http://mcloud-doc.lonxun.com/wiki/ 2026-04-23T04:52:27+00:00 http://mcloud-doc.lonxun.com/wiki/ http://mcloud-doc.lonxun.com/wiki/_media/logo.png text/html 2024-02-05T03:47:58+00:00 Anonymous (anonymous@undisclosed.example.com) http://mcloud-doc.lonxun.com/wiki/pwmat/pwmlff/appendix1?rev=1707104878&do=diff Appendix I: features Wiki This section provides a brief introduction on the features used in PWMLFF. The related literature is also listed, for readers’ reference. What are features? Features (or descriptors) are quantities that describe the local atomic environment of an atom. They are required preserve the translational, rotational, and permutational symmetries. Features are usually used as the input of various regressors(linear model, NN, .etc), which output atomic energies and forces.$$ … text/html 2024-02-05T03:13:54+00:00 Anonymous (anonymous@undisclosed.example.com) http://mcloud-doc.lonxun.com/wiki/pwmat/pwmlff/install?rev=1707102834&do=diff Installation PWMLFF 安装 (以Mcloud为例) PWMLFF包含Fortran、Python和CUDA加速等，需要在包含Python环境、gcc编译器、GPU硬件条件下进行安装 方式一：使用Mcloud现有环境 mcloud已有配置好的conda环境，可以直接调用，避免自己安装anaconda, cudatoolkit, pytorch等极度耗时的过程，具体步骤如下： text/html 2024-02-05T03:46:52+00:00 Anonymous (anonymous@undisclosed.example.com) http://mcloud-doc.lonxun.com/wiki/pwmat/pwmlff/param_dtls?rev=1707104812&do=diff Parameter details 本节介绍了所有模型中可由用户定义的参数，可以分为必要参数和高级参数两类。必要参数需要用户指定，高级参数采用了默认值，用户可以在 json 文件中根据需求手动修改。在下面的参数中，“相对路径（relative path）” 表示相对于当前工作目录的路径，而 “绝对路径（absolute path）” 表示从根目录开始的文件或目录的完整路径。$6.0 \text{\AA}$$0.5 \text{\AA}$$6.0 \text{\AA}$$0.5 \text{\AA}$$6.0 \text{\AA}$$0.5 \text{\AA}$$$ \text{decay\_rate} = \exp\left(\frac{\log(\text{stop\_lr}/\text{start\_lr})}{\text{stop\_step}/\text{decay\_step}}\right) $$$$ \text{real\_lr} = \text{start\_lr} \cdot \left(\mathrm{decay\_rate}\right)^{\left(\le… text/html 2024-02-05T03:00:44+00:00 Anonymous (anonymous@undisclosed.example.com) http://mcloud-doc.lonxun.com/wiki/pwmat/pwmlff/start?rev=1707102044&do=diff Machine Learning Force Field PWmat Machine Learning Force Field (PWMLFF) is an open source software under GNU license. It aims at generating force fields with accuracy comparable to Ab Initio Molecular Dynamics (AIMD). It is compatible with AIMD data with either