[https://nvbugs/6422343][test] Unwaive DeepSeek V3 Lite BF16 test

[yizhang-nv]

Dev Engineer Review

• Removed eight platform-specific waiver entries for DeepSeekV3Lite BF16 tests with v2_kv_cache=False and v2_kv_cache=True.
• Limited the change to tests/integration/test_lists/waives.txt.
• No code, API, performance, or error-handling changes were made.
• Waiver formatting, sorting, bug references, and duplicate checks passed.
• The change restores coverage on B200, B300, GB300, H100, and H20.

QA Engineer Review

• No test-db/ or qa/ files were modified.
• Removed eight entries from tests/integration/test_lists/waives.txt.
• The tests remain in the QA functional test list.
• The failure did not reproduce during 100 consecutive runs on H100 and H200.
• Diff, formatting, sorting, duplicate-waiver, and AST-based test-list checks passed.
• Full accuracy validation was not run because it requires model weights and a QA GPU environment.
• Verdict: needs follow-up because CBTS coverage data is unavailable.

Description

Remove all eight platform-specific waivers associated with NVBug 6422343 for the two DeepSeek V3 Lite BF16 test variants (v2_kv_cache=False and v2_kv_cache=True). The cases remain in the QA functional test list, so removing the waivers re-enables coverage on B200, B300, GB300, H100, and H20.

The NVBug investigation reports that the failure could not be reproduced in 100 consecutive runs each on H100 and H200. This is intentionally an unwaive-only change so CI and QA can surface any remaining platform-specific failures.

Test Coverage

• git diff --check
• Test-list file sorting and formatting pre-commit hooks
• Duplicate waiver check
• AST-based test-list validation

The full DeepSeek V3 Lite accuracy test was not rerun as part of this change because it requires the model weights and QA GPU environment; the repeated H100/H200 runtime validation is recorded in NVBug 6422343.

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]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 75120a1c-50a4-4ffa-85dc-3a761c09b335

📥 Commits

Reviewing files that changed from the base of the PR and between cd572e1 and 60e4320.

📒 Files selected for processing (1)

• tests/integration/test_lists/waives.txt

💤 Files with no reviewable changes (1)

• tests/integration/test_lists/waives.txt

Included review availability: Your plan includes up to 12 reviews per rolling hour; 9 remain after this review.

---

Walkthrough

The integration waiver list removes eight DeepSeekV3Lite bfloat16 skip entries for B200, B300, GB300, H100, and H20 environments.

Changes

Cohort / File(s)	Summary
Integration waiver cleanup tests/integration/test_lists/waives.txt	Removes eight DeepSeekV3Lite bfloat16 waiver entries across five GPU environments.

Estimated code review effort: 1 (Trivial) | ~2 minutes

Merge Risk: ⚪ Minimal · up to 60e43

This change only re-enables the two specified DeepSeek V3 Lite BF16 test variants across the listed platforms; no actionable merge-blocking risk remains beyond normal checks and review.

Possibly related PRs

• NVIDIA/TensorRT-LLM#17789: Removes related DeepSeekV3Lite test waiver entries.
• NVIDIA/TensorRT-LLM#17819: Removes DeepSeekV3Lite test skip entries in other environments.
• NVIDIA/TensorRT-LLM#17868: Removes a related DeepSeekV3Lite waiver entry.

Suggested reviewers: xinhe-nv

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly identifies the NVBug, test-related change, and unwaiving of the DeepSeek V3 Lite BF16 test.
Description check	✅ Passed	The description explains the change, rationale, test coverage, limitations, and checklist status using the required sections.
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.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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

[yizhang-nv]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #66700 [ run ] triggered by Bot. Commit: ee4ad51 Link to invocation

[tensorrt-cicd]

PR_Github #66700 [ run ] completed with state SUCCESS. Commit: ee4ad51
/LLM/main/L0_MergeRequest_PR pipeline #54313 completed with status: 'SUCCESS'

CI Report

Link to invocation

[coderabbitai]

Note

GitHub couldn't provide a complete incremental comparison for this pull request, so CodeRabbit is performing a full review instead. This review may take a little longer.

[yizhang-nv]

/bot run --disable-fail-fast

[coderabbitai]

Note

GitHub couldn't provide a complete incremental comparison for this pull request, so CodeRabbit is performing a full review instead. This review may take a little longer.

[tensorrt-cicd]

PR_Github #66991 [ run ] triggered by Bot. Commit: 60e4320 Link to invocation

[tensorrt-cicd]

PR_Github #66991 [ run ] completed with state SUCCESS. Commit: 60e4320
/LLM/main/L0_MergeRequest_PR pipeline #54537 completed with status: 'SUCCESS'
Pipeline passed with automatic retried tests. Check the rerun report for details.

CI Report

Link to invocation