GeoFuseDiff/
├── DatasetUS.py
├── Network.py
├── TrainDiffusion.py
├── Inference.py
├── README.md
We recommend using Conda to create a virtual environment and installing the following dependencies: conda create -n geofusediff python=3.9 conda activate geofusediff
conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia
pip install rasterio numpy scipy matplotlib tqdm tensorboard
Please organize your .tif raster data according to the following directory structure. Filenames must be in the format YYYYMMDDHH.tif (e.g., 2019010100.tif) so that DatasetUS.py can automatically parse the time information (month and daytime):
data_dir/
├── ERA5_2m/ # Coarse Resolution
├── CLDAS_2m/ # Fine Resolution - training
├── DEM/
│ └── dem_1km.tif
└── Land_use/
└── LinYi/