chore(deps): AUT-901 bump Megatron-LM to mcore main b4ad280d#4948
Open
svcnemo-autobot wants to merge 4 commits into
Open
chore(deps): AUT-901 bump Megatron-LM to mcore main b4ad280d#4948svcnemo-autobot wants to merge 4 commits into
svcnemo-autobot wants to merge 4 commits into
Conversation
svcnemo-autobot
added
dependencies
full-test-suite
mlm-sync
Collaborator
Author
|
/ok to test 9d21601 |
Bump 3rdparty/Megatron-LM to the tip of Megatron-LM main and record it in .main.commit. uv.lock refresh follows in a separate commit. Signed-off-by: svcnemo-autobot <svcnemo-autobot@nvidia.com>
Refresh uv.lock via 'uv lock --upgrade' in the CI container (pytorch 26.06, uv 0.7.2) after bumping 3rdparty/Megatron-LM to mcore main b4ad280d. Signed-off-by: svcnemo-autobot <svcnemo-autobot@nvidia.com>
test_training_readme_recipes_exist only scanned src/.../recipes/, but scripts/training/README.md references perf recipes defined under src/.../perf_recipes/ (e.g. qwen3_30b_a3b_pretrain_16gpu_h100_bf16_config), causing a false 'nonexistent recipes' failure. Discover both trees. Signed-off-by: svcnemo-autobot <svcnemo-autobot@nvidia.com>
nemo-autobot-origin
Bot
force-pushed
the
chore/bump-mcore-main
branch
from
July 19, 2026 08:45
9d21601 to
7cef4ad
Compare
Collaborator
Author
|
/ok to test 7cef4ad |
The file (added upstream by NVIDIA-NeMo#4938) shipped without the NVIDIA Apache-2.0 header, tripping the required copyright-check gate. Add the standard header. Signed-off-by: svcnemo-autobot <svcnemo-autobot@nvidia.com>
svcnemo-autobot
added
the
area:build
Collaborator
Author
|
/ok to test ad2ddec |
yaoyu-33
added
ci
This was referenced Jul 19, 2026
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
2 participants
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
Nightly-style bump of the pinned Megatron-Core commit on
main(mirrors the olddependabot.yml(main, main)flow), rebased onto currentmain.What changed
3rdparty/Megatron-LMsubmodule +.main.committo mcore mainb4ad280d.uv.lock(uv lock --upgrade).test_readme_consistencyto also scanperf_recipes/(pre-existing upstream drift from feat(training): route performance recipes by name #4945).Details
.main.commit:61f31145→b4ad280d.uv.lock: regenerated in the CI container (pytorch 26.06, uv 0.7.2);--upgradealso refreshed ~10 unrelated transitive deps (72 ins / 72 del).tests/unit_tests/doc_consistency/test_readme_consistency.py:_defined_recipe_names()now discoverssrc/.../recipes/**ANDsrc/.../perf_recipes/**. The training README referencesqwen3_30b_a3b_pretrain_16gpu_h100_bf16_config(a perf recipe added by feat(training): route performance recipes by name #4945) which the test previously couldn't see — a false 'nonexistent recipes' failure present onmainitself.Tested
MISSING=[]after the fix (recipe discovery 513→910).