Metadata-Version: 2.1
Name: mlipdockers
Version: 0.0.8
Summary: Request to docker containers in which the python enviroments for different machine learning potential usages are implemented. Using this package, one can get the predicted potential energy for any structure using any MLIP without needing to change python environments.
Home-page: https://github.com/HouGroup/ys_mlipdk/
Author: Yaoshu Xie
Author-email: jasonxie@sz.tsinghua.edu.cn
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.10
Description-Content-Type: text/markdown
Requires-Dist: docker

A docker socket allowing for multiple MLIP usages within the same python environment.  
`mlipdockers`是一个实现在同一个python环境中使用不同机器学习原子间势（MLIP）的docker接口。

Install: `pip install mlipdockers`

Integrated machine learning interatomic potentials (MLIPs) including `grace-2l` `chgnet` `mace` `orb-models` `sevenn` `eqv2`. Details can be find in https://matbench-discovery.materialsproject.org/

Our images are uploaded in the Alibaba Cloud. Therefore, to use our package, you need to register an Alibaba Cloud account at https://account.alibabacloud.com/ and install docker.  
docker镜像上传在阿里云，因此需要注册阿里云账号才能使用。

After you register your Alibaba Cloud account, go to the `Container Registry/Instances` page, follow the instruction to register for a totally free `Instance of Personal Edition`, and get your countainer registry [username] and [password] which you will need to login in to the docker registry.  
注册账号以后，进入`容器镜像服务`页面，根据提示注册免费的`个人实例`，在`个人实例`-`访问凭证`获得`docker login ...`命令，复制到本地运行进行登录便获得了访问阿里云上公开镜像的权限。

![image](https://github.com/user-attachments/assets/bd4240f8-f9d2-4f36-990b-579963a7462a)

Finally, execute the `docker login` command provided in your own `Container Registry/Instances` page, and try to run tutorial.ipynb.

Try [examples](https://github.com/HouGroup/mlipdockers/tree/main/examples) now!
