dxtb

Efficient And Fully Differentiable Extended Tight-Binding

https://github.com/grimme-lab/dxtb

Science Score: 75.0%

This score indicates how likely this project is to be science-related based on various indicators:

✓
CITATION.cff file
Found CITATION.cff file
✓
codemeta.json file
Found codemeta.json file
✓
.zenodo.json file
Found .zenodo.json file
✓
DOI references
Found 10 DOI reference(s) in README
○
Academic publication links
✓
Committers with academic emails
2 of 6 committers (33.3%) from academic institutions
✓
Institutional organization owner
Organization grimme-lab has institutional domain (www.chemie.uni-bonn.de)
○
JOSS paper metadata
○
Scientific vocabulary similarity
Low similarity (9.6%) to scientific vocabulary

Keywords

automatic-differentiation computational-chemistry pytorch quantum-chemistry tight-binding

Keywords from Contributors

ase

Last synced: 6 months ago · JSON representation ·

Repository

Efficient And Fully Differentiable Extended Tight-Binding

Basic Info

Host: GitHub
Owner: grimme-lab
License: apache-2.0
Language: Python
Default Branch: main
Homepage: https://dxtb.readthedocs.io
Size: 3.34 MB

Statistics

Stars: 100
Watchers: 4
Forks: 20
Open Issues: 5
Releases: 5

Topics

automatic-differentiation computational-chemistry pytorch quantum-chemistry tight-binding

Created about 4 years ago · Last pushed 6 months ago

Metadata Files

Readme Contributing License Citation

Fully Differentiable Extended Tight-Binding

- Combining semi-empirical quantum chemistry with machine learning in PyTorch -

The xTB methods (GFNn-xTB) are a series of semi-empirical quantum chemical methods that provide a good balance between accuracy and computational cost.

With dxtb, we provide a re-implementation of the xTB methods in PyTorch, which allows for automatic differentiation and seamless integration into machine learning frameworks.

NOTE: If you encounter any bugs or have questions on how to use dxtb, feel free to open an issue.

Installation

pip

dxtb can easily be installed with pip.

sh pip install dxtb[libcint]

Installing the libcint interface is highly recommended, as it is significantly faster than the pure PyTorch implementation and provides access to higher-order multipole integrals and their derivatives (required for GFN2-xTB). However, the interface is currently only available on Linux.

conda

dxtb is also available on conda from the conda-forge channel.

sh mamba install dxtb

Don't forget to install the libcint interface (not on conda) via pip install tad-libcint. The libcint interface is required for GFN2-xTB.

For Windows, dxtb is not available via conda, because PyTorch itself is not registered in the conda-forge channel.

Other

For more options, see the installation guide in the documentation.

