ParentFull threadpopinman322·I also liked the graph that showed only training loss vs Adam. Would have been nice to see validation loss as well; are we just helping Adam overfit the data?View on HN