Input preprocessing used to generate this hindcast dataset
#1
by spark618 - opened
Hi, thank you for sharing this dataset.
We are trying to reproduce these hindcasts with the released ONNX model and ERA5. Could you tell us which input preprocessing was used to generate them: the units (physical units or standardized), whether log1p was applied to precipitation, and how the daily means were formed from hourly ERA5?
We have posted the details on the GitHub repository: https://github.com/tpys/FuXi-S2S/issues/5 Thank you!