You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
One option would be to try to setup a PRIYA archive ignoring the fact that there are two different resolutions, but this could be bad.
Alternatively one could try to modify LaCE to be able to work with multiple resolutions (maybe the resolution could be a label in the training, a 7th emulator parameter). Looks like a lot of work though.
So not clear how to proceed in the short term, but I'm adding the Issue here to keep this in mind
The text was updated successfully, but these errors were encountered:
Oa simpler option, of course, is to only use the "low fidelity" resolutions in the training. There are 48 of these, so it is a very large training set. This could be particularly useful for a 3D emulator (their box size is 120 Mpc/h), but we should be careful that these have a resolution that is even lower than our Gadget sims. An interesting test regardless!
It looks like the files are available: https://github.com/mafern/InferenceLyaData
Unfortunately, the emulator itself is not yet public: https://github.com/sbird/lya_emulator_full
Options:
So not clear how to proceed in the short term, but I'm adding the Issue here to keep this in mind
The text was updated successfully, but these errors were encountered: