databricks/lilac
A dataset exploration and curation tool for improving LLM training and fine-tuning data quality.
Not currently ranked — collecting fresh signals.
star history
Lilac provides visualization and quality control capabilities for LLM training datasets. It helps teams explore, quantify, and improve pre-training and fine-tuning data through a Python API and on-device UI. The tool integrates open-source LLMs to run computations locally, supporting data filtering, clustering, and semantic search across large datasets.
Frequently asked
- What is databricks/lilac?
- A dataset exploration and curation tool for improving LLM training and fine-tuning data quality.
- Is lilac open source?
- Yes — databricks/lilac is open source, released under the Apache-2.0 license.
- What language is lilac written in?
- databricks/lilac is primarily written in Python.
- How popular is lilac?
- databricks/lilac has 1.1k stars on GitHub.
- Where can I find lilac?
- databricks/lilac is on GitHub at https://github.com/databricks/lilac.