Skip to content

Tinrry/thermal_conductivity

Repository files navigation

Thermal Conductivity Project

项目介绍

本项目用于热导率计算与贝叶斯分析。部分输入文件因体积过大未上传至 GitHub:


如果你正在查找论文 'On-the-fly machine learning-assisted high accuracy second-principles model for BaTiO3' 的相关代码,请参考 Bilibili 视频 了解如何使用本项目。

相关代码可在 GitH pymc 分支 获取。


If you are coming here for the code of 'On-the-fly machine learning-assisted high accuracy second-principles model for BaTiO3', please refer to Bilibili to know how to use the codes.

And the code related to that parper can be obtained via: pymc branch


快速开始

1. 热导率计算

运行以下命令,输出 heat.txt

python workflow.py

2. 结果展示

基于 heat.txt 展示结果:

python heat-plot.py

热导率结果图


运行流程说明

详细流程与说明请参考:Notion 文档


备注

  • BZO 示例在 mpirunnp=10,共 517 个 case,约需 1 小时。

  • 以下命令根据 --mesh 参数耗时不同:

    phono3py --fc3 --fc2 --dim="1 1 1" --mesh="11 11 11" --br --tmin=10 --tmax=1000 > Heat.txt &

贝叶斯实现(bayesian implement)

相关代码均放在 bayesian 目录下。

1. 数据格式转换

将 txt 文件转换为 csv:

python dataset.py

数据转换示例

2. 使用 pymc3 实现贝叶斯分析

运行:

python lr_poly.py

贝叶斯分析结果

About

总共有3个分支

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published