data.py里的各个文件的生成

data.py里的各个文件的生成

Load forcing data

ERA5_LAND_label_4_1990？？？哪生成的

ERA5-Land_forcing {sr} spatial resolution {year}.npy？？哪来的

lat_{s}.npy #109

lon_{s}.npy#110

filter Glacial region

Mask with {sr} spatial resolution.npy#204

Determine whether to perform memmap mapping

# Memmap is to deal with the problem of excessive amount of data

forcing_memmap.npy？？？#120通过flush

Load land_surface data

ERA5-Land_land_surface {sr} spatial resolution {year}.npy？？哪来的

land_surface_memmap.npy？哪生成的！#172通过flush

Load label data

# filter Glacial region

Mask with {sr} spatial resolution.npy#204

Load static_norm data

static_norm.npy#224

Partition and create datasets

Create a training dataset；

The default training data is 1990 to 2019.

x_train.npy？？？哪生成？通过memmap里的flush

y_train.npy#265

x_test.npy#275 通过flush

normalize

# Adopt maximum and minimum normalization
# There are two forms of normalization： region and gloabl

scaler_x.npy#303，flush

scaler_y.npy#304，flush

Save the normalized dataset

x_train_norm_shape.npy#383

x_test_norm_shape.npy#384

x_test_norm.npy#396

x_train_norm.npy#385,memmap

y_test_norm.npy#401

y_train_norm.npy#402