Example

The following example demonstrates how to compute the energy and forces using GFN1-xTB.

```python import torch import dxtb

dd = {"dtype": torch.double, "device": torch.device("cpu")}

LiH

numbers = torch.tensor([3, 1], device=dd["device"]) positions = torch.tensor([[0.0, 0.0, 0.0], [0.0, 0.0, 1.5]], **dd)

instantiate a calculator

calc = dxtb.calculators.GFN1Calculator(numbers, **dd)

compute the energy

pos = positions.clone().requiresgrad(True) energy = calc.get_energy(pos)

obtain gradient (dE/dR) via autograd

(g,) = torch.autograd.grad(energy, pos)

Alternatively, forces can directly be requested from the calculator.

(Don't forget to manually reset the calculator when the inputs are identical.)

calc.reset() pos = positions.clone().requiresgrad(True) forces = calc.get_forces(pos)

assert torch.equal(forces, -g) ```

All quantities are in atomic units.

For more examples and details, check out the documentation.

Compatibility

| PyTorch \ Python | 3.8 | 3.9 | 3.10 | 3.11 | 3.12 | |------------------|--------------------|--------------------|--------------------|--------------------|--------------------| | 1.11.0 | :whitecheckmark: | :whitecheckmark: | :x: | :x: | :x: | | 1.12.1 | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | :x: | :x: | | 1.13.1 | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | :x: | | 2.0.1 | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | :x: | | 2.1.2 | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | :x: | | 2.2.2 | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | | 2.3.1 | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | | 2.4.1 | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | | 2.5.1 | :x: | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | | 2.6.0 | :x: | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | | 2.7.1 | :x: | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | | 2.8.0 | :x: | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: | :whitecheckmark: |

Note that only the latest bug fix version is listed, but all preceding bug fix minor versions are supported. For example, although only version 2.2.2 is listed, version 2.2.0 and 2.2.1 are also supported.

Restriction for macOS and Windows:

On macOS and Windows, PyTorch<2.0.0 does only support Python<3.11.

The libcint interface is not available for macOS and Windows. Correspondingly, the integral evaluation can be considerably slower. Moreover, higher-order multipole integrals (dipole, quadrupole, ...) are not implemented. While macOS support may be considered in the future, native Windows support is not possible, because the underlying libcint library does not work under Windows.

Citation

If you use dxtb in your research, please cite the following paper:

M. Friede, C. Hölzer, S. Ehlert, S. Grimme, dxtb -- An Efficient and Fully Differentiable Framework for Extended Tight-Binding, J. Chem. Phys., 2024, 161, 062501. (DOI)

The Supporting Information can be found here.

For details on the xTB methods, see

C. Bannwarth, E. Caldeweyher, S. Ehlert, A. Hansen, P. Pracht, J. Seibert, S. Spicher, S. Grimme, WIREs Comput. Mol. Sci., 2020, 11, e01493. (DOI)
C. Bannwarth, S. Ehlert, S. Grimme, J. Chem. Theory Comput., 2019, 15, 1652-1671. (DOI)
S. Grimme, C. Bannwarth, P. Shushkov, J. Chem. Theory Comput., 2017, 13, 1989-2009. (DOI)

Contributing

This is a volunteer open source projects and contributions are always welcome. Please, take a moment to read the contributing guidelines.

License

This project is licensed under the Apache License, Version 2.0 (the "License"); you may not use this project's files except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

Owner

Name: Grimme lab
Login: grimme-lab
Kind: organization
Email: xtb@thch.uni-bonn.de
Location: University of Bonn

Website: https://www.chemie.uni-bonn.de/grimme/de
Twitter: GrimmeLab
Repositories: 53
Profile: https://github.com/grimme-lab

Quantum chemistry software - Made in Bonn.

Citation (CITATION.bib)

@article{dxtb.2024,
    title = {dxtb -- An Efficient and Fully Differentiable Framework for Extended Tight-Binding},
    author = {Friede, Marvin and H{\"o}lzer, Christian and Ehlert, Sebastian and Grimme, Stefan},
    journal = {The Journal of Chemical Physics},
    volume = {161},
    number = {6},
    pages = {062501},
    year = {2024},
    month = {08},
    issn = {0021-9606},
    doi = {10.1063/5.0216715},
