Intel® Optimization for Horovod* is the distributed training framework for TensorFlow*. The goal is to make distributed Deep Learning workload run faster and easier to use on Intel GPU devices. It's ...