Why not merging this work in kymatio?

[edouardoyallon]

Hi,
I read your algorithm, I think the acceleration is super dependent on the resources available, and probably, the way the tree of scattering coefficient should be computed could be automated, depending on them. I'm not sure to understand the rational to use our low level routines to derive the same algorithm in a different package. It could be nice to merge those accelerations in kymatio, couldn't it?
Best,
E

[jecampagne]

Hi!
Erwan Allys point me your work, and I agree with @edouardoyallon and would support to use kymatio. Notice that the JAX backend would be a very nice by-product, indeed.
Best.
JE