Repository navigation
docs(skills): clarify container::lts is the older LTS PyTorch base - #6008
Merged
Merged
Conversation
- Lead the dev-vs-lts section with the defining fact: `lts` pins the previous long-term-support NGC PyTorch/CUDA base, `dev` tracks the latest. The label is a backward-compat lane, not a "stable dependency subset" — the dropped extras (ModelOpt, the CUDA-13 TransformerEngine build) follow from the older base, they are not the point. - Annotate the base-image pins as latest vs older in the variants table. - Fix the mcore-cicd label cell to describe the same backward-compat intent instead of a vague "use the LTS base image". Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Signed-off-by: oliver könig <okoenig@nvidia.com>
svcnvidia-nemo-ci
marked this pull request as draft
July 23, 2026 22:55
Contributor
|
This PR has been automatically converted to draft because all PRs must start as drafts. When you are ready for review, click Ready for Review to begin the review process. This will:
See the contribution guide for more details. |
|
Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually. Contributors can view more details about this message here. |
Contributor
Author
|
/ok to test |
- Rewrite the "use dev" guidance and add an Answer-First constant so `lts` is a hard opt-in: never attach `container::lts`, build the LTS image, or run the `lts` uv group on your own initiative — only when the user explicitly asks for LTS validation, not even for a container/dependency change. - Drop the auto `container::lts` recommendation from the mcore-cicd "which label to attach" table; add it only on explicit request. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Signed-off-by: oliver könig <okoenig@nvidia.com>
Contributor
|
/nvskills-ci |
ko3n1g
marked this pull request as ready for review
July 23, 2026 23:07
Contributor
Author
|
/nvskills-ci |
shjwudp
pushed a commit
to shjwudp/Megatron-LM
that referenced
this pull request
Jul 30, 2026
…VIDIA#6008) Signed-off-by: oliver könig <okoenig@nvidia.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
terminator123
pushed a commit
to 021ai/Megatron-LM
that referenced
this pull request
Aug 3, 2026
…VIDIA#6008) Signed-off-by: oliver könig <okoenig@nvidia.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
svcnvidia-nemo-ci
pushed a commit
to dimapihtar/Megatron-LM
that referenced
this pull request
Aug 4, 2026
…VIDIA#6008) Signed-off-by: oliver könig <okoenig@nvidia.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Signed-off-by: Dmytro Pykhtar <dpykhtar@nvidia.com>
This branch was previously deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Background / motivation
The
mcore-build-and-dependencyandmcore-cicdskills describedcontainer::ltsas "stability testing" that "excludes ModelOpt and other bleeding-edge extras" /
"use the LTS base image". That buries the defining fact and reads as if
ltsis acurated stable dependency subset. It isn't —
ltsis the older,long-term-support NGC PyTorch/CUDA base image; the dropped extras are a
consequence of that older base, not the reason the lane exists. The old wording
also let an agent auto-attach
container::ltsfor any container/dep change, whichis wrong: LTS validation should require an explicit request.
Evidence (this branch):
docker/.ngc_version.dev→nvcr.io/nvidia/pytorch:26.06-py3(latest)docker/.ngc_version.lts→nvcr.io/nvidia/pytorch:25.09-py3(older / long-term support)What changed
skills/mcore-build-and-dependency/SKILL.md— lead the "dev vs lts" section withthe base-container difference; reframe the
lts"when used" cell as abackward-compat lane; annotate the pins as latest vs older.
container::ltsbehind an explicit user request — add an Answer-Firstconstant and rewrite the "use dev" guidance so
ltsis a hard opt-in: neverattach the label, build the LTS image, or run the
ltsuv group on your owninitiative, not even for a container/dependency change.
skills/mcore-cicd/SKILL.md— rewrite thecontainer::ltslabel cell for thebackward-compat intent, add the opt-in-only caveat to the Answer-First facts, and
drop the auto-
container::ltsrecommendation from the "which label to attach"table (add it only on explicit request).
Details
@pytest.mark.flaky/flaky_in_devmarker sentence was already correct(
tests/unit_tests/run_ci_test.shgatesnot flakyin the lts env andnot flaky_in_devin the dev env) — left unchanged.nvidia-resiliency-extis present indocker/lts/requirements.txt, so it isdeliberately not listed among the dropped extras; only ModelOpt and the
CUDA-13 TransformerEngine build are called out.
skill.oms.sigwill be re-attached bynv-skills-ci[bot]via
/nvskills-ci.Tested
docker/.ngc_version.*anddocker/lts/requirements.txt.🤖 Generated with Claude Code