FAQs¶

Which papers should I cite when using galsbi?¶

You can find out which papers to cite by using the following command in Python:

from galsbi import GalSBI
model = GalSBI("model_name")
model.cite()

This will print the bibtex entries of the papers that should be cited when using your configuration.

Can I use the emulator with any version of libraries (e.g. tensorflow, jax)?¶

The emulators are trained with the following versions of the libraries:

• tensorflow / keras version 2.12

• sklearn version 1.5.2

• jax version 0.4.31

• pzflow version 3.1.3

We tested that the emulator is working correctly for the following versions of the libraries:

• tensorflow version 2.12 to 2.18

• keras version 2.12 to 3.5

• sklearn version 1.2 to 1.5

• jax version 0.4.6 to 0.4.35

• pzflow version 3.1.0 to 3.1.3

Newer versions or older versions of the libraries might work as well, but we cannot guarantee that. Note that depending on your installed software, different emulators are loaded in the default config. If you adapt configs, make sure that you are using the correct emulator. If you encounter any problems or you need a specific library version with which the emulator is not working, please contact the developers. Retraining the emulator with your library version is potentially possible if no other solution can be found.

I run the code for the first time and it is taking a long time and printing a lot of messages. Is this normal?¶

Yes, this is normal. The first time you run the code, it will compile the necessary C code for PyCosmo and cache it. This process can take a few minutes, depending on your system. After the first run, the code will run much faster.

I get an error when I run the code for the first time during the compilation of PyCosmo. What should I do?¶

If you get an error when you run the code for the first time during the compilation of PyCosmo, (e.g. ModuleNotFoundError: No module named ‘_wrapper_1db8b055_fc3ec’), something went wrong during the compilation of the code. This can normally be resolved by deleting the cache and recompiling the code. To do this, run the following commands:

cd /path/to/cache
rm -rf PyCosmo
rm -rf gsl
rm -rf libf2c
rm -rf sympy2c

The cache is located under ~/Library/Cache on macOS and ~/_cache on Linux. After deleting the cache, recompile the code by running the following python code:

import PyCosmo
PyCosmo.build()

This should resolve the issue. If you still encounter problems, please contact the developers.