The number of elastic inference accelerators to attach to the instance.
The type of elastic inference accelerator. The possible values are eia1.medium, eia1.large, and eia1.xlarge.