Erin Sheldon
Erin Sheldon
Here is a tar ball with config file, instcat, and a script showing how to run it https://www.cosmo.bnl.gov/www/esheldon/data/lsst/test-artifacts.tar
This appears to be because this line: https://github.com/LLNL/MuyGPyS/blob/0dad6a882048bcf885c59a2a23ce09181b7e67f4/src/MuyGPyS/_src/config.py#L8 I do have jax installed and the JaxConfig does not have `define_bool_state` Is this a version compatibility issue?
Downgrading to jax 0.4.24 fixed this
I think I figured it out. ```python import numpy as np from MuyGPyS.gp import MuyGPS from MuyGPyS.neighbors import NN_Wrapper from MuyGPyS.gp.noise import HeteroscedasticNoise from MuyGPyS.gp.kernels import Matern from MuyGPyS.gp.hyperparameter import...
Maybe something like this could be added as an example discussed in #114
I can't get this to fail even either, even with the full serialized rng. It's complex because the same atmosphere gets used for all the objects in the image, and...
I'll try that. I'll also need to repr the photon_ops in that case, since that is how the PSF is implemented
It is not easy to get a perfect reproduction. There must be some internal settings from the imsim run that affect the results. Also note I had to turn off...
I'm aware that this can't be incorporated into the galsim test suite due to use of imsim, but it wasn't obvious to me how to reconstruct the photon_ops without using...
All the `im.photons.flux` are set to `nan` in all cases