-
The initial is passed into the set_up method, but it is only used in this line where it is overwritten with zeros. Is that correct?
On the other hand, in their set_up method GD, ADMM, SIRT, SPDHG do …
-
In FISTA and PDHG we copy x to x_old every iteration adding unnecessary overhead.
Instead we should pre-allocate 2 containers, and store a new variable referencing each of them. Each interaction we…
-
Now that the `pylops` FISTA works on matrices of data, it makes sense that the regularization parameter can be a vector of values (one per vector of data in the data matrix).
Implementing this feat…
-
Thanks for your contribution in group lasso question! What should I do if I want to solve Group lasso logistic regression using proximal gradident descent method? Can you give me some advice?
-
https://github.com/TomographicImaging/CIL/blob/3f0cb9dd363ac386832d3034717f022c3b2952a1/Wrappers/Python/cil/optimisation/algorithms/GD.py#L105
Hi, would it be possible to add some sort of optional …
-
Could we have a default `x_init` automatically set to the zero image of appropriate size? We almost always use that and would save the user from worrying about allocating one. Of course should allow t…
-
Reconstruction with mask can be implemented with two CIL operators: `CompositionOperator` and `MaskOperator`.
If `A = ProjectionOperator(ig, ag)` is the Projection operator using astra/tigre then we …
-
Fitting without an intercept using GlmValidation results in an error that traces back to sci-kit learn's safe_sparse_dot(X, coef, dense_output=True).
https://github.com/yaglm/yaglm/blob/b8cdd4346d5…
-
I am trying to solve a very simple version of the least-squares problem in which the matrix `A` is obtained by stacking `a` copies of the identity matrix of size `n`. For the purposes of this exercise…
-
im using win fista home premium
when i start bitcoing i get error msg windows saying, error loading blkindex.dat, when i hit ok: bitcoin-qt.exe has toped working, msg from microsoft windows
and to fix…