graph2mat


Namegraph2mat JSON
Version 0.0.6 PyPI version JSON
download
home_pageNone
SummaryUtility package to work with equivariant matrices and graphs.
upload_time2024-08-18 17:47:49
maintainerNone
docs_urlNone
authorNone
requires_python>=3.9
licenseMIT
keywords machine learning equivariance e3nn matrix
VCS
bugtrack_url
requirements No requirements were recorded.
Travis-CI No Travis.
coveralls test coverage No coveralls.
            graph2mat: Equivariant matrices meet machine learning
----------------------

![graph2mat_overview](https://raw.githubusercontent.com/BIG-MAP/graph2mat/main/docs/_static/images/graph2mat_overview.svg)

The aim of `graph2mat` is to pave your way into meaningful science by providing the **tools to interface to common machine learning frameworks** (e3nn, pytorch) **to learn equivariant matrices.**

**[Documentation](https://big-map.github.io/graph2mat/)**

It also provides a **set of tools** to facilitate the training and usage of the models created using the package:

- **Training tools**: It contains custom `pytorch_lightning` modules to train, validate and test the orbital matrix models.
- **Server**: A production ready server (and client) to serve predictions of the trained
    models. Implemented using `fastapi`.
- **Siesta**: A set of tools to interface the machine learning models with SIESTA. These include tools for input preparation, analysis of performance...

The package also implements a **command line interface** (CLI): `graph2mat`. The aim of this CLI is
to make the usage of `graph2mat`'s tools as simple as possible. It has two objectives:

- Make life easy for the model developers.
- Facilitate the usage of the models by non machine learning scientists, who just want
  good predictions for their systems.

Installation
------------

It can be installed with pip. Adding the tools extra will also install all the dependencies
needed to use the tools provided.

```
pip install graph2mat[tools]
```

If you want to use `graph2mat` with e3nn you can also ask for the `e3nn` extra dependencies:

```
pip install graph2mat[tools,e3nn]
```

You can also ask for

What is an equivariant matrix?
------------------------------

![water_equivariant_matrix](https://raw.githubusercontent.com/BIG-MAP/graph2mat/main/docs/_static/images/water_equivariant_matrix.png)


Contributions
--------------

We are very open to suggestions, contributions, discussions...

- If you have questions or want do discuss an idea, please [start a discussion](https://github.com/BIG-MAP/graph2mat/discussions)
- If you have a feature suggestion or bug report, please [open an issue](https://github.com/BIG-MAP/graph2mat/issues)

We look forward to your contributions!

            

Raw data

            {
    "_id": null,
    "home_page": null,
    "name": "graph2mat",
    "maintainer": null,
    "docs_url": null,
    "requires_python": ">=3.9",
    "maintainer_email": null,
    "keywords": "machine learning, equivariance, e3nn, matrix",
    "author": null,
    "author_email": "Pol Febrer <pfebrer96@gmail.com>, =?utf-8?q?Peter_B=2E_J=C3=B8rgensen?= <peterbjorgensen@gmail.com>",
    "download_url": "https://files.pythonhosted.org/packages/29/d2/12b5ebab58849bb0af6dcbcdae5e1e70b67e2d3d686ef11953abc3f14d3b/graph2mat-0.0.6.tar.gz",
    "platform": null,
    "description": "graph2mat: Equivariant matrices meet machine learning\n----------------------\n\n![graph2mat_overview](https://raw.githubusercontent.com/BIG-MAP/graph2mat/main/docs/_static/images/graph2mat_overview.svg)\n\nThe aim of `graph2mat` is to pave your way into meaningful science by providing the **tools to interface to common machine learning frameworks** (e3nn, pytorch) **to learn equivariant matrices.**\n\n**[Documentation](https://big-map.github.io/graph2mat/)**\n\nIt also provides a **set of tools** to facilitate the training and usage of the models created using the package:\n\n- **Training tools**: It contains custom `pytorch_lightning` modules to train, validate and test the orbital matrix models.\n- **Server**: A production ready server (and client) to serve predictions of the trained\n    models. Implemented using `fastapi`.\n- **Siesta**: A set of tools to interface the machine learning models with SIESTA. These include tools for input preparation, analysis of performance...\n\nThe package also implements a **command line interface** (CLI): `graph2mat`. The aim of this CLI is\nto make the usage of `graph2mat`'s tools as simple as possible. It has two objectives:\n\n- Make life easy for the model developers.\n- Facilitate the usage of the models by non machine learning scientists, who just want\n  good predictions for their systems.\n\nInstallation\n------------\n\nIt can be installed with pip. Adding the tools extra will also install all the dependencies\nneeded to use the tools provided.\n\n```\npip install graph2mat[tools]\n```\n\nIf you want to use `graph2mat` with e3nn you can also ask for the `e3nn` extra dependencies:\n\n```\npip install graph2mat[tools,e3nn]\n```\n\nYou can also ask for\n\nWhat is an equivariant matrix?\n------------------------------\n\n![water_equivariant_matrix](https://raw.githubusercontent.com/BIG-MAP/graph2mat/main/docs/_static/images/water_equivariant_matrix.png)\n\n\nContributions\n--------------\n\nWe are very open to suggestions, contributions, discussions...\n\n- If you have questions or want do discuss an idea, please [start a discussion](https://github.com/BIG-MAP/graph2mat/discussions)\n- If you have a feature suggestion or bug report, please [open an issue](https://github.com/BIG-MAP/graph2mat/issues)\n\nWe look forward to your contributions!\n",
    "bugtrack_url": null,
    "license": "MIT",
    "summary": "Utility package to work with equivariant matrices and graphs.",
    "version": "0.0.6",
    "project_urls": null,
    "split_keywords": [
        "machine learning",
        " equivariance",
        " e3nn",
        " matrix"
    ],
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "blake2b_256": "29d212b5ebab58849bb0af6dcbcdae5e1e70b67e2d3d686ef11953abc3f14d3b",
                "md5": "09175b41c838b7991c9521b0541fa8cb",
                "sha256": "5d1dc9b257dddbdbf988b0975024f4d700273b91ad4d6c2971f4bcc451763891"
            },
            "downloads": -1,
            "filename": "graph2mat-0.0.6.tar.gz",
            "has_sig": false,
            "md5_digest": "09175b41c838b7991c9521b0541fa8cb",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": ">=3.9",
            "size": 1123527,
            "upload_time": "2024-08-18T17:47:49",
            "upload_time_iso_8601": "2024-08-18T17:47:49.001735Z",
            "url": "https://files.pythonhosted.org/packages/29/d2/12b5ebab58849bb0af6dcbcdae5e1e70b67e2d3d686ef11953abc3f14d3b/graph2mat-0.0.6.tar.gz",
            "yanked": false,
            "yanked_reason": null
        }
    ],
    "upload_time": "2024-08-18 17:47:49",
    "github": false,
    "gitlab": false,
    "bitbucket": false,
    "codeberg": false,
    "lcname": "graph2mat"
}
        
Elapsed time: 1.78098s