Skip to content

Dataset · jinaai

120 thousand Python programming exercises

jinaai/code_exercises

A synthetic set of Python programming tasks in HumanEval format from ChatGPT 3.5 for distilling coding skills.

What's inside

The dataset contains about 120 thousand Python programming exercises (roughly 120 million tokens) in English. Examples were generated using ChatGPT 3.5 and formatted according to the HumanEval benchmark standard.

What you can do with it

  • Distilling coding skills into compact language models
  • SFT fine-tuning of LLMs for solving algorithmic Python tasks
Tasks
Text generation
Languages
English
Size
1M–10M rows
Downloads this month
1.4K
Likes
42

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("jinaai/code_exercises")