Force MaxResid appears wrong
Issue #84
resolved
I did a manual check to find the large force, and the value I found manually was much smaller than what was printed in Force MaxResid. In another one I just trained, it's reporting a Force MaxResid of 4.4061e+00 for a small training set of 12 images and an rmse goal of 0.003, so this seems unlikely.
It seems there's some bug in how this is calculated...
Comments (2)
-
-
- changed status to resolved
already fixed.
- Log in to comment
This should be fixed with the commit dedb298.