harlow
Adaptive surrogate modelling.
f_target(x) ~ f_surrogate(x)
Harlow is an Adaptive Surrogate Modeling package, in Python. The package offers a wide range of GPU-trainable Surrogate Models for single input-output and multi input-output pairs. Additionaly a series of Adaptive Samplers is implemented that work with multivariate data providing real-time web-logging. The package offers an intergration and benchmark test suite with numerous test functions and a real case study along with visualization functionality.
DISCLAIMER: This repository is in development. There's no guarantee in terms of code quality or output.
On using the repository
- Install dependencies and the code from this repo:
pip install -e .
- Install with additional dependencies for building documentation:
pip install -e .[docs]
- To build the documentation locally:
sphinx-build -b html docs/source docs/build
-
To view the documentation open
docs/build/index.html -
All code within this repository is expected to be run with the working directory set as the root directory of the repository.