    url = {https://doi.org/10.1063/5.0216715},
}

GitHub Events

Total

Create event: 33
Release event: 2
Issues event: 26
Watch event: 33
Delete event: 37
Issue comment event: 85
Push event: 98
Pull request review comment event: 22
Pull request review event: 15
Pull request event: 71
Fork event: 8

Last Year

Create event: 33
Release event: 2
Issues event: 26
Watch event: 33
Delete event: 37
Issue comment event: 85
Push event: 98
Pull request review comment event: 22
Pull request review event: 15
Pull request event: 71
Fork event: 8

Committers

Last synced: 9 months ago

All Time

Total Commits: 521
Total Committers: 6
Avg Commits per committer: 86.833
Development Distribution Score (DDS): 0.223

Past Year

Commits: 61
Committers: 4
Avg Commits per committer: 15.25
Development Distribution Score (DDS): 0.23

Top Committers

Name	Email	Commits
marvinfriede	5****e	405
hoelzerC	h**r@t**e	58
Sebastian Ehlert	2****k	42
pre-commit-ci[bot]	6****]	11
TCH-MGH	t**r@m**u	4
Ihor Neporozhnii	7****i	1

Committer Domains (Top 20 + Academic)

mgh.harvard.edu: 1 thch.uni-bonn.de: 1

Issues and Pull Requests

Last synced: 6 months ago

All Time

Total issues: 48
Total pull requests: 149
Average time to close issues: 4 months
Average time to close pull requests: 23 days
Total issue authors: 14
Total pull request authors: 6
Average comments per issue: 2.19
Average comments per pull request: 0.88
Merged pull requests: 134
Bot issues: 1
Bot pull requests: 19

Past Year

Issues: 22
Pull requests: 79
Average time to close issues: 10 days
Average time to close pull requests: 2 days
Issue authors: 11
Pull request authors: 5
Average comments per issue: 2.05
Average comments per pull request: 0.99
Merged pull requests: 70
Bot issues: 1
Bot pull requests: 16

View more stats

Top Authors

Issue Authors

hoelzerC (13)
marvinfriede (12)
ivan-pikulin (5)
njzjz (4)
Yufan121 (2)
MoleOrbitalHybridAnalyst (2)
blaschma (2)
vladislavivanistsev (1)
AdeeshKolluru (1)
MauriceDHanisch (1)
LiuGaoyong (1)
recisic (1)
awvwgk (1)

Pull Request Authors

marvinfriede (130)
pre-commit-ci[bot] (26)
hoelzerC (11)
ineporozhnii (2)
MauriceDHanisch (2)
recisic (2)

Top Labels

Issue Labels

enhancement (8) bug (7) performance (7) version (4) autograd (2) feature (2) documentation (1) question (1)

Pull Request Labels

version (4) enhancement (3) autograd (1) performance (1) bug (1)

Packages

Total packages: 1
Total downloads:
- pypi 84 last-month

Total dependent packages: 0
Total dependent repositories: 0
Total versions: 6
Total maintainers: 1

pypi.org: dxtb

Fully Differentiable Approach to Extended Tight Binding

Homepage: https://github.com/grimme-lab/dxtb
Documentation: https://github.com/grimme-lab/dxtb
License: Apache-2.0
Latest release: 0.3.0
published 6 months ago

Versions: 6
Dependent Packages: 0
Dependent Repositories: 0
Downloads: 84 Last month

Rankings

Dependent packages count: 10.7%

Average: 35.5%

Dependent repos count: 60.4%

Maintainers (1)

marvinfriede

Last synced: 6 months ago

Dependencies

.github/workflows/python.yml actions

actions/checkout v3 composite
actions/setup-python v4 composite
codecov/codecov-action v3 composite

.github/workflows/release.yaml actions

actions/checkout v3 composite
actions/download-artifact v3 composite
actions/upload-artifact v3 composite
pypa/gh-action-pypi-publish release/v1 composite

docs/requirements.txt pypi

jinja2 *
sphinx *
sphinx-book-theme *
sphinx-copybutton *
sphinx-design *
sphinx-togglebutton *
sphinx_autodoc_typehints *

environment.yaml pypi

tad-libcint *

pyproject.toml pypi

setup.py pypi

dxtb

Science Score: 75.0%

Keywords

Keywords from Contributors

Repository

Basic Info

Statistics

Topics

Metadata Files

README.md

Fully Differentiable Extended Tight-Binding

Installation

pip

conda

Other

Example

LiH

instantiate a calculator

compute the energy

obtain gradient (dE/dR) via autograd

Alternatively, forces can directly be requested from the calculator.

(Don't forget to manually reset the calculator when the inputs are identical.)

Compatibility

Citation

Contributing

License

Owner

Citation (CITATION.bib)

GitHub Events

Total

Last Year

Committers

All Time

Past Year

Top Committers

Committer Domains (Top 20 + Academic)

Issues and Pull Requests

All Time

Past Year

Top Authors

Issue Authors

Pull Request Authors

Top Labels

Issue Labels

Pull Request Labels

Packages

pypi.org: dxtb

Rankings

Maintainers (1)

Dependencies