[TAO-2511][Feature] Add Visual ChangeNet DINOv3 7B configuration

[nnagrajrao]

What changes are proposed in this pull request?

Add vit_7b_dinov3 to the tao-core Visual ChangeNet backbone enum and extend the downstream drift test with a Visual ChangeNet-specific 7B assertion. SegFormer remains limited to its existing five DINOv3 variants.

Why are the changes needed?

tao-core is the configuration source mirrored by downstream schemas. It must expose the same Visual ChangeNet DINOv3 variants as tao-pytorch so generated interfaces do not drift.

Related issues

JIRA: TAO-2511.

Companion implementation: tao-pytorch branch TAO-2511-frozen-dinov3-aoi.
Companion workflow/schema update: tao-skill-bank branch TAO-2511-frozen-dinov3-aoi.

Does this PR introduce any user-facing change?

Yes. Generated Visual ChangeNet configurations can select vit_7b_dinov3. Existing defaults and SegFormer options are unchanged.

How was this patch tested?

• Required TAO image, downstream DINOv3 config drift suite: 3 passed.
• Exact PR-range CI hooks: license, pylint, pydocstyle, and flake8 passed.
• git diff --check and no-Markdown checks passed.
• Independent panel review completed with no remaining blockers.

Was this patch authored or co-authored using generative AI tooling?

Generated-by: OpenAI Codex

Release note

Visual ChangeNet configuration now exposes the DINOv3 ViT-7B backbone option.

Checklist

• My commits are signed off per the DCO (git commit -s)
• I have read the contributing guidelines
• The code follows the project style, and lint and format checks pass locally
• I added or updated tests covering this change
• All applicable tests pass locally
• Inline configuration metadata is updated
• A version or changelog update is not required for this patch
• No secrets, credentials, proprietary data, or customer data are included
• I understand the contribution is licensed under the repository license

Notes for reviewers

No Markdown files are changed. The split regression expectation is intentional: five variants are shared with SegFormer, while 7B is Visual ChangeNet-only.

[tao-pr-bot]

For security reasons, CI does not run automatically on NVIDIA's runners — it must be triggered per commit.

• Comment /build to run CI on your latest commit, and make sure blossom-ci is green before merging.
• Re-run /build after each new push — CI is pinned to the head commit, so a stale run won't count.

Tip

Should this fix also ship in a release? Add a release/X.Y.Z label (matching the target release branch) and tao-cherry-pick-bot will backport it after merge.

Important

Before merging: every file needs a license header and every commit must be DCO signed-off — see the 📖 Contribution guide for details.

[nnagrajrao]

/build

[vpraveen-nv]

LGTM — enum addition is consistent with the tao-pytorch companion, and splitting the regression into shared-vs-ChangeNet-only expectations is a nice touch.