Amazon ECS#

Run containerized Neuron workloads on Amazon Elastic Container Service. ECS provides task-based container orchestration for inference and training on Inferentia and Trainium instances, with support for Neuron node problem detection and recovery.

Run inference on ECS

Deploy inference containers on ECS using Neuron Deep Learning Containers on Inferentia instances.

Run training on ECS

Deploy training containers on ECS using Neuron DLCs on Trainium instances.

Node problem detector for ECS

Monitor Neuron device health and automatically remediate issues on ECS clusters.