issue with the index.npy size
Hello,
I am having trouble understanding your dataset structure. The index.npy file is supposed to have a shape of N×9, but the one I obtained is N×12. Additionally, the first column is expected to contain geometry indices in the range [0, 4239], but instead I am seeing integer values in the range [0, 5871].
Is this expected, or am I possibly using a different dataset version or preprocessing configuration?
Thank you for your help.
Hi,
Thank you very much for the report. That true, the first column should be [0, 4239], and currently it is up to 5871 because we forgot to transfer the running index to a continues one (there are geometries that fails during generation). I'll update it asap but you can just cover the first column with a ascending contiunes index for each unique one. For the shape, I just update the introduction. The column 5-8 are actually the coefficients from the solver, and column 9-11 are the actuall coefficients used for ML models. Again, thanks for your feedback. If you have further question, we can figure them out together.
Thank you.
I replaced the first column with sequential indices for each unique value, and now everything works perfectly — I’m getting the correct number of wing shapes.