Published November 27, 2024 | Version v1

MMGAN Model and Samples

  • 1. ROR icon University College London
  • 2. CEA Saclay

Description

This dataset is for the results of Generative modelling for mass-mapping with fast uncertainty quantification (Whitney et al. 2024).

 

We include the weights of the trained MMGAN model used within the paper, as well as our main result: The COSMOS field convergence map reconstruction made with MMGAN. For this we include the MMGAN convergence map reconstruction, the associated uncertainty map, as well as the approximate posterior samples used to build the final reconstruction.

We also include 3 pseudo-COSMOS-style mock maps, made using the kappaTNG mock weak lensing suite and (Schrabback et al. 2010) COSMOS shape catalogue.

 

The paper can be found on ArXiv: https://arxiv.org/abs/2410.24197 and the code used can be found on Github: https://github.com/astro-informatics/rcGAN/tree/master

Files

MMGAN_results.zip

Files (1.5 GB)

Name Size
md5:158abfc482107fef9b62d168d16f8241
1.5 GB Preview Download

Additional details

Software

Repository URL
https://github.com/astro-informatics/rcGAN/tree/dev
Programming language
Python
Development Status
Active