mSANN model benchmarks

This data set contains scripts and dataset needed to reproduce the results in the following paper:

Mohamed Amine Bouhlel, Sicheng He, and Joaquim R. R. A. Martins. Scalable gradient-enhanced artificial neural networks forairfoil shape design in the subsonic and transonic regimes.

In this paper, we mainly produced three studies:
* An analytical test case stored in the repository "Rosenbrock": This repository contains the dataset for running the training and validation of the models. It also contains three repositories ANN, SANN, and mSANN that contains the scripts needed to rerun the models, respectively.
* The airfoil shape design analysis test case in both subsonic and transonic regimes in the repository "Analysis/mSANN": This repository contains three repositories "cd", "cl", and "cd" for training the mSANN model on the aerodynamic coefficients. Each sub-repository contains the dataset (for training, validation, and testing the model) and two script files run.py and prediction.py for training the neural network and make predictions.
* The airfoil shape design optimization test case in both subsonic and transonic regimes in the repossitory "Optimization": This repository contains two repositories "subsonic" and "transonic" for running an optimization either using the mSANN model or CFD.

本数据集包含复现下述论文实验成果所需的脚本与数据集：
Mohamed Amine Bouhlel、Sicheng He 与 Joaquim R. R. A. Martins. 适用于亚音速与跨音速翼型外形设计的可扩展梯度增强人工神经网络。

本论文主要开展了三项研究：
* 存储于仓库"Rosenbrock"中的解析测试用例：该仓库包含模型训练与验证所需的数据集，同时包含三个子仓库ANN、SANN与mSANN，分别存放复现对应模型的脚本。
* 存储于仓库"Analysis/mSANN"中的亚音速与跨音速翼型外形设计分析测试用例：该仓库包含三个子仓库"cd"、"cl"与"cd"，用于基于空气动力学系数训练mSANN模型。每个子仓库均包含模型训练、验证与测试所需的数据集，以及两个脚本文件run.py与prediction.py，分别用于神经网络训练与预测任务。
* 存储于仓库"Optimization"中的亚音速与跨音速翼型外形设计优化测试用例：该仓库包含两个子仓库"subsonic"与"transonic"，分别用于基于mSANN模型或计算流体动力学（Computational Fluid Dynamics，CFD）开展优化任务。