Anirudh

@meissnereric how close is Gibbs sampling to getting merged?

@zhenwendai My use case was to build a topic model like LDA( or similar PGMs) with MXFusion. And yes, having Gibbs sampling would be good.

@laurenyu the config is different for different nodes in the cluster. Are you suggesting we do something like this - ```python # Start of train.py if getHostname() == "algo-1": #...

@laurenyu I can try this, but I am not sure it will work, because `TF_CONFIG` looks something like this - ```python os.environ['TF_CONFIG'] = json.dumps({ 'cluster': { 'worker': [] }, 'task':...

@ThomasDelteil @thomelane can you please review this example. Thanks

@juliusshufan can you address the comments and CI test failures

@seujung can you please look at @ThomasDelteil comments. I think a few them are still pending.

this module has been moved into mxnet repo. Can you please raise this issue here - https://github.com/apache/incubator-mxnet/issues

@brandonwillard can you please share an example of a large `Scan` graph with which I can replicate this issue locally? I would like to attempt to solve this bug.

@KellenSunderland could you please review this PR?