---
dataset_info:
  features:
  - name: content
    dtype: string
  splits:
  - name: train
    num_bytes: 10335479660.938484
    num_examples: 2073959
  - name: validation
    num_bytes: 2583871161.0981674
    num_examples: 518490
  download_size: 4705432064
  dataset_size: 12919350822.036652
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
---

# Small Python Stack

This dataset contains a random sample of 20% of the Python code available in [the original Stack](https://huggingface.co/datasets/bigcode/the-stack).