[https://nvbugs/6184143][fix] AutoDeploy: Fix newly added unit tests for Transformers 5.5.3

[govind-ramnarayan]

Basically make sure we are using Eager / non-cached attention in all these unit tests still after the transformers upgrade changed some args.

Summary by CodeRabbit

• Tests
  • Re-enabled previously skipped model tests for improved coverage.
  • Updated test configurations to ensure accurate model behavior verification across different scenarios.

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)

• 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 enables previously-skipped model equivalence and export tests by fixing test configuration issues and removing their SKIP waivers. DeCiLM config uses eager attention, Granite MoE configs disable generation caching, and corresponding waive-list entries are removed.

Changes

Enable DeCiLM and Granite MoE model tests

Layer / File(s)	Summary
DeCiLM test config with eager attention tests/unittest/auto_deploy/singlegpu/models/test_decilm_modeling.py	_create_small_config() assigns the config to a local variable and explicitly sets _attn_implementation to "eager" before returning, ensuring eager attention behavior in tests.
Granite MoE test config with cache disabled tests/unittest/auto_deploy/singlegpu/models/test_granite_moe_hybrid_modeling.py	Both _create_small_config and _create_small_hybrid_config now set use_cache = False to disable generation caching during model export and equivalence tests.
Remove test waivers tests/integration/test_lists/waives.txt	SKIP entries for DeCiLM and Granite MoE hybrid full model tests are removed, allowing these previously-waived tests to run.

🎯 1 (Trivial) | ⏱️ ~3 minutes

Possibly Related PRs

• NVIDIA/TensorRT-LLM#14215: Adds the same DeCiLM and Granite MoE test SKIP waivers that this PR removes.
• NVIDIA/TensorRT-LLM#14228: Modifies tests/integration/test_lists/waives.txt to manage SKIP entries for auto-deploy model tests.
• NVIDIA/TensorRT-LLM#14103: Removes multiple waived entries from the same waive list.

Suggested Reviewers

• galagam
• MrGeva
• suyoggupta

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Description check	⚠️ Warning	PR description is minimal and lacks required sections. Only a brief summary is provided; Description, Test Coverage, and PR Checklist sections are incomplete or missing substantive content.	Complete the Description section with details about why the transformers upgrade required these changes. Add Test Coverage section listing affected test files. Provide substantive completion of unchecked PR Checklist items.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly references the bug ticket, indicates a fix type, and specifically describes the main change: addressing newly added unit tests for Transformers 5.5.3.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
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.}

[govind-ramnarayan]

/bot run --stage-list "A10-Build_Docs, A10-PackageSanityCheck-PY310-UB2204, A100X-PackageSanityCheck-PY312-UB2404, A30-AutoDeploy-1, H100_PCIe-AutoDeploy-1, DGX_B200-AutoDeploy-1, A100X-PyTorch-1, DGX_H100-4_GPUs-AutoDeploy-1, DGX_B200-4_GPUs-AutoDeploy-1"

[tensorrt-cicd]

PR_Github #48969 [ run ] triggered by Bot. Commit: 1d5a322 Link to invocation

[tensorrt-cicd]

PR_Github #48969 [ run ] completed with state FAILURE. Commit: 1d5a322
/LLM/main/L0_MergeRequest_PR pipeline #38714 (Partly Tested) 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

[govind-ramnarayan]

/bot run --stage-list "A10-Build_Docs, A10-PackageSanityCheck-PY310-UB2204, A100X-PackageSanityCheck-PY312-UB2404, A30-AutoDeploy-1, H100_PCIe-AutoDeploy-1, DGX_B200-AutoDeploy-1, A100X-PyTorch-1, DGX_H100-4_GPUs-AutoDeploy-1, DGX_B200-4_GPUs-AutoDeploy-1"

[tensorrt-cicd]

PR_Github #49255 [ run ] triggered by Bot. Commit: 1d5a322 Link to invocation

[tensorrt-cicd]

PR_Github #49255 [ run ] completed with state SUCCESS. Commit: 1d5a322
/LLM/main/L0_MergeRequest_PR pipeline #38924 (Partly Tested) 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

[govind-ramnarayan]

/bot run --stage-list "A10-Build_Docs, A10-PackageSanityCheck-PY310-UB2204, A100X-PackageSanityCheck-PY312-UB2404, A30-AutoDeploy-1, H100_PCIe-AutoDeploy-1, DGX_B200-AutoDeploy-1, A100X-PyTorch-1, DGX_H100-4_GPUs-AutoDeploy-1, DGX_B200-4_GPUs-AutoDeploy-1"

[tensorrt-cicd]

PR_Github #49263 [ run ] triggered by Bot. Commit: 1d5a322 Link to invocation

[tensorrt-cicd]

PR_Github #49263 [ run ] completed with state FAILURE. Commit: 1d5a322
/LLM/main/L0_MergeRequest_PR pipeline #38930 (Partly Tested) 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

[govind-ramnarayan]

/bot run --stage-list "A10-Build_Docs, A10-PackageSanityCheck-PY310-UB2204, A100X-PackageSanityCheck-PY312-UB2404, A30-AutoDeploy-1, H100_PCIe-AutoDeploy-1, DGX_B200-AutoDeploy-1, A100X-PyTorch-1, DGX_H100-4_GPUs-AutoDeploy-1, DGX_B200-4_GPUs-AutoDeploy-1"

[tensorrt-cicd]

PR_Github #49460 [ run ] triggered by Bot. Commit: 4f930eb Link to invocation

[tensorrt-cicd]

PR_Github #49460 [ run ] completed with state SUCCESS. Commit: 4f930eb
/LLM/main/L0_MergeRequest_PR pipeline #39102 (Partly Tested) completed with status: 'SUCCESS'

CI Report

Link to invocation

[govind-ramnarayan]

/bot run --disable-fail-fast --extra-stage "DGX_B200-4_GPUs-AutoDeploy-1,DGX_H100-4_GPUs-AutoDeploy-1"

[tensorrt-cicd]

PR_Github #49507 [ run ] triggered by Bot. Commit: 4f930eb Link to invocation

[tensorrt-cicd]

PR_Github #49507 [ run ] completed with state SUCCESS. Commit: 4f930eb
/LLM/main/L0_MergeRequest_PR pipeline #39143 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

[govind-ramnarayan]

/bot run --disable-fail-fast --extra-stage "DGX_B200-4_GPUs-AutoDeploy-1,DGX_H100-4_GPUs-AutoDeploy-1"

[tensorrt-cicd]

PR_Github #49762 [ run ] triggered by Bot. Commit: fa2aa55 Link to invocation

[tensorrt-cicd]

PR_Github #49762 [ run ] completed with state SUCCESS. Commit: fa2aa55
/LLM/main/L0_MergeRequest_PR pipeline #39362 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

[govind-ramnarayan]

/bot run --disable-fail-fast --extra-stage "DGX_B200-4_GPUs-AutoDeploy-1,DGX_H100-4_GPUs-AutoDeploy-1"

[tensorrt-cicd]

PR_Github #49968 [ run ] triggered by Bot. Commit: fa2aa55 Link to invocation

[tensorrt-cicd]

PR_Github #49968 [ run ] completed with state SUCCESS. Commit: fa2aa55
/LLM/main/L0_MergeRequest_PR pipeline #39534 completed with status: 'SUCCESS'

CI Report

Link to invocation