[![CI][ci:b]][ci]
[![License BSD3][license:b]][license]
![Python3.12][python:b]
[![pypi][pypi:b]][pypi]
[![codecov][codecov:b]][codecov]
[](https://zenodo.org/badge/latestdoi/227004259)
[ci]: https://github.com/dirichletcal/dirichlet_python/actions/workflows/ci.yml
[ci:b]: https://github.com/dirichletcal/dirichlet_python/workflows/CI/badge.svg
[license]: https://github.com/dirichletcal/dirichlet_python/blob/master/LICENSE.txt
[license:b]: https://img.shields.io/github/license/dirichletcal/dirichlet_python.svg
[python:b]: https://img.shields.io/badge/python-3.12-blue
[pypi]: https://badge.fury.io/py/dirichletcal
[pypi:b]: https://badge.fury.io/py/dirichletcal.svg
[codecov]: https://codecov.io/gh/dirichletcal/dirichlet_python
[codecov:b]: https://codecov.io/gh/dirichletcal/dirichlet_python/branch/master/graph/badge.svg
# Dirichlet Calibration Python implementation
This is a Python implementation of the Dirichlet Calibration presented in
__Beyond temperature scaling: Obtaining well-calibrated multi-class probabilities
with Dirichlet calibration__ at NeurIPS 2019. The original version used Python
3.8 and reached version `0.4.2`. The code started using Python 3.12 from
version `0.5.0`, you can see the other version in the GitHub history, tags, or
in Pypi.
# Installation
```
# Clone the repository
git clone git@github.com:dirichletcal/dirichlet_python.git
# Go into the folder
cd dirichlet_python
# Create a new virtual environment with Python3
python3.12 -m venv venv
# Load the generated virtual environment
source venv/bin/activate
# Upgrade pip
pip install --upgrade pip
# Install all the dependencies
pip install -r requirements.txt
pip install --upgrade jaxlib
```
# Unittest
```
python -m unittest discover dirichletcal
```
# Cite
If you use this code in a publication please cite the following paper
```
@inproceedings{kull2019dircal,
title={Beyond temperature scaling: Obtaining well-calibrated multi-class probabilities with Dirichlet calibration},
author={Kull, Meelis and Nieto, Miquel Perello and K{\"a}ngsepp, Markus and Silva Filho, Telmo and Song, Hao and Flach, Peter},
booktitle={Advances in Neural Information Processing Systems},
pages={12295--12305},
year={2019}
}
```
# Examples
You can find some examples on how to use this package in the folder
[examples](examples)
# Pypi
To push a new version to Pypi first build the package
```
python3.12 setup.py sdist
```
And then upload to Pypi with twine
```
twine upload dist/*
```
It may require user and password if these are not set in your home directory a
file __.pypirc__
```
[pypi]
username = __token__
password = pypi-yourtoken
```
Raw data
{
"_id": null,
"home_page": "https://github.com/dirichletcal/dirichlet_python",
"name": "dirichletcal",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.12",
"maintainer_email": null,
"keywords": "classifier, calibration, dirichlet, multiclass, probability",
"author": "Miquel Perello Nieto and Hao Song",
"author_email": "perello.nieto@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/5a/a2/5006b6d405fb2b1beb81515c315454eeef0255cff49fc82e0dadeb09b377/dirichletcal-0.5.2.tar.gz",
"platform": null,
"description": "[![CI][ci:b]][ci]\n[![License BSD3][license:b]][license]\n![Python3.12][python:b]\n[![pypi][pypi:b]][pypi]\n[![codecov][codecov:b]][codecov]\n[](https://zenodo.org/badge/latestdoi/227004259)\n\n[ci]: https://github.com/dirichletcal/dirichlet_python/actions/workflows/ci.yml\n[ci:b]: https://github.com/dirichletcal/dirichlet_python/workflows/CI/badge.svg\n[license]: https://github.com/dirichletcal/dirichlet_python/blob/master/LICENSE.txt\n[license:b]: https://img.shields.io/github/license/dirichletcal/dirichlet_python.svg\n[python:b]: https://img.shields.io/badge/python-3.12-blue\n[pypi]: