Amazon EC2

Run training and inference workloads directly on Amazon EC2 instances with Neuron hardware. Use a Deep Learning AMI for the fastest setup, or install the Neuron SDK manually on a supported base AMI.

Train on EC2

Set up a Trn1 instance with a Deep Learning AMI and run your first training job using PyTorch.

Run inference on EC2 (Inf2, Trn1)

Compile and deploy models for inference on Inf2 and Trn1 instances.

Run inference with DLC on EC2

Pull a Neuron Deep Learning Container and run inference on EC2 with Docker.

Run inference on EC2 (Inf1)

Compile and deploy models for inference on Inf1 instances.