Bojaxns is:
- a Bayesian Optimisation package for easily performing advanced non-myopic Bayesian optimisation.
- using JAXNS under the hood to marginalise over multiple models.
- using multi-step lookahead to plan out your next step.
- available for academic use and non-commercial use (without permission) read the license.
For examples, check out the documentation (still in progress).
Notes:
- Bojaxns requires >= Python 3.9.
- It is always highly recommended to use a unique virtual environment for each project.
To use
miniconda
, have it installed, and run
# To create a new env, if necessary
conda create -n bojaxns_py python=3.11
conda activate bojaxns_py
Install directly from PyPi,
pip install bojaxns
Clone repo git clone https://www.github.com/JoshuaAlbert/bojaxns.git
, and install:
cd bojaxns
pip install -r requirements.txt
pip install -r requirements-tests.txt
pip install .
15 Jan, 2024 -- Bojaxns 1.1.0/1 released. Bumped to jaxns 2.4.3/4. 20 July, 2023 -- Bojaxns 1.0.0 released