{"id":14970743,"url":"https://github.com/a-ws-m/unlocknn","last_synced_at":"2026-02-28T20:02:47.848Z","repository":{"id":38689492,"uuid":"276433339","full_name":"a-ws-m/unlockNN","owner":"a-ws-m","description":"A Python package for adding uncertainties to neural network models of chemical systems.","archived":false,"fork":false,"pushed_at":"2022-07-05T19:52:36.000Z","size":134210,"stargazers_count":22,"open_issues_count":7,"forks_count":6,"subscribers_count":5,"default_branch":"master","last_synced_at":"2024-10-11T04:41:30.477Z","etag":null,"topics":["chemistry","keras","machine-learning","materials-science","python","tensorflow-probability"],"latest_commit_sha":null,"homepage":"","language":"PureBasic","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/a-ws-m.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null}},"created_at":"2020-07-01T16:50:18.000Z","updated_at":"2024-06-13T22:24:41.000Z","dependencies_parsed_at":"2022-09-24T12:11:23.141Z","dependency_job_id":null,"html_url":"https://github.com/a-ws-m/unlockNN","commit_stats":null,"previous_names":[],"tags_count":5,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/a-ws-m%2FunlockNN","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/a-ws-m%2FunlockNN/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/a-ws-m%2FunlockNN/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/a-ws-m%2FunlockNN/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/a-ws-m","download_url":"https://codeload.github.com/a-ws-m/unlockNN/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":219862885,"owners_count":16555951,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["chemistry","keras","machine-learning","materials-science","python","tensorflow-probability"],"created_at":"2024-09-24T13:44:04.722Z","updated_at":"2025-10-26T13:31:18.947Z","avatar_url":"https://github.com/a-ws-m.png","language":"PureBasic","funding_links":[],"categories":[],"sub_categories":[],"readme":"# unlockNN\n\n[![Documentation Status](https://readthedocs.org/projects/unlocknn/badge/?version=latest)](https://unlocknn.readthedocs.io/en/latest/?badge=latest)\n![PyPi](https://img.shields.io/pypi/v/unlockNN)\n[![status](https://joss.theoj.org/papers/b00df538a159c4b6816ec24d4d1716fb/status.svg)](https://joss.theoj.org/papers/b00df538a159c4b6816ec24d4d1716fb)\n[![made-with-python](https://img.shields.io/badge/Made%20with-Python-1f425f.svg)](https://www.python.org/)\n[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)\n[![Code style: black](https://img.shields.io/badge/code%20style-black-000000.svg)](https://github.com/psf/black)\n[![codecov](https://codecov.io/gh/a-ws-m/unlockNN/branch/master/graph/badge.svg?token=TBDX3P6OZ3)](https://codecov.io/gh/a-ws-m/unlockNN)\n\nA Python package for interpreting and extracting uncertainties in neural network models of chemical systems based upon Gaussian processes.\n\n## Statement of need\n\nNeural networks (NNs) are powerful tools for materials property prediciton (MPP)\nbased on structural information. After training, they offer a cheaper\nalternative to density function theory (DFT) and are therefore promising for\nhigh throughput screening of materials. However, most current implementations of\nNNs for MPP lack uncertainty quantifiers. Knowledge of the certainty in an\nestimate is particularly important for machine learning models, as the\nreliability of a prediction depends on the existence of functionally similar\nstructures in the training dataset, which cannot be readily determined.\n\nUnlockNN contains utilities for adding uncertainty quantification to Keras-based\nmodels. This is achieved by replacing the last layer of the model with a\n*variational Gaussian process* (VGP), a modification of a Gaussian process that\nimproves scalability to larger data sets. The caveat is that the modified model\nmust undergo further training in order to calibrate the uncertainty quantifier;\nhowever, this typically only requires a small number of training iterations.\n\nUnlockNN also contains a specific configuration for adding uncertainty\nquantification to [MEGNet](https://github.com/materialsvirtuallab/megnet/): a\npowerful graph NN model for predicting properties of molecules and crystals.\n\n## Installation\n\n\nThe package can be installed by cloning this repository and building it using either anaconda or pip,\nor it can be downloaded directly from PyPi.\n\nTo install from PyPi, run `pip install unlockNN`.\nTo install from source:\n\n```bash\ngit clone https://github.com/a-ws-m/unlockNN.git\ncd unlockNN\nconda env create -f environment.yml  # Optional: create a virtual environment with conda\npip install .\n```\n\nThe `dev_environment.yml` contains additional dependencies for development, testing and building documentation.\nIt can be installed using `conda env create -f dev_environment.yml`.\n\n## Documentation\n\nFull documentation is available for the project [here](https://unlocknn.readthedocs.io/en/latest/).\n\n## Benchmarks\n\nBenchmark details and results can be found in the [benchmarking](benchmarking/) subdirectory.\n\n## License and attribution\n\nCode licensed under the MIT License.\n\n## Development notes\n\n### Reporting issues\n\nPlease use the Issue tracker to report bugs in the software, suggest feature improvements, or seek support.\n\n### Contributing to unlockNN\n\nContributions are very welcome as we look to make unlockNN more flexible and efficient.\nPlease use the [Fork and Pull](https://guides.github.com/activities/forking/) workflow to make contributions and follow the contribution guidelines:\n\n- Use the environment defined in `dev_environment.yml`. This installs `black`, the formatter used for this project, as well as utilities for building documentation (`sphinx` and the `insegel` theme), enabling the testing suite (`pytest` and `pytest-cov`) and publishing to PyPi (`build`, but this will be handled by the package maintainer).\n- Use `black` to format all Python files that you edit: `black {edited_file.py}` or `python -m black {edited_file.py}`.\n- Write tests for new features in the appropriate directory. Run tests using `pytest tests/`, or optionally with `pytest --cov=unlocknn tests/` to generate coverage on the fly.\n- After testing that `pytest` works for your current environment, run `tox` in the root directory of the project to check that all versions of Python are compatible.\n- Use [Google-style Docstrings](https://sphinxcontrib-napoleon.readthedocs.io/en/latest/example_google.html). Check docstrings with `pydocstyle`: `pydocstyle {edited_file.py}`.\n- Check docstrings are valid Sphinx RST and that the documentation compiles without errors: in the `docs` directory, run `make html`.\n- Feel free to clean up others' code as you go along.\n\n### List of developers\n\nContributors to unlockNN:\n\n- [Alexander Moriarty](https://github.com/a-ws-m)\n\nHuge thanks to [Keith Butler](https://github.com/keeeto), [Aron Walsh](https://github.com/aronwalsh) and [Kazuki Morita](https://github.com/KazMorita) for supervising the project at its inception and for their immense support.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fa-ws-m%2Funlocknn","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fa-ws-m%2Funlocknn","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fa-ws-m%2Funlocknn/lists"}