 examples

### Corpus
- `corpus`: 5,416,568 examples

## Original Dataset

This dataset is migrated from: BeIR/fever

## Citation

If you use this dataset, please cite the original BEIR paper:

```bibtex
@article{thakur2021beir,
  title={BEIR: A Heterogeneous Benchmark for Zero-shot Evaluation of Information Retrieval Models},
  author={Thakur, Nandan and Reimers, Nils and Ruckle, Andreas and Srivastava, Abhishek and Gurevych, Iryna},
  journal={arXiv preprint arXiv:2104.08663},
  year={2021}
}
```