Closed mrynik closed 1 week ago
Hey, just spotted a bug https://github.com/ENSTA-U2IS-AI/torch-uncertainty/blob/ff355d979b73e7c6acc8361e2616fdcc26d7fb6a/torch_uncertainty/metrics/classification/calibration_error.py#L63
This will fail if you pass in the ax object to the _ce_plot function because in that case fig is initialized to None. https://github.com/ENSTA-U2IS-AI/torch-uncertainty/blob/ff355d979b73e7c6acc8361e2616fdcc26d7fb6a/torch_uncertainty/metrics/classification/calibration_error.py#L18
_ce_plot
fig
None
Hi mrynik!
Thank you for the pointer! I pushed a quick fix on dev. We intend to rework the plots a bit in the following releases (#54)!
Hey, just spotted a bug https://github.com/ENSTA-U2IS-AI/torch-uncertainty/blob/ff355d979b73e7c6acc8361e2616fdcc26d7fb6a/torch_uncertainty/metrics/classification/calibration_error.py#L63
This will fail if you pass in the ax object to the
_ce_plot
function because in that casefig
is initialized toNone
. https://github.com/ENSTA-U2IS-AI/torch-uncertainty/blob/ff355d979b73e7c6acc8361e2616fdcc26d7fb6a/torch_uncertainty/metrics/classification/calibration_error.py#L18