Distributed Tensorflow

While TensorFlow can at times be verbose, it offers a tremendous amount of flexibility in terms of device placement, which allows for parallelization of training across multiple GPUs or compute nodes. In this post, we'll explore different manners of leveraging this flexibility.