feat: Add g7e instance types to HyperPod helm chart

[PremiumSpider]

What's changing and why?

Adding g7e instance types to the HyperPod helm chart values.yaml to enable g7e support on EKS clusters.

Without this change, users cannot install the HyperPod helm charts on EKS clusters with g7e nodes because:

• The NVIDIA device plugin won't be scheduled on g7e nodes (missing from nodeAffinity)
• The EFA device plugin won't recognize EFA-capable g7e instances (missing from supportedInstanceLabels)

Changes:

• nvidia-device-plugin nodeAffinity: added all 6 g7e sizes (2xlarge, 4xlarge, 8xlarge, 12xlarge, 24xlarge, 48xlarge)
• aws-efa-k8s-device-plugin supportedInstanceLabels: added 4 EFA-capable g7e sizes (8xlarge, 12xlarge, 24xlarge, 48xlarge)

EFA support verified via aws ec2 describe-instance-types:

Instance	EFA	Max EFA Interfaces
g7e.2xlarge	No	-
g7e.4xlarge	No	-
g7e.8xlarge	Yes	1
g7e.12xlarge	Yes	1
g7e.24xlarge	Yes	2
g7e.48xlarge	Yes	4

Before/After UX

Before: Helm chart installation on g7e EKS clusters results in NVIDIA device plugin pods not being scheduled on g7e nodes, and EFA devices not being recognized.

After: NVIDIA device plugin is correctly scheduled on all g7e nodes, and EFA device plugin recognizes EFA-capable g7e instances.

How was this change tested?

Config-only change (YAML values). Verified instance types and EFA support via aws ec2 describe-instance-types.

Are unit tests added?

N/A — config-only change, no code logic modified.

Are integration tests added?

N/A — config-only change.

Reviewer Guidelines

‼️ Merge Requirements: PRs with failing integration tests cannot be merged without justification.

One of the following must be true:

• All automated PR checks pass
• Failed tests include local run results/screenshots proving they work
• Changes are documentation-only