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?
Why do you have an aversion to Google Cloud Platform? Surely that's like having an aversion to AWS.
pytorch, not tf, seems to be winning the hearts and minds of ml researchers. then again, things change quickly.
Tensorflow has a great model deployment story compared to alternatives like pytorch. However, pytorch tends to be more convenient when researching new techniques and architectures.