Workload Example
Teaching notebook that introduces how Ray Data feeds distributed Ray Train workloads through preprocessing, streaming sharding, and iter_torch_batches().