[https://nvbugs/6272673][chore] Unwaive DeepSeek V3 Lite NVFP4 test

[yuxianq]

Summary by CodeRabbit

• Tests
  • Updated test waivers for DeepSeek V3 Lite accuracy scenarios using the CUTLASS backend.
  • Adjusted coverage for tensor and pipeline parallel configurations, including FP8 key/value cache settings.

Description

Remove the NVBug 6272673 waiver for the DeepSeek V3 Lite NVFP4 four-GPU test so the configuration runs in CI again.

The reported intermittent failure could not be reproduced in 100 consecutive executions on B300 x4.

Test Coverage

• Built TensorRT-LLM for B300/SM103.
• Ran the following test in 100 fresh pytest processes across two B300 x4 nodes; all 100 executions passed:
  accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_nvfp4_4gpus[moe_backend=CUTLASS-mtp_nextn=0-tp4-fp8kv=True-attention_dp=True-cuda_graph=True-overlap_scheduler=True-low_precision_combine=False-torch_compile=True]
• Pre-commit waiver validation, duplicate detection, source-entry validation, and DCO checks passed.

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.

[yuxianq]

/bot run --disable-fail-fast

[coderabbitai]

📝 Walkthrough

Walkthrough

Updates integration-test SKIP waivers for TestDeepSeekV3Lite::test_nvfp4_4gpus across CUTLASS parallelism and fp8kv configurations.

Changes

DeepSeekV3Lite waiver updates

Layer / File(s)	Summary
Adjust NVFP4 waiver entries tests/integration/test_lists/waives.txt	Removes the tp4 with fp8kv=True waiver and adds waivers for tp2pp2 with both fp8kv values and tp4 with fp8kv=False.

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

Possibly related PRs

• NVIDIA/TensorRT-LLM#15774: Updates overlapping DeepSeekV3Lite integration-test waiver entries.

Suggested reviewers: dongxuy04

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
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.
Title check	✅ Passed	The title matches the change and uses the repository's required ticket/type format.
Description check	✅ Passed	The description includes the required Description, Test Coverage, and PR Checklist sections with specific details.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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

[tensorrt-cicd]

PR_Github #59336 [ run ] triggered by Bot. Commit: 0a11b84 Link to invocation

[tensorrt-cicd]

PR_Github #59336 [ run ] completed with state SUCCESS. Commit: 0a11b84
/LLM/main/L0_MergeRequest_PR pipeline #47817 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

CI Agent Failure Analysis

Link to invocation

[yuxianq]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #59396 [ run ] triggered by Bot. Commit: 0a11b84 Link to invocation

[tensorrt-cicd]

PR_Github #59396 [ run ] completed with state SUCCESS. Commit: 0a11b84
/LLM/main/L0_MergeRequest_PR pipeline #47870 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

CI Agent Failure Analysis

Link to invocation

[yuxianq]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #59457 [ run ] triggered by Bot. Commit: 0a11b84 Link to invocation

[tensorrt-cicd]

PR_Github #59457 [ run ] completed with state SUCCESS. Commit: 0a11b84
/LLM/main/L0_MergeRequest_PR pipeline #47925 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

CI Agent Failure Analysis

Link to invocation

[yuxianq]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #59588 [ run ] triggered by Bot. Commit: 97895f5 Link to invocation

[tensorrt-cicd]

PR_Github #59588 [ run ] completed with state SUCCESS. Commit: 97895f5
/LLM/main/L0_MergeRequest_PR pipeline #48032 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

CI Agent Failure Analysis

Link to invocation

[yuxianq]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #59849 [ run ] triggered by Bot. Commit: 9b2b8dd Link to invocation

[tensorrt-cicd]

PR_Github #59849 [ run ] completed with state SUCCESS. Commit: 9b2b8dd
/LLM/main/L0_MergeRequest_PR pipeline #48255 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

CI Agent Failure Analysis

Link to invocation

[yuxianq]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #59874 [ run ] triggered by Bot. Commit: 9b2b8dd Link to invocation

[tensorrt-cicd]

PR_Github #59874 [ run ] completed with state SUCCESS. Commit: 9b2b8dd
/LLM/main/L0_MergeRequest_PR pipeline #48277 completed with status: 'SUCCESS'

CI Report

Link to invocation