Distributed TensorFlow

deep_learning

#1

Is this possible to control the distribution of delays in asynchronous training in TensorFlow?