# Overview
`seqinf` is a Python library for performing sequential simulation-based
inference. It aims to complement the popular [`sbi`][1] package by providing
top-level abstractions that enable running methods like SNPE/SNLE/etc without
additional boilerplate. It also provides additional inference diagnostics and
support for Bayesian NDEs and active methods (e.g., ASNPE).
# Install
Installation is available through PyPI:
```sh
pip install seqinf
```
## Dependencies
Note the version clash between `sbibm`'s requirement for `sbi`, and the latest
`sbi`. This can more or less be ignored, and if this presents an issue when
installing locally, just make sure to install `sbi` _after_ `sbibm`.
# Citing this package
Although `seqinf` is intended for general pipelines, it was principally written
to wrap [ASNPE][2]. If you use this library in your work, please cite the
following:
```
@misc{griesemer2024activesequentialposteriorestimation,
title={Active Sequential Posterior Estimation for Sample-Efficient Simulation-Based Inference},
author={Sam Griesemer and Defu Cao and Zijun Cui and Carolina Osorio and Yan Liu},
year={2024},
eprint={2412.05590},
archivePrefix={arXiv},
primaryClass={cs.LG},
url={https://arxiv.org/abs/2412.05590},
}
```
[1]: https://github.com/sbi-dev/sbi
[2]: https://arxiv.org/abs/2412.05590
Raw data
{
"_id": null,
"home_page": null,
"name": "seqinf",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.10",
"maintainer_email": null,
"keywords": "simulation, machine learning, statistics",
"author": null,
"author_email": "Sam Griesemer <samgriesemer+git@gmail.com>",
"download_url": "https://files.pythonhosted.org/packages/64/e1/8da981d0ba27008700dc7d96e7cbb2bd6f62b0068c28414c5f5747694a77/seqinf-0.2.1.tar.gz",
"platform": null,
"description": "# Overview\n`seqinf` is a Python library for performing sequential simulation-based\ninference. It aims to complement the popular [`sbi`][1] package by providing\ntop-level abstractions that enable running methods like SNPE/SNLE/etc without\nadditional boilerplate. It also provides additional inference diagnostics and\nsupport for Bayesian NDEs and active methods (e.g., ASNPE).\n\n# Install\nInstallation is available through PyPI:\n\n```sh\npip install seqinf\n```\n\n## Dependencies\nNote the version clash between `sbibm`'s requirement for `sbi`, and the latest\n`sbi`. This can more or less be ignored, and if this presents an issue when\ninstalling locally, just make sure to install `sbi` _after_ `sbibm`.\n\n\n\n# Citing this package\nAlthough `seqinf` is intended for general pipelines, it was principally written\nto wrap [ASNPE][2]. If you use this library in your work, please cite the\nfollowing:\n\n```\n@misc{griesemer2024activesequentialposteriorestimation,\n title={Active Sequential Posterior Estimation for Sample-Efficient Simulation-Based Inference}, \n author={Sam Griesemer and Defu Cao and Zijun Cui and Carolina Osorio and Yan Liu},\n year={2024},\n eprint={2412.05590},\n archivePrefix={arXiv},\n primaryClass={cs.LG},\n url={https://arxiv.org/abs/2412.05590}, \n}\n```\n\n\n[1]: https://github.com/sbi-dev/sbi\n[2]: https://arxiv.org/abs/2412.05590\n",
"bugtrack_url": null,
"license": null,
"summary": "Sequential inference abstractions",
"version": "0.2.1",
"project_urls": {
"Documentation": "https://doc.olog.io/<project-name>",
"Homepage": "https://doc.olog.io/<project-name>",
"Issues": "https://git.olog.io/olog/<project-name>/issues",
"Repository": "https://git.olog.io/olog/<project-name>"
},
"split_keywords": [
"simulation",
" machine learning",
" statistics"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "57f2cc788e5a2a706cc44d5ad9cd85a4282952526d8b0ea86417f4c84658702d",
"md5": "3a448a027be3d9cfc509048fcdf29e56",
"sha256": "0fcee56e62ba9c7e1ab0ab64ec553b6bafa2c6d9b297f68c6d59b1c41695bf20"
},
"downloads": -1,
"filename": "seqinf-0.2.1-py3-none-any.whl",
"has_sig": false,
"md5_digest": "3a448a027be3d9cfc509048fcdf29e56",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.10",
"size": 72044,
"upload_time": "2024-12-11T05:44:40",
"upload_time_iso_8601": "2024-12-11T05:44:40.607655Z",
"url": "https://files.pythonhosted.org/packages/57/f2/cc788e5a2a706cc44d5ad9cd85a4282952526d8b0ea86417f4c84658702d/seqinf-0.2.1-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "64e18da981d0ba27008700dc7d96e7cbb2bd6f62b0068c28414c5f5747694a77",
"md5": "365358979946324d3761c80d0e977a0b",
"sha256": "faba68c1334a1c82717b72eacd3c578711d79e36224b52d0ffce27a8d4fa03c6"
},
"downloads": -1,
"filename": "seqinf-0.2.1.tar.gz",
"has_sig": false,
"md5_digest": "365358979946324d3761c80d0e977a0b",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.10",
"size": 61050,
"upload_time": "2024-12-11T05:44:43",
"upload_time_iso_8601": "2024-12-11T05:44:43.046789Z",
"url": "https://files.pythonhosted.org/packages/64/e1/8da981d0ba27008700dc7d96e7cbb2bd6f62b0068c28414c5f5747694a77/seqinf-0.2.1.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-12-11 05:44:43",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "seqinf"
}
JSON