Thank you for your interest in our papers:
Hector J. Hortua, et.al Parameters Estimation for the Cosmic Microwave Background with Bayesian Neural Networks arXiv:1911.08508, https://arxiv.org/abs/1911.08508
Hector J. Hortua, et.al Constraining the reionization history using Bayesian normalizing flows,Mach. Learn.: Sci. Technol. 1 035014, 2020 https://doi.org/10.1088/2632-2153/aba6f1
Please consider citing the paper when any of the material is used for your research.
Data_generator provide a dataset for either CMB Temperature(+Polarization) or 21cm maps. Additionally, the DropourBNN_example.py shows an example of BNN using Dropout. For a more general treatment of BNN, download the Argo library https://github.com/rist-ro/argo
Chains.zip and PS_POL...npy, PS_TT...npy are the chains generated from Cobaya and BNNs(Argo Lib.) respectively reported in the Sec.II of the paper "Parameters Estimation for the Cosmic Microwave Background with Bayesian Neural Networks".