Name | bayesian-models JSON |
Version |
0.1.0
JSON |
| download |
home_page | |
Summary | A package for building common bayesian models in pymc |
upload_time | 2023-04-05 09:14:06 |
maintainer | |
docs_url | None |
author | |
requires_python | |
license | |
keywords |
bayesian
models
statistics
pymc
|
VCS |
|
bugtrack_url |
|
requirements |
No requirements were recorded.
|
Travis-CI |
No Travis.
|
coveralls test coverage |
No coveralls.
|
# bayesian-models
`bayeian-models` is a small library build on top of `pymc` that
implements common statistical models
`bayesian_models` aims to implement `sklearn` style classes,
representing general types of models a user may wish to specify. Since
there is a very large variety of statistical models available, only some
are included in this library in a somewhat ad-hoc manner. The following
models are planned for implementation:
* BEST (Bayesian Estimation Superceeds the t Test) := Statistical
comparisons' between groups, analoguous to hypothesis testing
(COMPLETED)
## Installation
`bayesian-models` can be installed with pip
```
pip install bayesian-models
```
Newer releases are first published to TestPyPI. They are installable as
follows
```
pip install --index-url https://test.pypi.org/simple/ --extra-index-url https://pypi.org/simple bayesian-models
```
To install from git:
```
pip install git+ssh://git@github.com/AlexRodis/bayesian-models.git
```
To install the developement version run:
```
pip install 'bayesian_models[dev]@ git+ssh://git@github.com/AlexRodis/bayesian_models.git@dev-main'
```
It is often desirable to run models with a GPU if available. At present,
there are known issues with the `numpyro` dependency. Only these
versions are supported:
```
jax==0.4.1
jaxlib==0.4.1
```
To attempt to install with GPU support run:
```
pip install 'bayesian_models[GPU]@git+ssh://git@github.com/AlexRodis/bayesian-models.git'
```
Note: the GPU version is unstable
You must also set the following environment variable prior to all other
commands, including imports
```
XLA_PYTHON_CLIENT_PREALLOCATE=false
```
These dependencies are only required with
`pymc.sampling.jax.sample_numpyro_nuts` and if using the default options
can be ignored
Raw data
{
"_id": null,
"home_page": "",
"name": "bayesian-models",
"maintainer": "",
"docs_url": null,
"requires_python": "",
"maintainer_email": "Alexander Rodis <alexanderrhodis@gmail.com>",
"keywords": "bayesian,models,statistics,pymc",
"author": "",
"author_email": "Alexander Rodis <alexanderrhodis@gmail.com>",
"download_url": "https://files.pythonhosted.org/packages/8a/2d/4ce6b4a58bd863bde0d84a6f0cbedcaf48442c654a4f0c17a9b4414625ff/bayesian-models-0.1.0.tar.gz",
"platform": null,
"description": "# bayesian-models\n`bayeian-models` is a small library build on top of `pymc` that\nimplements common statistical models\n\n`bayesian_models` aims to implement `sklearn` style classes,\nrepresenting general types of models a user may wish to specify. Since\nthere is a very large variety of statistical models available, only some\nare included in this library in a somewhat ad-hoc manner. The following\nmodels are planned for implementation:\n\n* BEST (Bayesian Estimation Superceeds the t Test) := Statistical\n comparisons' between groups, analoguous to hypothesis testing\n (COMPLETED)\n\n## Installation\n\n`bayesian-models` can be installed with pip\n\n```\npip install bayesian-models\n```\n\nNewer releases are first published to TestPyPI. They are installable as\nfollows\n\n```\npip install --index-url https://test.pypi.org/simple/ --extra-index-url https://pypi.org/simple bayesian-models\n```\n\nTo install from git:\n\n```\npip install git+ssh://git@github.com/AlexRodis/bayesian-models.git\n```\n\nTo install the developement version run:\n```\npip install 'bayesian_models[dev]@ git+ssh://git@github.com/AlexRodis/bayesian_models.git@dev-main'\n```\n\nIt is often desirable to run models with a GPU if available. At present,\nthere are known issues with the `numpyro` dependency. Only these\nversions are supported:\n\n```\njax==0.4.1\njaxlib==0.4.1\n```\n\n\nTo attempt to install with GPU support run:\n\n```\npip install 'bayesian_models[GPU]@git+ssh://git@github.com/AlexRodis/bayesian-models.git'\n```\n\nNote: the GPU version is unstable\n\nYou must also set the following environment variable prior to all other\ncommands, including imports\n\n```\nXLA_PYTHON_CLIENT_PREALLOCATE=false\n```\n\nThese dependencies are only required with\n`pymc.sampling.jax.sample_numpyro_nuts` and if using the default options\ncan be ignored\n\n",
"bugtrack_url": null,
"license": "",
"summary": "A package for building common bayesian models in pymc",
"version": "0.1.0",
"split_keywords": [
"bayesian",
"models",
"statistics",
"pymc"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "1aa30695c312ff45ffaa6138eae1c64914e588ac285b7f7fee50c3ffecdc9bc2",
"md5": "4183ccf9e2ce5d49f857a4f38b8688a1",
"sha256": "95a99c1f93b0ead2b11b5220bcbf1c9fb3bd4e2e445e0ccd27f79986f792e66b"
},
"downloads": -1,
"filename": "bayesian_models-0.1.0-py3-none-any.whl",
"has_sig": false,
"md5_digest": "4183ccf9e2ce5d49f857a4f38b8688a1",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": null,
"size": 62284,
"upload_time": "2023-04-05T09:14:04",
"upload_time_iso_8601": "2023-04-05T09:14:04.414050Z",
"url": "https://files.pythonhosted.org/packages/1a/a3/0695c312ff45ffaa6138eae1c64914e588ac285b7f7fee50c3ffecdc9bc2/bayesian_models-0.1.0-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "8a2d4ce6b4a58bd863bde0d84a6f0cbedcaf48442c654a4f0c17a9b4414625ff",
"md5": "63749a861907698bf839b0ba30225b45",
"sha256": "d2772931c5cef8cb894054cbbcf388614462e69dcf6a95339bcf1e8a98f7f030"
},
"downloads": -1,
"filename": "bayesian-models-0.1.0.tar.gz",
"has_sig": false,
"md5_digest": "63749a861907698bf839b0ba30225b45",
"packagetype": "sdist",
"python_version": "source",
"requires_python": null,
"size": 59802,
"upload_time": "2023-04-05T09:14:06",
"upload_time_iso_8601": "2023-04-05T09:14:06.394917Z",
"url": "https://files.pythonhosted.org/packages/8a/2d/4ce6b4a58bd863bde0d84a6f0cbedcaf48442c654a4f0c17a9b4414625ff/bayesian-models-0.1.0.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-04-05 09:14:06",
"github": false,
"gitlab": false,
"bitbucket": false,
"lcname": "bayesian-models"
}