Closed tkoskela closed 1 year ago
My hypothesis is that we are not setting something we should in https://github.com/astro-informatics/purify/blob/31dcff9be82dafcf6c44fd10cd5644579a2a1f34/cpp/purify/algorithm_factory.h#L169-L190
We could compare to e.g. tests in sopt https://github.com/astro-informatics/sopt/blob/ae4c48e397b99d6b27b727744da292cb3f1817ce/cpp/tests/inpainting.cc#L59-L76 to figure out what is missing.
@mmcleod89 I assume this should have been closed by #315?
Yeah this is one of the things that was fixed, although it wasn't actually fixed by #315 but by a separate bug fix within SOPT itself. Do we want to close it now, or wait for the fix to be propagated to the main?
I suppose we can link this to ~#315~, I mean #302
We get a segfault from
is_converged()
function. The problem appears to be happening inside the g_proximal proximal_norm within the objective_convergence function