I want to use TensorFlow to solve a distributed classification problem. I need an example that explains me how can I do this.
I have installed TensorFlow on Windows 7 using Anaconda (4.6.4) and Python (3.6.4).
Even if I'm not sure if this question is best suited here or on Stack Overflow, I'll try to help
First, you need a Distributed Computing Framework and then a Deep Learning Framework adaptation which provide the capability to run on top of this Distributed Computing Platform
I have prepared a simple example for you, where
Please check this Notebook
The Distributed Training is managed via the high level TFCluster API which requires, among other things (set of params, training set, ...), a Notebook specific Spark Context (i.e. a Notebook Object providing the APIs to interact with Spark)
This Notebook shows the minimal procedure to run the Cluster and provides references to get to an actually working example
Please also have a look at TensorflowOnSpark - Conversion Guide explaining how to convert a standard Tensorflow Application into a TensorflowOnSpark one
Hope it helps