Skip to content

Dataset · bigcode

Small code slice across 30 languages

bigcode/the-stack-smol

A subset of The Stack: 10K random code examples for each of 30 languages, 2.6 GB.

What's inside

The dataset contains code texts—a small slice (~0.1%) of the original The Stack. For each of the 30 programming languages, 10,000 random examples were taken; the total volume is about 2.6 GB. Data is presented in tabular and text formats.

What you can do with it

  • Pre-training or fine-tuning small code models
  • Testing training pipelines on code without the full The Stack
  • Quick experiments with tokenizers for 30 languages
Tasks
Text generation
Languages
Code
Size
100K–1M rows
Downloads this month
31.5K
Likes
107
Access
on request on Hugging Face

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("bigcode/the-stack-smol")