Hello all,
I am trying to create training tiles from a classified raster dataset using the Export Training Data for Deep Learning tool. The training data is a classified raster that has class values from 1 to 58 with 24 total classes. When I run the tool, the labeled tiles are labeled according to the object ID (they contain numbers 1-24) rather than the classvalue feature that has the correct classvalues. Has anyone run into this issue before? I need to use the wider range of values to maintain consistency with other training sets that use the same schema.
Thanks!
