s2fft

s2fft copied to clipboard

Gaussian random coefficients from power spectra

3

This PR adds a new function `s2fft.utils.signal_generator.generate_flm_from_spectra()` that generates a stack of Gaussian $f_{lm}$ from a stack of power spectra. Closes #201. The function is implemented in terms of the...

ntessore

PyTorch gradient tests are taking up a lot of test time

1

I noticed as part of working on #277 that the gradient checks on the PyTorch implementations of the precompute transforms are very slow and infact end up constituting a significant...

matt-graham

Updating Healpix CUDA primitive

11

Adding a few updates - [x] Updating to the newest custom call API (API 4) using [FFI](https://docs.jax.dev/en/latest/ffi.html) - [x] implementing a grad rule for healpix cuda FFT - [x] Implementing...

ASKabalan

Questions related to HEALPix support

1

Hello, I have been working on interfacing s2fft with https://github.com/CMBSciPol/jax-healpy (currently only the transform for spin 0 is included, I am working on the interface towards spin=-2 and 2 and...

Magwos

Support for Healpix polarisation

3

I have been trying to reproduce the [healpix implementation of the E/B modes](https://healpix.jpl.nasa.gov/html/subroutinesnode12.htm#sub:alm2map_spin) computation using s2fft tools in this script, using the map provided by healpy for their tests available...

Magwos

Significant round trip error when using the method = "jax" and sampling = "healpix"

1

Hi , I was checking the round trip errors in a similar manner to that of the tutorial notebooks with method =" jax" and sampling= 'healpix.' I noticed that the...

sijil-jose

Check autodiff and batching support for `healpix_fft_cuda` primitive and add if needed

3

I think the primitive added in #204 may not support automatic differentiation and batching transforms as we did not define Jacobian vector product and transpose operations (for autodiff) and a...

matt-graham

Add notebook demonstrating equivalence of equiangular samplings

2

jasonmcewen

Support additional cylindrical (aka equiangular) sampling schemes

6

In CMB data analysis (aiming at probing small angular scales), an interest has been growing towards cylindrical griding, named CAR in https://phy-act1.princeton.edu/public/saiola/act_dr4_C20.pdf or ECP in https://www.aanda.org/articles/aa/full_html/2013/08/aa21575-13/aa21575-13.html. It corresponds to a...

Magwos

Add support for the optimal sampling scheme outlined in [this paper](https://arxiv.org/abs/1403.4661). For bandlimit L, this scheme requires the optimal number of L^2 samples (matching the dimensionality of the bandlimited signal...

jasonmcewen