https://badge.fury.io/py/dirichletcal\n[pypi:b]: https://badge.fury.io/py/dirichletcal.svg\n[codecov]: https://codecov.io/gh/dirichletcal/dirichlet_python\n[codecov:b]: https://codecov.io/gh/dirichletcal/dirichlet_python/branch/master/graph/badge.svg\n\n# Dirichlet Calibration Python implementation\n\nThis is a Python implementation of the Dirichlet Calibration presented in\n__Beyond temperature scaling: Obtaining well-calibrated multi-class probabilities\nwith Dirichlet calibration__ at NeurIPS 2019. The original version used Python\n3.8 and reached version `0.4.2`. The code started using Python 3.12 from\nversion `0.5.0`, you can see the other version in the GitHub history, tags, or\nin Pypi.\n\n# Installation\n\n```\n# Clone the repository\ngit clone git@github.com:dirichletcal/dirichlet_python.git\n# Go into the folder\ncd dirichlet_python\n# Create a new virtual environment with Python3\npython3.12 -m venv venv\n# Load the generated virtual environment\nsource venv/bin/activate\n# Upgrade pip\npip install --upgrade pip\n# Install all the dependencies\npip install -r requirements.txt\npip install --upgrade jaxlib\n```\n\n# Unittest\n\n```\npython -m unittest discover dirichletcal\n```\n\n\n# Cite\n\nIf you use this code in a publication please cite the following paper\n\n\n```\n@inproceedings{kull2019dircal,\n title={Beyond temperature scaling: Obtaining well-calibrated multi-class probabilities with Dirichlet calibration},\n author={Kull, Meelis and Nieto, Miquel Perello and K{\\\"a}ngsepp, Markus and Silva Filho, Telmo and Song, Hao and Flach, Peter},\n booktitle={Advances in Neural Information Processing Systems},\n pages={12295--12305},\n year={2019}\n}\n```\n\n# Examples\n\nYou can find some examples on how to use this package in the folder\n[examples](examples)\n\n# Pypi\n\nTo push a new version to Pypi first build the package\n\n```\npython3.12 setup.py sdist\n```\n\nAnd then upload to Pypi with twine\n\n```\ntwine upload dist/*\n```\n\nIt may require user and password if these are not set in your home directory a\nfile __.pypirc__\n\n```\n[pypi]\nusername = __token__\npassword = pypi-yourtoken\n```\n",
"bugtrack_url": null,
"license": null,
"summary": "Python code for Dirichlet calibration",
"version": "0.5.2",
"project_urls": {
"Download": "https://github.com/dirichletcal/dirichlet_python/archive/0.5.2.tar.gz",
"Homepage": "https://github.com/dirichletcal/dirichlet_python"
},
"split_keywords": [
"classifier",
" calibration",
" dirichlet",
" multiclass",
" probability"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "5aa25006b6d405fb2b1beb81515c315454eeef0255cff49fc82e0dadeb09b377",
"md5": "02c0d865324daa22ac7166c90fdbba5b",
"sha256": "48d69089c127121543bba7abef6327a10e7c6008cec9823b01da5aaa94b5eeb7"
},
"downloads": -1,
"filename": "dirichletcal-0.5.2.tar.gz",
"has_sig": false,
"md5_digest": "02c0d865324daa22ac7166c90fdbba5b",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.12",
"size": 12677,
"upload_time": "2024-08-05T18:52:13",
"upload_time_iso_8601": "2024-08-05T18:52:13.475317Z",
"url": "https://files.pythonhosted.org/packages/5a/a2/5006b6d405fb2b1beb81515c315454eeef0255cff49fc82e0dadeb09b377/dirichletcal-0.5.2.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-08-05 18:52:13",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "dirichletcal",
"github_project": "dirichlet_python",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"requirements": [
{
"name": "numpy",
"specs": [
[
"==",
"2.0.1"
]
]
},
{
"name": "scipy",
"specs": [
[
"==",
"1.14.0"
]
]
},
{
"name": "scikit-learn",
"specs": [
[
"==",
"1.5.1"
]
]
},
{
"name": "jax",
"specs": [
[
"==",
"0.4.31"
]
]
},
{
"name": "jaxlib",
"specs": [
[
"==",
"0.4.31"
]
]
}
],
"lcname": "dirichletcal"
}
JSON
