Quite obviously Google is investing to push tensor flow to mass audience.
As an ML programmer with deep aversion to Google products I have never jumped from Apache spark, deeplearning4j to tensor flow.
What's the general consensus regarding tensor flow in the ML community?