Skip to content

chore(beep boop 🤖): Bump uv.lock (main, mcore-dev) (2026-06-27)#4539

Open
svcnvidia-nemo-ci wants to merge 1 commit into
mainfrom
bump-ci-container-2026-06-27-main-dev
Open

chore(beep boop 🤖): Bump uv.lock (main, mcore-dev) (2026-06-27)#4539
svcnvidia-nemo-ci wants to merge 1 commit into
mainfrom
bump-ci-container-2026-06-27-main-dev

Conversation

@svcnvidia-nemo-ci

Copy link
Copy Markdown
Contributor

🚀 PR to bump uv.lock in main.

🤖 This PR will be merged automatically once CI passes.

Signed-off-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
@svcnvidia-nemo-ci

Copy link
Copy Markdown
Contributor Author

/ok to test 80bc77b

@copy-pr-bot

copy-pr-bot Bot commented Jun 27, 2026

Copy link
Copy Markdown

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

@yaoyu-33

Copy link
Copy Markdown
Contributor

MCore bump auto-fix status for dev:

Classification: MCore broke Bridge
Evidence: PR #4539 Launch_Unit_Tests_Core failed in run 28286323896, job 83814573271, completed on 2026-06-27 04:33 PDT. The failed tests are tests/unit_tests/models/gemma/test_gemma4_modeling.py::TestGemma4PLEHelpers::test_patch_ple_block_threading_injects_layer_inputs_and_restores_state, tests/unit_tests/models/gemma/test_gemma4_modeling.py::TestGemma4PLEHelpers::test_patch_ple_block_threading_wraps_checkpointed_forward, and tests/unit_tests/models/gemma/test_gemma4_provider.py::TestGemma4PLEBlockThreading::test_threads_per_layer_inputs_to_each_layer; all fail with AttributeError: module 'megatron.core.transformer.transformer_block' has no attribute 'checkpointed_forward'. The bump moves 3rdparty/Megatron-LM from da42015c8033495cf6cc6523f8525fdb139a21d2 to d963266282a625dce4e9252d03ae8605d52b45da; at the old commit transformer_block.py imports module-level checkpointed_forward, while at the new commit only TransformerBlock._checkpointed_forward remains.
Fix PR: #4445 (existing open PR already covers this same dev failure; I did not open a duplicate fix PR)
Guards: existing PR #4445 adds a narrow module-level vs instance-level checkpointed_forward feature guard with TODO: remove the guard when both MCore main and dev call TransformerBlock._checkpointed_forward directly.
Validation: no new branch or tests were run for this handoff because #4445 already covers the failure. #4445 reports CW interactive validation on 2026-06-22 America/Los_Angeles for the focused Gemma4 tests and full Gemma4 unit files, plus pre-commit, all passing. Latest #4445 CI has Launch_Unit_Tests_Core passing in run 28256574194; its current blockers are L1_Launch_models_stepfun and gb200_L1_Launch_models_stepfun, not the Gemma4 unit failure in #4539.
Next action: review and land #4445 after its StepFun L1 blockers are resolved, then rerun or supersede the main/mcore-dev bump.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants