another machine learning library
Project description
ubergauss
another machine learning library
Install
pip install ubergauss
conda install -c conda-forge ubergauss
Example
kneepoint detection
import ubergauss as ug
# kneepoint gaussians:
>>>ug.between_gaussians([.1,.1,.2,.2,.4,.7])
4
# max dist to diagonal
>>>ug.diag_maxdist([.1,.1,.2,.2,.4,.7])
3
hyperparameter optimization
embedding of distance matrix
collection of small tools
sigma boxplot
optimization
blackboxBORE is the one that works best
def test_ga():
ut.nuke()
def example_function(data, x=None, y=None, some_boolean=None,**kwargs):
score_from_x = - (x - 0.5)**2 # Max at x=0.5
score_from_y = - (y - 10)**2 / 100.0 # Max at y=10
score_from_bool = .1*some_boolean
score_noise = np.random.normal(0, .1)
return score_noise + score_from_x + score_from_y + score_from_bool
example_space = """
x 0.0 1.0
y 1 20 1
some_boolean [1, 0]
"""
o = gatype.nutype(example_space,
example_function,
data=[[0]],
numsample=16)
[o.opti() for _ in range(5)]
o.print()
# o.print_more()
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
ubergauss-0.0.141.tar.gz
(38.1 kB
view details)
File details
Details for the file ubergauss-0.0.141.tar.gz.
File metadata
- Download URL: ubergauss-0.0.141.tar.gz
- Upload date:
- Size: 38.1 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/4.0.2 CPython/3.11.5
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
474022864617ca289b91626befbe91b1664cde1689852e43d265360f1619099e
|
|
| MD5 |
413380a553cbb027cc419181e3f8685b
|
|
| BLAKE2b-256 |
85b2bfcc7e07fb1844b63534935e74aaf052ed1db89b30a48b07ee6c5e949f68
|