docs(examples): add ML training and CI pull-request examples

[JPZ13]

What

Add two runnable examples to the repository:

• pytorch-fashion-mnist: a four-stage ML training pipeline (prep, parallel train, register) authored with the Argo Workflows Python SDK and as a native Workflow.
• ci-pr-check: a pull-request check pipeline (checkout, parallel lint and tests, and a combined status on exit).

Both run in the free trial cluster.

Why

The docs IA refactor (pipekit/pipeline-mono#12347) adds ML Pipelines and CI/CD use-case pages. Those pages need runnable examples to point readers at.

Testing

Both workflows were submitted to the free trial cluster and completed:

• pytorch-fashion-mnist: trains three hyperparameter configs in parallel and registers the most accurate (test accuracy about 0.72). The winning config is exposed as an output parameter, so it stays visible after the pods are cleaned up.
• ci-pr-check: checkout, ruff lint, and the forecast unit test pass; the exit handler reports the combined status.

ruff, cspell, and markdownlint pass. The ML example includes a local test (test_model.py).

Notes

• PyTorch installs from download.pytorch.org (the CPU wheel), so the only Docker Hub image is python:3.11-slim. This avoids the anonymous rate limit on the trial cluster's shared egress IP.

[pipekit-ci]

✅⚠️MegaLinter analysis: Success with warnings

Descriptor	Linter	Files	Fixed	Errors	Warnings	Elapsed time
✅ JSON	jsonlint	3		0	0	0.21s
✅ JSON	prettier	1		0	0	0.42s
✅ MARKDOWN	markdownlint	11		0	0	0.82s
⚠️ MARKDOWN	markdown-table-formatter	11		1	0	0.32s
⚠️ PYTHON	isort	10		2	0	0.48s
✅ PYTHON	ruff	12		0	0	0.01s
✅ REPOSITORY	git_diff	yes		no	no	0.21s
✅ REPOSITORY	grype	yes		no	no	82.33s
✅ REPOSITORY	secretlint	yes		no	no	1.22s
✅ REPOSITORY	syft	yes		no	no	2.21s
✅ REPOSITORY	trivy-sbom	yes		no	no	0.59s
✅ REPOSITORY	trufflehog	yes		no	no	4.94s
✅ SPELL	cspell	39		0	0	5.08s
✅ YAML	prettier	9		0	0	0.76s

Detailed Issues

⚠️ PYTHON / isort - 2 errors

Linter output file not found

⚠️ MARKDOWN / markdown-table-formatter - 1 error

Linter output file not found

See detailed reports in MegaLinter artifacts

Your project could benefit from a custom flavor, which would allow you to run only the linters you need, and thus improve runtime performances. (Skip this info by defining FLAVOR_SUGGESTIONS: false)

• Documentation: Custom Flavors
• Command: npx mega-linter-runner@9.0.1 --custom-flavor-setup --custom-flavor-linters PYTHON_ISORT,PYTHON_RUFF,JSON_JSONLINT,JSON_PRETTIER,MARKDOWN_MARKDOWNLINT,MARKDOWN_MARKDOWN_TABLE_FORMATTER,REPOSITORY_GIT_DIFF,REPOSITORY_GRYPE,REPOSITORY_SECRETLINT,REPOSITORY_SYFT,REPOSITORY_TRIVY_SBOM,REPOSITORY_TRUFFLEHOG,SPELL_CSPELL,YAML_PRETTIER