[https://nvbugs/5813192][fix] Make trust_remote_code opt-in in MultimodalModelRunner

[yibinl-nvidia]

Summary by CodeRabbit

• New Features
  • Added --trust_remote_code CLI flag for multimodal model loading, enabling users to control remote code execution during model initialization (disabled by default for enhanced security).

Description

Test Coverage

PR Checklist

Please review the following before submitting your PR:

• PR description clearly explains what and why. If using CodeRabbit's summary, please make sure it makes sense.

• PR Follows TRT-LLM CODING GUIDELINES to the best of your knowledge.

• Test cases are provided for new code paths (see test instructions)

• Any new dependencies have been scanned for license and vulnerabilities

• CODEOWNERS updated if ownership changes

• Documentation updated as needed

• Update tava architecture diagram if there is a significant design change in PR.

• The reviewers assigned automatically/manually are appropriate for the PR.

• Please check this after reviewing the above items as appropriate for this PR.

GitHub Bot Help

To see a list of available CI bot commands, please comment /bot help.

[coderabbitai]

📝 Walkthrough

Walkthrough

A CLI flag --trust_remote_code is added to the argument parser, and the corresponding runtime code is updated to use this configurable flag via getattr instead of always hardcoding trust_remote_code=True when loading Hugging Face models, tokenizers, and processors.

Changes

Cohort / File(s)	Summary
CLI Argument Definition examples/models/core/multimodal/utils.py	Added --trust_remote_code boolean flag (action='store_true', default=False) to the argument parser in add_common_args().
Runtime Configuration tensorrt_llm/runtime/multimodal_model_runner.py	Updated model, tokenizer, and processor loading to use configurable trust_remote_code via getattr(self.args, "trust_remote_code", False) instead of hardcoding trust_remote_code=True across multiple Hugging Face loading call sites.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

🚥 Pre-merge checks | ✅ 1 | ❌ 2

❌ Failed checks (2 warnings)

Check name	Status	Explanation	Resolution
Description check	⚠️ Warning	The PR description is incomplete. It contains only the template structure with placeholder comments and a checked checklist item, but lacks any actual description of the issue, solution, or test coverage information.	Provide a detailed description of the issue being fixed, the solution implemented, and list the relevant tests that validate these changes.
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (1 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly and specifically describes the main change: making trust_remote_code opt-in in MultimodalModelRunner, and properly includes the NVBugs ticket and fix type.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@examples/models/core/multimodal/utils.py`:
- Around line 130-136: The return parser is incorrectly indented inside the
parser.add_argument(...) call causing a syntax error; move the return statement
out to the same indentation level as the parser.add_argument(...) call so that
return parser executes after the add_argument call (locate the
parser.add_argument invocation for '--trust_remote_code' and un-indent the
return parser to be outside that call).

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 52b92403-9ee6-4ffd-bd92-919bc661cc65

📥 Commits

Reviewing files that changed from the base of the PR and between b427d3b and d5b0ce4.

📒 Files selected for processing (2)

• examples/models/core/multimodal/utils.py
• tensorrt_llm/runtime/multimodal_model_runner.py

[yibinl-nvidia]

/bot run

[tensorrt-cicd]

PR_Github #41571 [ run ] triggered by Bot. Commit: 555dbf1 Link to invocation

[tensorrt-cicd]

PR_Github #41571 [ run ] completed with state SUCCESS. Commit: 555dbf1
/LLM/main/L0_MergeRequest_PR pipeline #32482 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

• Please check the failed tests and fix your PR
• If you cannot view the failures, ask the CI triggerer to share details
• Once fixed, request an NVIDIA team member to trigger CI again

Link to invocation

[yibinl-nvidia]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #41685 [ run ] triggered by Bot. Commit: 555dbf1 Link to invocation

[tensorrt-cicd]

PR_Github #41685 [ run ] completed with state SUCCESS. Commit: 555dbf1
/LLM/main/L0_MergeRequest_PR pipeline #32587 completed with status: 'SUCCESS'

CI Report

Link to invocation