Skip to content

Dataset · MoreThought

SmolPyEdu clean Python script corpus

MoreThought/SmolPyEdu

Large array of cleaned short Python scripts for language model pre-training.

What's inside

A corpus containing between 1M and 10M English examples, created based on the Avelina/python-edu-cleaned dataset. Only working scripts were filtered from the source data, all comments were removed, and short clean pieces of code were left.

What you can do with it

  • Pre-training general-purpose language models and specialized SLMs/TLMs for code tasks
  • Fine-tuning models for Python program code generation
Tasks
Text generation, Question answering
Languages
English
Size
1M–10M rows
Downloads this month
782
Likes
4

How to load it

from datasets import load_dataset

ds = load_dataset("MoreThought/SmolPyEdu")