Skip to content

chore(beep boop 🤖): Bump uv.lock (main, mcore-dev) (2026-05-22)#3934

Open
svcnvidia-nemo-ci wants to merge 1 commit into
mainfrom
bump-ci-container-2026-05-22-main-dev
Open

chore(beep boop 🤖): Bump uv.lock (main, mcore-dev) (2026-05-22)#3934
svcnvidia-nemo-ci wants to merge 1 commit into
mainfrom
bump-ci-container-2026-05-22-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 78d009f

@copy-pr-bot
Copy link
Copy Markdown

copy-pr-bot Bot commented May 22, 2026

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 yaoyu-33 added area:build Dependencies, packaging, images, and environment setup ci CI, automation, test queue, or workflow infrastructure work needs-review PR is ready for code review and waiting on a reviewer labels May 22, 2026
@yaoyu-33
Copy link
Copy Markdown
Contributor

MCore bump auto-fix status for dev:

Classification: MCore broke Bridge
Evidence: Current main / mcore-dev bump #3934 fails Launch_Unit_Tests_Core job 77372646987 in run 26282501817. The failed log shows six tests/unit_tests/training/test_tokenizer.py failures, all from 3rdparty/Megatron-LM/megatron/core/tokenizers/utils/build_tokenizer.py:106 with AttributeError: 'TokenizerConfig' object has no attribute 'make_vocab_size_divisible_by'. MCore dev 56481b0501cf7b3719e1869c495e2680ef0f3456 still reads padding attrs during tokenizer build while its TokenizerConfig no longer declares them.
Fix PR: #3945
Guards: added one local guard in src/megatron/bridge/training/tokenizers/tokenizer.py; removal TODO says to remove it when Megatron-LM tokenizer build_tokenizer either restores pad_vocab_size handling or no longer requires padding attrs on TokenizerConfig. No stale guards were removed after auditing current MCore compatibility guards.
Validation: local python3 -m py_compile src/megatron/bridge/training/tokenizers/tokenizer.py tests/unit_tests/training/test_tokenizer.py passed. CW interactive focused test passed with UV_CACHE_DIR=... PYTHONPATH=... uv run --no-sync python -m pytest tests/unit_tests/training/test_tokenizer.py -q (6 passed, 41 warnings in 4.24s). CW interactive pre-commit passed with UV_NO_SYNC=1 uv run pre-commit run --all-files. Full-sync CW attempts with plain uv run python -m pytest ... were blocked before pytest by nvidia-cusolver download failures and then a 30-minute environment setup timeout, so I used the warmed no-sync environment for the focused validation.
Next action: review fix PR #3945, then rerun CI for #3934 after the fix lands. Handoff time: 2026-05-22 08:11 PDT.

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

Labels

area:build Dependencies, packaging, images, and environment setup ci CI, automation, test queue, or workflow infrastructure work full-test-suite needs-review PR is ready for code review and waiting on a reviewer

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants