Louis Desdoigts
Louis Desdoigts
Yah keen for the changes. Although small optimisation, rather than start and end, just calculate the end value as 'k' and linspace from -k, k
Ah yeah of course, this is perfect then. Probs wont merge this till I get some time to actually do some proper work on dLux. This isn't holding up anything...
This would be a Zodiax side fix, probably either by pre-processing the leaves into arrays or using a different equinox filter function. Will need to check how it interacts with...
Oh yeah this is a really cool idea - definitely something we build and do in due course
Batching can be done with `jax.lax.map` https://jax.readthedocs.io/en/latest/_autosummary/jax.lax.map.html
This is essentially the same issue as #264. The enforced float was originally chosen as a way to enhance the readability of the models (as arrays appear as 'f64[]'), so...
Yep good point. It might be worth a small review over what we want the wavefront numerical operations to do. I think how it is setup now is a little...
Yeah this is a known issue as you point out - I would prefer to keep the dLux side as it is, since having a float leaf allows us to...
So this function is designed to be used with a set of input parameter paths as defined in Zodiax. So once we have defined what leaves we want to take...
Ultimately it should live within the zodiax filter function, but applying before being passed to the filter function was just my hacky way around not modifying zodiax. Looking back over...