1. What Ray Is In 2016, UC Berkeley’s RISELab released a new distributed computing framework called Ray. In 2017, after the Ray paper was published, it drew broad attention across the industry; in China it was mainly Ant Group that adopted and contributed to Ray. In 2020, Ray released version

1. What TensorBoard Is TensorBoard is mainly used to monitor how a model’s various metrics change — accuracy, loss, the weight distributions of each layer, and so on. TensorBoard is a visualization tool from TensorFlow that supports visualizing scalar, text, image, audio, video, and embedding data, among other types. But PyTorch can use TensorBoard too.

1. Create the Training Script Create a training script mnist.py with the following contents: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

The DLRover version used in this article is 0.3.7 1. DLRover Operator 1.1 Starting the ElasticJob and ScalePlan Controllers Implementation code: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 // 创建 ElasticJob 的控制器 if err = controllers.NewElasticJobReconciler(mgr, masterImage).SetupWithManager(mgr); err

1. Parameter Server Architecture In the Parameter Server architecture, the nodes in the cluster are divided into two categories: parameter server nodes (Parameter Server) and worker server nodes (Worker). 1.1 Parameter Server The Parameter Server is used to store the model’s parameters. Each parameter server node is responsible for managing and updating one part of the model’s parameters, while each worker node only processes the subset of data corresponding to it.