@aaronmarkham I agree, but after a previous PR[1] where @indhub had written a Model parallelism tutorial, we had an offline discussion and decided that the penalty incurred by doing model parallelism when you should actually be doing data parallelism outweighed the benefit for the few users that would actually need model parallelism.
[1] https://github.com/apache/incubator-mxnet/pull/10568/files [ Full content available at: https://github.com/apache/incubator-mxnet/pull/12298 ] This message was relayed via gitbox.apache.org for [email protected]
