Datasets:
Tasks:
Text Generation
Modalities:
Text
Formats:
parquet
Languages:
English
Size:
10K - 100K
Tags:
code
License:
| license: apache-2.0 | |
| task_categories: | |
| - text-generation | |
| tags: | |
| - code | |
| size_categories: | |
| - 10K<n<100K | |
| language: | |
| - en | |
| # Dataset description | |
| This dataset consists of sequences of Python code followed by a a docstring explaining its function. It was constructed by concatenating code and text pairs | |
| from this [dataset](https://huggingface.co/datasets/codeparrot/github-jupyter-text-code-pairs) that were originally code and markdown cells in Jupyter Notebooks. | |
| The content of each example the following: | |
| ```` | |
| [CODE] | |
| """ | |
| Explanation: [TEXT] | |
| End of explanation | |
| """ | |
| [CODE] | |
| """ | |
| Explanation: [TEXT] | |
| End of explanation | |
| """ | |
| ... | |
| ```` | |
| # How to use it | |
| ```python | |
| from datasets import load_dataset | |
| ds = load_dataset("codeparrot/github-jupyter-code-to-text", split="train") | |
| ```` | |
| ```` | |
| Dataset({ | |
| features: ['repo_name', 'path', 'license', 'content'], | |
| num_rows: 47452 | |
| }) | |
| ```` |