Muawiz Sajjad Chaudhary
Muawiz Sajjad Chaudhary
@eugenium when your back from vacation please update the archive version! Thanks to @janden @edouardoyallon @eickenberg for making sure I'm on the paper : )
@eugenium Thanks for fixing it. Unfortunately, seems like [Joakim's](https://arxiv.org/abs/1812.11214) last name is undercased. Can we get this fixed?
This would be sick. Let's aim for a 0.4 release.
#817 is not really an instance of the examples failing. The first instance is rather an example of the documentation not being created on a machine with GPU. The scattering...
We shouldn't need to check if numpy arrays are contiguous, me thinks. There seems to be multiple definitions for contiguity in NumPy.
It looks like we exclude skcuda backend if its not available. Similarly, we don't test on GPU device if there is no GPU. @edouardoyallon @eickenberg @janden reopen if you disagree.
Hi Dr. Oyallon, it appears that our examples/documentation are ran on a CPU machine. In the code for the scattering inversion if the machine is a gpu we set the...
@edouardoyallon @anakin-datawalker I'm not sure what is the fix to to scattering disk. do you have an image of what is supposed to be shown there/whats the correct/expected output? It...
I see your concern. I think fixing the rotation being off by pi/4 should be easy enough, but fixing the oscillations maybe a bit hard. I'll see what I can...
@edouardoyallon @anakin-datawalker This seems hard. It seems like the ordering might be different than what's shown in Bruna's paper. The oscillations are a bit confusing. Is the ordering of our...