Deep learning insights into non-universality in the halo mass function

Data used in paper Deep learning insights into non-universality in the halo mass function (DOI: doi.org/10.1093/mnras/stae1696). Model_outputs.zip contains the saved latents, predictions, ground truths, and the relevant cosmological quantities used in the analysis (see Model_outputs_readme.md). Trained_model.zip has the trained baseline IVE model. It contains the saved model weights and code needed to build and load the model weights, as well as the test dataset (see Trained_model_readme.md).