[tylerpoon]

• I, the PR author, have personally reviewed every line of this PR.

What does this PR do?

Support initializing the language component of a multimodal model from a language-only checkpoint while leaving vision parameters locally initialized. Preserve legacy checkpoint enum deserialization compatibility and add regression tests.

Replaces #7251, which was automatically closed when its bot-managed base branch was deleted after the closure of #7250. The original discussion remains in #7251. This replacement reuses the original source branch and commits without code changes.

Production run scripts, data YAMLs, checkpoint conversion tooling, and separate CI compatibility fixes are outside this series.

Dependencies and review scope

Depends on #7315. Merge #7315 first, then this PR. #7249 has already merged.

Both PRs intentionally target main, not bot mirror branches. The author will manage merge order. Until #7315 is integrated and this branch is refreshed, the diff includes its prerequisite vision context-parallelism/metadata changes as well as this PR's checkpoint-initialization and legacy-deserialization changes. The retained history may also show inherited #7249 changes because that parent was squash-merged.

Validation

• Reuses the existing signed and signed-off head b48769523b47732e89ec0110baa3afb4a685eee3.
• Prior local Black, isort, pylint, Ruff, and Python syntax checks passed for the original branch; mypy was unavailable locally.
• Combined-stack Super SFT validation is not an isolated-PR test result. Creating this replacement does not add a new test result or establish that current-main integration is green.
• Kept as a draft pending dependency integration, review, and CI. No functional-test label or manual full-CI trigger is requested, to limit CI load. Reviewers are selected by the author.

[copy-pr-bot]

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.

[tylerpoon]

/ok to test 7f47fac

[nemo-automation-bot]

🔄 Merge queue validation started!

You can track the progress here: https://github.com/NVIDIA/Megatron-LM/actions/runs/35655152711