Fully compatible with Tensorflow 2
Model training/inference should be much faster by using Tensorflow v2 eager execution (see: https://github.com/tensorflow/tensorflow/issues/33024#issuecomment-551184305)
Improved continuous integration testing with Github Actions, now actually test models learn properly with real world data instead of checking no syntax error with random data
Support sample_weight in all losss functions and training
Improved catalog coordinates matching
New documentation webpages
~15% faster in Bayesian neural network inference by using parallelized loop
Loss/metrics functions and normalizer now check for NaN too
Updated many of notebooks to be compable with the latest Tensorflow