Activity
-
-
b4edf83c · add test
-
-
Hi Arinze, I don't really have multiple experiments since I don't genrate the meausrements randomly but by using a quadrature rule (to get "moments")....
-
I am not sure there is a problem. However, I feel that you might have run multiple experiments and plotted the entire convergence results instead o...
-
-
09ed866d · fix scripts
-
-
Thanks, now I found it....
-
I checked the hosvd_gd pdf you sent. It looks fine for me. So, what did you say the problem is with the result? The algorithm implementation as you...
-
Hi Arinze, I invted you also with the other email arinze@aims.ac.za....
-
@lamBOO What optimal step-size did you use? I am a bit lost, though. Do you mean the version you have isn't fast enough? The speed might also depen...
-
-
commented on issue #1 "Low rank recovery for a kinetic problem" at Lambert Theisen / SparseTensorMoments.jl
Ahhh okay, so you say that you got a solution after waiting enough time? Maybe I should also try that.
-
commented on issue #1 "Low rank recovery for a kinetic problem" at Lambert Theisen / SparseTensorMoments.jl
I did successfully get recovery for 111111 tensors before. But this method is based on the semi-definite program(SDP), it really takes time and spa...
-
commented on issue #1 "Low rank recovery for a kinetic problem" at Lambert Theisen / SparseTensorMoments.jl
I tried the solver and it seems to work when the number of DOFs is low....
-
-
cf6ab3c9 · fix projected GD
-
-
-
517f99bb · Add projected GD and torch CP + Tucker solutions
-
-
-
c9c077bc · fix ci by blacklist
-
-
-
9d350718 · fix ci
-
-
-
ba02b657 · fix ci
-
-
-
3e1b4a0c · update pkgs
-
-
-
fd7a1720 · update
-
-
-
685795de · add projected gradient descent with hosvd projection
- ... and 1 more commit. Compare 62e929d5...685795de
-