AWS Open Source Blog
Distributed TensorFlow training using Kubeflow on Amazon EKS
Training heavy-weight deep neural networks (DNNs) on large datasets ranging from tens to hundreds of GBs often takes an unacceptably long time. Business imperatives force us to search for solutions that can reduce the training time from days to hours. Distributed data-parallel training of DNNs using multiple GPUs on multiple machines is often the right […]
