[skills] chore: rename offloading and resiliency skills with nemo mbridge prefix - #4034
Conversation
…efix Signed-off-by: Chen Cui <chcui@nvidia.com>
|
/nvskills-ci |
|
/nvskills-ci |
![claude[bot]](https://avatars.githubusercontent.com/in/1236702?s=60&v=4){kind=small}
|
|
||
| Stable docs: @docs/training/resiliency.md, @docs/training/checkpointing.md | ||
| Card: @skills/resiliency/card.yaml | ||
| Stable docs: @docs/training/nemo-mbridge-resiliency.md, @docs/training/checkpointing.md |
There was a problem hiding this comment.
Bug: This file was not renamed in this PR — docs/training/resiliency.md still exists at the old path. Same issue on lines 225, 227, 234, 285–286, 294: all nemo-mbridge-resiliency substitutions in code-anchor and verification paths point to non-existent files/dirs.
Either revert these path references to the originals (resiliency), or rename the actual docs/scripts/examples in a follow-up and reference the correct future names consistently.
| @@ -103,16 +103,16 @@ evidence: | |||
| - src/megatron/bridge/training/config.py | |||
| - src/megatron/bridge/training/utils/sig_utils.py | |||
| - src/megatron/bridge/recipes/run_plugins.py | |||
| - scripts/performance/resiliency_plugins.py | |||
| - scripts/performance/mbridge-resiliency_plugins.py | |||
| - tests/unit_tests/training/test_fault_tolerance.py | |||
| - tests/unit_tests/training/test_nvrx_straggler.py | |||
| - tests/unit_tests/training/test_inprocess_restart.py | |||
| - tests/unit_tests/recipes/test_run_plugins.py | |||
| - tests/functional_tests/training/test_nvrx_straggler.py | |||
| - tests/functional_tests/training/test_inprocess_restart.py | |||
| - tests/functional_tests/training/test_local_checkpointing.py | |||
| - examples/training_features/resiliency/fault_tolerance/ | |||
| - examples/training_features/resiliency/straggler_detection/ | |||
| - examples/training_features/mbridge-resiliency/fault_tolerance/ | |||
| - examples/training_features/mbridge-resiliency/straggler_detection/ | |||
There was a problem hiding this comment.
Bug: Two problems here:
-
Broken paths — the underlying files were not renamed in this PR.
docs/training/mbridge-resiliency.md,scripts/performance/mbridge-resiliency_plugins.py, andexamples/training_features/mbridge-resiliency/do not exist. The actual files are still atdocs/training/resiliency.md,scripts/performance/resiliency_plugins.py, andexamples/training_features/resiliency/. -
Inconsistent prefix — SKILL.md uses
nemo-mbridge-resiliencyfor these same paths, but card.yaml usesmbridge-resiliency. They should match.
Recommend reverting these evidence paths to the originals until the underlying files are actually renamed.
yaoyu-33
added
area:misc
|
Light Code Review -- Broken references in nemo-mbridge-resiliency (bug): The skill directory was renamed, but SKILL.md and card.yaml also updated references to docs, scripts, and example directories that were NOT renamed in this PR (docs/training/resiliency.md, scripts/performance/resiliency_plugins.py, examples/training_features/resiliency/). These now point to non-existent paths. Additionally, SKILL.md uses nemo-mbridge-resiliency prefix while card.yaml uses mbridge-resiliency - they should match. Recommend reverting the path references to the original names since the underlying files were not renamed. nemo-mbridge-perf-cpu-offloading looks clean. Suggested test cases: No perf tests impacted. |
|
See detailed inline comments. Summary: (1) SKILL.md and card.yaml in nemo-mbridge-resiliency updated references to docs/scripts/examples that were NOT renamed -- all those paths are now broken (e.g. docs/training/nemo-mbridge-resiliency.md does not exist, actual file is docs/training/resiliency.md). (2) SKILL.md uses nemo-mbridge-resiliency prefix while card.yaml uses mbridge-resiliency for the same paths -- inconsistent even with each other. Recommend reverting path references to originals. nemo-mbridge-perf-cpu-offloading is clean. Suggested test cases: No perf tests impacted. |
Signed-off-by: nvskills-svc-account <svc-nvskills-signing@nvidia.com>
Signed-off-by: Chen Cui <chcui@nvidia.com>
|
/nvskills-ci |
Signed-off-by: nvskills-svc-account <svc-nvskills-signing@nvidia.com>
|
/ok to test 7e449de |
Signed-off-by: Chen Cui <chcui@nvidia.com>
…idge prefix (NVIDIA-NeMo#4034) Signed-off-by: Chen Cui <chcui@nvidia.com> Signed-off-by: nvskills-svc-account <svc-nvskills-signing@nvidia.com> Co-authored-by: nvskills-svc-account <svc-nvskills-signing@nvidia.com> Signed-off-by: Vasudevan Rengasamy <vrengasamy@nvidia.com>
3 participants
Summary
nemo-mbridge-*prefix:nemo-mbridge-perf-cpu-offloadingnemo-mbridge-resiliencylicense: Apache-2.0frontmatter for these public skills.evals/evals.jsonfiles where needed for NVSkills signing.docs/skills-index.mdreferences for the renamed skills.Validation
SKILL.mdfrontmatter names.license: Apache-2.0.evals/evals.jsonfiles.SKILL.mdfiles are <= 500 lines.git diff --checkpassed.uv run --no-sync pre-commit run --all-filespassed.