[None][fix] Uncomment Qwen3.5 and DSR1 from model registry so that they can run f…

[taylor-yb-lee]

Summary by CodeRabbit

• New Features
  • DeepSeek-R1 model variant is now available
  • Qwen3.5 MoE models (35B and 400B variants) are now available

Description

In PR #14686, Qwen3.5 and DSR1 are commented out from model registry.
At the same time, those acc tests were unwaived and post merge test is crashing because of the missing parmeter.
This PR uncommenting those models from the registry

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)

• If PR introduces API changes, an appropriate PR label is added - either api-compatible or api-breaking. For api-breaking, include BREAKING in the PR title.

• 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

This PR adds three new model entries to the auto-deploy model registry: the DeepSeek-R1-0528 variant and two Qwen3.5 MoE models (35B and 397B). Each entry includes configuration references for dashboard, world-size mappings, model-specific YAML, and IR sharding enablement.

Changes

Model Registry Additions

Layer / File(s)	Summary
DeepSeek-R1 and Qwen3.5 MoE Model Entries examples/auto_deploy/model_registry/models.yaml	Registers deepseek-ai/DeepSeek-R1-0528 under config_id: deepseek_r1 with IR sharding and dashboard configs. Adds two Qwen3.5 MoE variants: Qwen/Qwen3.5-35B-A3B (qwen3_5_moe_35b) and Qwen/Qwen3.5-397B-A17B (qwen3_5_moe_400b), each with world-size-specific and IR-sharding configuration references.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~8 minutes

Possibly related PRs

• NVIDIA/TensorRT-LLM#14870: Unwaives DeepSeek R1 accuracy test for the same deepseek-ai/DeepSeek-R1-0528 model variant added in this registry.

Suggested reviewers

• nvchenghaoz

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 inconclusive)

Check name	Status	Explanation	Resolution
Description check	❓ Inconclusive	The description explains the issue and solution (uncommenting models to fix post-merge test crashes) but the Test Coverage section is empty, which is a required section per the template.	Complete the Test Coverage section by clearly listing the relevant acceptance tests that validate the uncommented models work correctly.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly summarizes the main change: uncommenting Qwen3.5 and DSR1 from the model registry, matching the file-level summary and PR objectives.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ 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]

🧹 Nitpick comments (1)

examples/auto_deploy/model_registry/models.yaml (1)

379-381: 💤 Low value

Clarify the intentional 397B → 400B naming in this config entry

• name: Qwen/Qwen3.5-397B-A17B maps to config_id: qwen3_5_moe_400b and yaml_extra includes qwen3.5_moe_400b.yaml.
• A search only finds qwen3.5_moe_400b.yaml (no qwen3.5_moe_397b.yaml/other 397b variants), so switching config_id to ..._397b would break the existing mapping; document the rounding convention instead.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@examples/auto_deploy/model_registry/models.yaml` around lines 379 - 381, The
entry's display name "Qwen/Qwen3.5-397B-A17B" doesn't match the config_id
"qwen3_5_moe_400b" and yaml_extra file "qwen3.5_moe_400b.yaml"; update the
models.yaml entry to add a clarifying comment or field next to name/config_id
that states the project uses a rounding/labeling convention (e.g., display shows
exact parameter count 397B while internal config and filenames use rounded 400B)
so readers know this mismatch is intentional; reference the "name", "config_id",
and "yaml_extra" values and the file "qwen3.5_moe_400b.yaml" when adding the
single-line clarification.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Nitpick comments:
In `@examples/auto_deploy/model_registry/models.yaml`:
- Around line 379-381: The entry's display name "Qwen/Qwen3.5-397B-A17B" doesn't
match the config_id "qwen3_5_moe_400b" and yaml_extra file
"qwen3.5_moe_400b.yaml"; update the models.yaml entry to add a clarifying
comment or field next to name/config_id that states the project uses a
rounding/labeling convention (e.g., display shows exact parameter count 397B
while internal config and filenames use rounded 400B) so readers know this
mismatch is intentional; reference the "name", "config_id", and "yaml_extra"
values and the file "qwen3.5_moe_400b.yaml" when adding the single-line
clarification.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: f8e519b2-3f94-443b-86ae-c186d11f1803

📥 Commits

Reviewing files that changed from the base of the PR and between fdcdcb3 and 7e01610.

📒 Files selected for processing (1)

• examples/auto_deploy/model_registry/models.yaml

[taylor-yb-lee]

/bot run --disable-fail-fast --stage-list "DGX_B200-8_GPUs-AutoDeploy-Post-Merge-1"

[tensorrt-cicd]

PR_Github #52316 [ run ] triggered by Bot. Commit: 7e01610 Link to invocation

[tensorrt-cicd]

PR_Github #52316 [ run ] completed with state SUCCESS. Commit: 7e01610
/LLM/main/L0_MergeRequest_PR pipeline #41622 (Partly Tested) completed with status: 'SUCCESS'

CI Report

Link to invocation

[galagam]

/bot skip --comment "autodeploy model registry change, affected tests passed in pipeline #52316"

[tensorrt-cicd]

PR_Github #52342 [ skip ] triggered by Bot. Commit: 7e01610 Link to invocation

[tensorrt-cicd]

PR_Github #52342 [ skip ] completed with state SUCCESS. Commit: 7e01610
Skipping testing for commit 7e01610

Link to invocation