Datasets:

Languages:
Chinese
ArXiv:
License:
T2Ranking / README.md
qian's picture
update readme
5504a5c
|
raw
history blame
1.93 kB
---
license: apache-2.0
task_categories:
- text-retrieval
language:
- zh
size_categories:
- 1M<n<10M
---
# T2Ranking
T2Ranking is a large-scale Chinese benchmark for passage ranking. The details about T2Ranking are elaborated in [this paper](https://github.com/THUIR/T2Ranking).
The download links and data formats are presented in the following table.
| Description | Filename | File size | Num Records | Format |
|-------------------------------------------------------|-------------------------------------------------------------------------------------------------------------------------|----------:|-----------------------------------:|----------------------------------------------------------------|
| Collection | collection.tsv | 3.5 GB | 2,303,643 | tsv: pid, passage |
| Queries Train | queries.train.tsv | 9.8 MB | 258,042 | tsv: qid, query |
| Qrels Train | qrels.train.tsv | 29 MB | 1,613,421 | TREC qrels format |
<!-- | Qrels Dev | qrels.dev.tsv | 13 MB | 781,949 | TREC qrels format |
| Queries Dev | queries.dev.tsv | 1.9 MB | 49,662 | tsv: qid, query |
| Qrels Retrieval | retrieval_qrels.tsv | 14 MB | 980,060 | tsv:qid, pid | -->