metadata
license: mit
dataset_info:
features:
- name: document_id
dtype: int64
- name: text
dtype: string
splits:
- name: train
num_bytes: 4691943775
num_examples: 8841823
download_size: 2121157232
dataset_size: 4691943775
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
This dataset is used with akhooli/ar_mmarco_queries (see card) and akhooli/ar_mmarco_scores.