leezu commented on issue #16900: introduce gradient update handler to the base estimator URL: https://github.com/apache/incubator-mxnet/pull/16900#issuecomment-558438234 @ptrendx the estimator API is currently experimental and is not yet up to more complex real-world use-cases. @liuzh91 is currently converting GluonNLP training scripts to make use of the Estimator API and ran into several remaining shortcomings of the API. We shouldn't release 1.6 with a known broken version of Estimator *if fixes are available*. Thus if this PR is approved and merged to master in time for the 1.6 release, it may be good to backport. Given the scope of the estimator API and the current shortcomings we can consider these PRs as bugfixes. If we don't fix the API with the 1.6 release, we may be bound to the broken API due to backwards compatibility commitments. Do you think backporting the fix would be reasonable or do you have concerns? When do you plan to tag a release candidate? These extends the fixes commited started November 2019 https://github.com/apache/incubator-mxnet/commits/master/python/mxnet/gluon/contrib/estimator.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
