Matlab implementation of the Adam stochastic gradient descent optimisation algorithm
Hi Dylan, thanks for sharing. One question regarding your implementation; why are you rejecting bad steps or what's your motivation for that? I guess that the original Adam paper is not suggesting it.
This issue appears to be discussing a feature request or bug report related to the repository. Based on the content, it seems to be still under discussion. The issue was opened by mh510 and has received 1 comments.