4. Run FourCastNetGFS
We use weights and normalization statistics archived at ECMWF to run FourCastNet in inference mode. These files can be downloaded at:
wget https://get.ecmwf.int/repository/test-data/ai-models/fourcastnetv2/small/weights.tar
wget https://get.ecmwf.int/repository/test-data/ai-models/fourcastnetv2/small/global_means.npy
wget https://get.ecmwf.int/repository/test-data/ai-models/fourcastnetv2/small/global_stds.npy
Run the model in inference mode:
python inference.py startdate -w </path/to/weightsandstats> -i </path/to/input/input_YYYYMMDDHH.npy> -o </path/to/output/> -l <forecast-hours>
Arguments
Requried:
startdate: string, yyyymmddhh
Optional:
-w or –weights: path to downloaded files (weights.tar, global_means.npy, global_stds.npy)
-i or –input: path to the input file
-o or –output: path to save output files
-l or –length: lead time in hours