| license: mit | |
| language: | |
| - zh | |
| task_categories: | |
| - text-generation | |
| tags: | |
| - dialogue | |
| - chinese | |
| pretty_name: LCCC Large (mirror) | |
| size_categories: | |
| - 10M<n<100M | |
| # LCCC Large (mirror for HySI Lab) | |
| Redistributed mirror of the **LCCC large** Chinese dialogue corpus for internal / cluster use. | |
| ## Source | |
| - Original paper dataset: THU-CoAI LCCC | |
| - Loader script repo: [`thu-coai/lccc`](https://huggingface.co/datasets/thu-coai/lccc) | |
| - Data files originally hosted at [`silver/lccc`](https://huggingface.co/datasets/silver/lccc) | |
| ## Files | |
| | file | description | | |
| |------|-------------| | |
| | `lccc_large.jsonl` | LCCC large split (~12M dialogues) | | |
| | `lccc_base_train.jsonl` | LCCC base train | | |
| | `lccc_base_valid.jsonl` | LCCC base valid | | |
| | `lccc_base_test.jsonl` | LCCC base test | | |
| Each line is a JSON list of utterance strings. | |
| ## License | |
| MIT (per upstream). Please cite the original LCCC paper if you use this data. | |