Datasets:
| license: apache-2.0 | |
| language: | |
| - zh | |
| task_categories: | |
| - text-generation | |
| tags: | |
| - dialogue | |
| - chinese | |
| - knowledge-grounded | |
| pretty_name: DuConv (mirror) | |
| size_categories: | |
| - 10K<n<100K | |
| # DuConv (mirror for HySI Lab) | |
| Redistributed mirror of **DuConv** (Baidu PaddleNLP) for internal / cluster use. | |
| ## Source | |
| - Loader script repo: [`PaddlePaddle/duconv`](https://huggingface.co/datasets/PaddlePaddle/duconv) | |
| - Upstream archive: `https://bj.bcebos.com/paddlenlp/datasets/DuConv.zip` | |
| ## Files | |
| | file | split | | |
| |------|-------| | |
| | `train.txt` | train | | |
| | `dev.txt` | dev | | |
| | `test_1.txt` | test_1 | | |
| | `test_2.txt` | test_2 | | |
| ## License | |
| Apache-2.0 (per upstream). Please cite the original DuConv paper if you use this data. | |