README.md

metadata

license: apache-2.0
task_categories:
  - text-generation
language:
  - en
tags:
  - conversational-search
  - query-rewriting
  - cqr

Conversational Query Rewriting (CQR) Dataset

Unified dataset for training CQR models. Combines human-annotated rewrites from QReCC, CANARD, and Redis examples.

Schema

Field	Type	Description
context	list[str]	Previous conversation turns
question	str	Context-dependent question
rewrite	str	Self-contained rewrite
dataset	str	Source dataset
split	str	Original split
index	int	Original index

Splits

Split	Count
train	83702
validation	2196
test	19824

Sources

QReCC (apple/ml-qrecc) - Apache 2.0

@inproceedings{anantha-etal-2021-qrecc,
    title = "Open-Domain Question Answering Goes Conversational via Question Rewriting",
    author = "Anantha, Raviteja and Vakulenko, Svitlana and Tu, Zhucheng and Longpre, Shayne and Pulman, Stephen and Chappidi, Srinivas",
    booktitle = "NAACL",
    year = "2021",
    url = "https://aclanthology.org/2021.naacl-main.44",
}

CANARD (uva-irlab/canard_quretec) - CC BY-SA 4.0

@inproceedings{elgohary-etal-2019-canard,
    title = "Can You Unpack That? Learning to Rewrite Questions-in-Context",
    author = "Elgohary, Ahmed and Peskov, Denis and Boyd-Graber, Jordan",
    booktitle = "EMNLP-IJCNLP",
    year = "2019",
    url = "https://aclanthology.org/D19-1605",
}

Redis Examples - Hand-crafted examples by Redis - Apache 2.0

Citation

If you use this dataset, please cite this Hugging Face entry and the source datasets above.

@misc{zhechev2025cqr,
    title = "Conversational Query Rewriting Dataset",
    author = "Zhechev, Iliya and Ralev, Rado and Rajamohan, Srijith and Redis",
    year = "2025",
    publisher = "Hugging Face",
    url = "https://huggingface.co/datasets/redis/langcache-cqr-v1",
}