Metadata-Version: 2.1
Name: hgdl
Version: 2.0.3
Summary: HGDL Optimization
Home-page: https://github.com//hgdl
Author: David Perryman, Marcus Noack
Author-email: MarcusNoack@lbl.gov
License: BSD3
Classifier: Development Status :: 2 - Pre-Alpha
Classifier: Natural Language :: English
Classifier: Programming Language :: Python :: 3
Requires-Python: >=3.6
Description-Content-Type: text/markdown
Provides-Extra: tests
Provides-Extra: docs
License-File: LICENSE
License-File: AUTHORS.rst

# HGDL

[![PyPI](https://img.shields.io/pypi/v/HGDL)](https://pypi.org/project/hgdl/)
[![Documentation Status](https://readthedocs.org/projects/gpcam/badge/?version=latest)](https://gpcam.readthedocs.io/en/latest/?badge=latest)
[![HGDL CI](https://github.com/lbl-camera/HGDL/actions/workflows/HGDL-CI.yml/badge.svg)](https://github.com/lbl-camera/fvGP/actions/workflows/HGDL-CI.yml)
[![Codecov](https://img.shields.io/codecov/c/github/lbl-camera/HGDL)](https://app.codecov.io/gh/lbl-camera/HGDL)
[![PyPI - License](https://img.shields.io/pypi/l/HGDL)](https://pypi.org/project/hgdl/)
[<img src="https://img.shields.io/badge/slack-@gpCAM-purple.svg?logo=slack">](https://gpCAM.slack.com/)
[![DOI](https://zenodo.org/badge/434769975.svg)](https://zenodo.org/badge/latestdoi/434769975)


HGDL is an API for HPC distributed constrained function optimization.
At the core, the algorithm uses local and global optimization
and bump-function-based deflation to provide a growing list of unique optima of a differentiable function.
This tackles the common problem of non-uniquness of optimization problems, especially in machine learning.

## Usage

The following demonstrates a simple usage of the HGDL API.

```python
import numpy as np
from hgdl.hgdl import HGDL as hgdl
from hgdl.support_functions import *
import dask.distributed as distributed

bounds = np.array([[-500,500],[-500,500]])
#dask_client = distributed.Client("10.0.0.184:8786")
a = hgdl(schwefel, schwefel_gradient, bounds,
        hess = schwefel_hessian,
        global_optimizer = "genetic",
        local_optimizer = "dNewton",
        number_of_optima = 30000,
        args = (1,1), radius = None, num_epochs = 100)

x0 = np.random.uniform(low = bounds[:, 0], high = bounds[:,1],size = (20,2))
a.optimize(x0 = x0)

###the thread is now released, but the work continues in the background

a.get_latest() ##prints the current result whenever queried

a.kill_client() ##stops the execution and returns the result
```


## Credits

Main Developers: Marcus Noack ([MarcusNoack@lbl.gov](mailto:MarcusNoack@lbl.gov)) and David Perryman.
Several people from across the DOE national labs have given insights
that led to the code in its current form.
See [AUTHORS](AUTHORS.rst) for more details on that.
HGDL is based on the [HGDN](https://www.sciencedirect.com/science/article/pii/S037704271730225X) algorithm by Noack and Funke.

