Require --instance-type when specifying accelerator resources (#317)

[FarhanTejani]

What's changing and why?

Closes #317 (the --node-count mutual exclusivity part was fixed in #383)

When --accelerators or --accelerators-limit is specified without --instance-type, the CLI silently produces nvidia.com/gpu: 0 in the k8s resource spec instead of the user's requested value. This happens because _set_default_accelerators_val needs the instance type to determine the accelerator key (nvidia.com/gpu vs aws.amazon.com/neuroncore), and when it can't, it silently returns None for both values.

This change adds an early validation that raises a clear error when accelerators are specified without --instance-type, and removes the dead code path in _get_limits that hardcoded nvidia.com/gpu: 0.

Before/After UX

Before:

$ hyp create hyp-pytorch-job --accelerators 4
# Silently generates: nvidia.com/gpu: '0'

After:

$ hyp create hyp-pytorch-job --accelerators 4
# Error: --instance-type is required when specifying accelerator resources

How was this change tested?

• Unit tests pass

Are unit tests added?

Updated 3 tests that asserted the old behavior (nvidia.com/gpu: 0 on CPU-only / invalid instances)

Are integration tests added?

N/A

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

[mollyheamazon]

Can you reenable these unit tests as part of your fix:

sagemaker-hyperpod-cli/test/unit_tests/training/test_pytorch_job_template_model.py

Lines 6 to 47 in 4406d08

	class TestPyTorchJobConfigEFA(unittest.TestCase):
	"""Test EFA resource allocation in PyTorchJobConfig"""
	# def test_single_node_no_efa(self):
	# """Test that single-node jobs don't get EFA resources"""
	# config = PyTorchJobConfig(
	# job_name="test-single-node",
	# image="pytorch:latest",
	# node_count=1,
	# accelerators=2,
	# instance_type="ml.p4d.24xlarge"
	# )
	# job = config.to_domain()
	# container = job.replicaSpecs[0].template.spec.containers[0]
	# # Should not have EFA resources
	# self.assertNotIn("vpc.amazonaws.com/efa", container.resources.requests)
	# self.assertNotIn("vpc.amazonaws.com/efa", container.resources.limits)
	# # Should have GPU resources
	# self.assertEqual(container.resources.requests["nvidia.com/gpu"], "2")
	# def test_multi_node_with_efa(self):
	# """Test that multi-node jobs automatically get EFA resources"""
	# config = PyTorchJobConfig(
	# job_name="test-multi-node",
	# image="pytorch:latest",
	# node_count=4,
	# accelerators=8,
	# instance_type="ml.p4d.24xlarge"
	# )
	# job = config.to_domain()
	# container = job.replicaSpecs[0].template.spec.containers[0]
	# # Should have EFA resources
	# self.assertEqual(container.resources.requests["vpc.amazonaws.com/efa"], "1")
	# self.assertEqual(container.resources.limits["vpc.amazonaws.com/efa"], "1")
	# # Should also have GPU resources
	# self.assertEqual(container.resources.requests["nvidia.com/gpu"], "8")

sagemaker-hyperpod-cli/test/unit_tests/training/test_pytorch_job_template_model.py

Lines 114 to 150 in 4406d08

	# def test_multi_node_with_memory_and_cpu(self):
	# """Test EFA with other resource types"""
	# config = PyTorchJobConfig(
	# job_name="test-multi-resources",
	# image="pytorch:latest",
	# node_count=2,
	# accelerators=4,
	# vcpu=16.0,
	# memory=64.0,
	# instance_type="ml.p4d.24xlarge"
	# )
	# job = config.to_domain()
	# container = job.replicaSpecs[0].template.spec.containers[0]
	# # Should have all resources including EFA
	# self.assertEqual(container.resources.requests["vpc.amazonaws.com/efa"], "1")
	# self.assertEqual(container.resources.requests["nvidia.com/gpu"], "4")
	# self.assertEqual(container.resources.requests["cpu"], "16.0")
	# self.assertEqual(container.resources.requests["memory"], "64.0Gi")
	# def test_accelerators_without_instance_type(self):
	# """Test that accelerators work without instance_type (fixes the main issue)"""
	# config = PyTorchJobConfig(
	# job_name="test-no-instance-type",
	# image="pytorch:latest",
	# accelerators=4
	# # No instance_type specified
	# )
	# job = config.to_domain()
	# container = job.replicaSpecs[0].template.spec.containers[0]
	# # Should respect accelerators value even without instance_type
	# self.assertEqual(container.resources.requests["nvidia.com/gpu"], "4")
	# # Limits should default to "0" since accelerators_limit not specified
	# self.assertEqual(container.resources.limits["nvidia.com/gpu"], "0")

They were commented out previously for a merge conflict resolution.

[FarhanTejani]

Re-enabled all 4 tests with corrected assertions and refactored to use the full pipeline. Also moved the validation before the early return per the inline comment.