Skip to content
Willump is a compiler for model serving.
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.


Willump is under active development! Please report any bugs or unusual behavior to

Willump is a compiler for model serving. It speeds up model serving pipelines written in Python using Numpy and Pandas. To speed up a function using Willump, simply wrap it in the Willump decorator:

import willump.evaluation.willump_executor

def make_me_faster(...):

To install Willump, first install the llvm-st branch of our Weld fork, weld-willump. Its repository and installation instructions are available here.

Next, install Python, Mypy, NumPy, scikit-learn, and Pandas. Willump requires Python version 3.6 or later.

Next, define the WILLUMP_HOME environment variable to point to the Willump root directory (this one) and include the Willump root directory on your PYTHONPATH. Willump should now work!

To confirm Willump works, run the Willump unit tests:

python3 -m unittest discover -s tests -p *.py

To run our Willump benchmarks, first train the benchmark models:

python3 tests/test_scripts/
python3 tests/test_scripts/

Then run the benchmarks themselves using the trained models:

python3 tests/test_scripts/
python3 tests/test_scripts/

Both benchmarks should run dramatically faster than they would without Willump.

You can’t perform that action at this time.