I see that LBSGD is not listed as an optimizer with momentum in incubator-mxnet/example/image-classification/common/fit.py, but it indeed has the functions to create states for it. I was wondering if it was disabled because of some reason.
See this issue, which was closed in favor of this discussion topic: https://github.com/apache/incubator-mxnet/issues/11620#issuecomment-403648924