Datasets:
Tasks:
Text Generation
Modalities:
Text
Formats:
parquet
Sub-tasks:
language-modeling
Languages:
code
Size:
100K - 1M
License:
| annotations_creators: [] | |
| language: | |
| - code | |
| license: | |
| - other | |
| multilinguality: | |
| - monolingual | |
| size_categories: | |
| - unknown | |
| task_categories: | |
| - text-generation | |
| task_ids: | |
| - language-modeling | |
| pretty_name: github-jupyter-text-code-pairs | |
| This is a parsed version of [github-jupyter-parsed](https://huggingface.co/datasets/codeparrot/github-jupyter-parsed), with markdown and code pairs. We provide the preprocessing script in [preprocessing.py](https://huggingface.co/datasets/codeparrot/github-jupyter-parsed-v2/blob/main/preprocessing.py). The data is deduplicated and consists of 451662 examples. | |
| For similar datasets with text and Python code, there is [CoNaLa](https://huggingface.co/datasets/neulab/conala) benchmark from StackOverflow, with some samples curated by annotators. |