Closed Luke-Pratley closed 8 years ago
The reweighting was not correctly implemented for complex-valued input: the type of the weights should always be real-valued, and this was not done right. The wrapper for easily creating a reweighted padmm has now been updated and used in the example.
@mdavezac The reweighting method seems to be working for padmm, from what I have tested. Is it okay to close this issue? Are there any other problems left?
We need a working example of l1 reweighting in purify. In particular, we need a reweighting example that uses PADMM.
I have started an example, adapted from the sdmm reweighting sopt example. However, it does not compile for various reasons. The example I have started is https://github.com/astro-informatics/purify/blob/cpp-gridding-refactor/cpp/example/sara_padmm_random_coverage.cc
Note, I have not added it to the
CMakeLists.txt
because it does not compile.