[shin0343]

What's changing and why?

Handle explicit CPU and memory units during training quota allocation. Previously, CPU 500m raised a conversion error and memory 1024Mi was interpreted as 1024 GiB before clipping. Parse explicit Kubernetes units, preserve exponent notation through quota calculations, and emit fixed-point GiB quantities.

Retain unitless memory's existing GiB meaning, including values emitted by already-installed hyperpod-pytorch-job-template versions. Reject negative/non-finite/overflowing quantities before mutating the job.

Before/After UX

Before: A request of cpu=500m, memory=1024Mi failed or allocated the wrong memory amount. Small memory-only requests could also corrupt the derived CPU value.

After: The request resolves to 0.5 CPU and 1 GiB. Ki/small-memory quantities preserve their value, and existing CLI/template GiB inputs retain their allocations.

How was this change tested?

• Full unit suite on Python 3.11: 1,286 passed.
• 39 quantity cases cover installed-template conversion without replacing its resources, legacy/exponent GiB values, explicit units, memory-only derived CPU, defaults, malformed quantities and arithmetic/float overflow.
• Changed files parse with Python 3.8 grammar; new tests pass Ruff lint/format; changed files pass git-secrets and git diff --check.
• Command: PYTHONPATH=src python -m pytest --confcutdir=test/unit_tests -o addopts='' -q test/unit_tests using the prepared environment. The flag avoids the parent template-install hook; default coverage options are separated.

Are unit tests added?

Yes, 39 parameterized/public-allocation cases.

Are integration tests added?

No live AWS/EKS tests. Existing and new tests run locally without provisioning infrastructure.

Reviewer Guidelines

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

[shin0343]

The current head is 7d6b205067822b486b417e8c8a9d92d7b734cb6b. collab-check passed; wait-for-approval is still waiting before the external-contributor CI jobs can run.

Could a maintainer approve CI for this head? Local validation recorded in the PR includes 1,286 passing unit tests on Python 3.11 and 39 quantity regression cases. No live AWS/EKS validation was performed.