siboehm/lleaves

Compiler for LightGBM gradient-boosted trees, based on LLVM. Speeds up prediction by ≥10x.

57
/ 100
Established

This project helps data scientists and machine learning engineers significantly speed up predictions from their LightGBM models. You provide a trained LightGBM model, and it outputs an optimized version that makes predictions much faster, whether for individual requests or large batches of data. It's designed for those who need low-latency or high-throughput model inference.

463 stars. Available on PyPI.

Use this if you have LightGBM models in production and need to reduce their prediction time by a factor of 10 or more.

Not ideal if your application's performance is not bottlenecked by LightGBM model inference speed or if you are not using LightGBM models.

machine-learning-inference real-time-prediction model-deployment data-science-optimization
Maintenance 6 / 25
Adoption 10 / 25
Maturity 25 / 25
Community 16 / 25

How are scores calculated?

Stars

463

Forks

43

Language

Python

License

MIT

Last pushed

Jan 01, 2026

Commits (30d)

0

Dependencies

2

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/siboehm/lleaves"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.