2808条流场数据，取自51个超临界翼型在Ma=0.73和不同攻角范围内（-2.0~4.6）的流场数据。分辨率为192,384）。
Firstly, to use the dataset, you need to install the following packege
1.MindSpore https://www.mindspore.cn/install/
2.MindFlow https://www.mindspore.cn/mindflow/docs/zh-CN/master/mindflow_install.html
The dataset contains 2808 flow field data for 51 supercritical airfoils at Ma=0.73 and in the range of different angles of attack (-2.0~4.6). Where, the data dimensions of input are (13, 192, 384), 192 and 384 are the grid resolution after Jacobi conversion, and 13 are different feature dimensions, respectively $$(AOA, x, y, {{x}_{i,0}},{{y}_{i,0}},{{\xi }_{x}},{{\xi }_{y}},{{\eta }_{x}},{{\eta }_{y}},{{x}_{\xi }},{{x}_{\eta }},{{y}_{\xi }},{{y}_{\eta }}$$
The data dimension of Label is (288,768). The flow field data (u,v,p) obtained after the patchify function (16×16) operation can be restored to (3, 192, 384) through the unpatchify operation. Users can customize and select according to their own network input and output design.
the following code is the example of loading data
mindrecord_name = "flowfiled_000_050.mind"
dataset = ds.MindDataset(dataset_files=mindrecord_name, shuffle=False)
dataset = dataset.project(["inputs", "labels"])
print("samples:", dataset.get_dataset_size())
for data in dataset.create_dict_iterator(num_epochs=1, output_numpy=False):
input = data["inputs"]
label = data["labels"]
print(input.shape)
print(label.shape)
break
