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 | |
}) | |
```` |