Jakob Sauer Jørgensen

We don't normally require dask do we? Can we avoid it? On Mon, 4 Mar 2024 at 15:58, Casper da Costa-Luis ***@***.***> wrote: > ***@***.**** commented on this pull request....

I was starting to look at this and see it contains some of the same code that is provided by #1550, on which changes are currently being implemented. What is...

This seems to require update of our loaders to return (up to) three different DataAcquisition objects: projections, flats and darks. If multiple flats and darks, what should the dimension holding...

Yes, should be seeded for reproducibility. On Mon, 22 Mar 2021 at 09:03, evelinaametova ***@***.***> wrote: > Assigned #806 to > @jakobsj . > > — > You are receiving...

Is this PR ready for review? If still in progress, please make draft.

Thank you very much! Great to have the minimal working example. That does look like a bug. A possible solution would be to have a default label, if no vectorgeometry...

Good points. The points you mention convince me that removing the tolerance is the best choice of action - to be more consistent with other algorithms, and because I think...

For the step size, it sounds like the "safety factor" of 0.99 is sensible. For you question 1 on AA^T having gone missing. Maybe I misunderstand, but I read the...

Eq. (4a) contains the gradient of l^*. I don't see that in the code?

For reference the link to the published version: https://link.springer.com/article/10.1007/s10915-018-0680-3 In the lines just below eq (4) it says l^* is considered equal to zero. So perhaps this implementation assumes l^*=0....