Integrate and test deep learning (DL) and physics-informed deep learning (PIML) streamflow models and data assimilation into NOAA’s Next Generation Water Modeling Engine and Framework Prototype.
There is a bug here where the raster is flipped then the data to process is selected from the wrong dataset -- this should be data = data.sel(...) but that seems to lead to another error where the dataset has a 0 length dimension. The raster really does need flipped to be consistent with the weight file generated from from exact extract (and to be consistent with any other standard geo-processing...).
I am working through this bug and the subsequent 0 length dim problem identified by @glitt13 when using the proper flipped raster. I should have a patch pushed in the next day or so.