Skip to content


Switch branches/tags

Latest commit

…is null in model file (#14)

* extract_pandas_traintime_categories: return empty list if pandas_categorical is null in model file

* Test: Prediction for df with empty categoricals

Co-authored-by: chenglin <>
Co-authored-by: Simon Boehm <>

Git stats


Failed to load latest commit information.

lleaves 🍃

CI Documentation Status Downloads

A LLVM-based compiler for LightGBM decision trees.

lleaves converts trained LightGBM models to optimized machine code, speeding-up prediction by ≥10x.


lgbm_model = lightgbm.Booster(model_file="NYC_taxi/model.txt")
%timeit lgbm_model.predict(df)
# 12.77s

llvm_model = lleaves.Model(model_file="NYC_taxi/model.txt")
%timeit llvm_model.predict(df)
# 0.90s 

Why lleaves?

  • Speed: Both low-latency single-row prediction and high-throughput batch-prediction.
  • Drop-in replacement: The interface of lleaves.Model is a subset of LightGBM.Booster.
  • Dependencies: llvmlite and numpy. LLVM comes statically linked.


conda install -c conda-forge lleaves or pip install lleaves (Linux and MacOS only).


Ran on a dedicated Intel i7-4770 Haswell, 4 cores. Stated runtime is the minimum over 20.000 runs.

Dataset: NYC-taxi

mostly numerical features.

batchsize 1 10 100
LightGBM 52.31μs 84.46μs 441.15μs
ONNX Runtime 11.00μs 36.74μs 190.87μs
Treelite 28.03μs 40.81μs 94.14μs
lleaves 9.61μs 14.06μs 31.88μs

Dataset: MTPL2

mix of categorical and numerical features.

batchsize 10,000 100,000 678,000
LightGBM 95.14ms 992.47ms 7034.65ms
ONNX Runtime 38.83ms 381.40ms 2849.42ms
Treelite 38.15ms 414.15ms 2854.10ms
lleaves 5.90ms 56.96ms 388.88ms

Advanced usage

To avoid any Python overhead during prediction you can link directly against the generated binary. See benchmarks/c_bench/ for an example of how to do this. The function signature can change between major versions.


conda env create
conda activate lleaves
pip install -e .
pre-commit install