Skip to content

llm-datasets

Curated datasets and tools for LLM finetuning

Visit official site

What llm-datasets is

The project is described as a GitHub collection of curated datasets and tools for large language model finetuning, created and maintained by mlabonne. It groups general mixes, math and code data with tools for evaluating and expanding them.

What you can do with it

  • Pick a dataset and clone or download it locally
  • Screen and refine data with the quality evaluation tools
  • Generate extra data to fill gaps in a collection

Application scenarios

  • Researchers train models on math datasets
  • Developers train models on code datasets
  • Teams train models on a general mixed dataset

Pros & cons

✓ What we like

  • Data generation tools help fill gaps in a dataset
  • Exploration tools reveal patterns in the data

! What to watch out for

  • Datasets are picked and downloaded by the user
  • Quality screening runs through the provided tools

FAQ

Who maintains it?

The material states the collection is created and maintained by mlabonne.

What tools are included?

Data quality evaluation, data generation and data exploration tools are listed.

Last reviewed: 2026-11-15

More AI coding tools tools

View all →

How we review