This PR allows one to set the dtype and max memory allowed for the CovarianceMatrix---handy when response matrices start getting large.
I branched off of my other PR branch sparse_poisson by mistake, but if there's interest in merging this without that branch, it can be redone.
Tests pass on my machine with dtype=np.float64, but not with dtype=np.float32---small precision differences apparently fall outside the rtol of some of the comparisons.
flake8 pyunfoldThis PR allows one to set the
dtypeand max memory allowed for theCovarianceMatrix---handy when response matrices start getting large.I branched off of my other PR branch
sparse_poissonby mistake, but if there's interest in merging this without that branch, it can be redone.Tests pass on my machine with
dtype=np.float64, but not withdtype=np.float32---small precision differences apparently fall outside thertolof some of the comparisons.As before, CI tests here fail due to timeouts