Datasets:
Tasks:
Text Generation
Modalities:
Text
Formats:
parquet
Languages:
English
Size:
10K - 100K
Tags:
code
License:
metadata
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 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
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
})