[https://nvbugs/5521949][fix] Replace test_codellama_fp8_with_bf16_lora with test_llama_3_1_8b_fp8_with_bf16_lora

[amitz-nv]

Description

The issue: The test was failing because it tried to quantize the codellama model through LLM api, by passing QuantConfig, whose support for pytorch flow is removed.

The solution: Make the FP8 model with BF16 LoRA adapter use an already quantized model. So I found Llama-3.1-8B-Instruct-FP8 with llama-3-chinese-8b-instruct-v2-lora that's with dtype of BF16.

Test Coverage

• tests/unittest/llmapi/test_llm_pytorch.py::test_llama_3_1_8b_fp8_with_bf16_lora

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

• 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

/bot [-h] ['run', 'kill', 'skip', 'reuse-pipeline'] ...

Provide a user friendly way for developers to interact with a Jenkins server.

Run /bot [-h|--help] to print this help message.

See details below for each supported subcommand.

run [--reuse-test (optional)pipeline-id --disable-fail-fast --skip-test --stage-list "A10-PyTorch-1, xxx" --gpu-type "A30, H100_PCIe" --test-backend "pytorch, cpp" --add-multi-gpu-test --only-multi-gpu-test --disable-multi-gpu-test --post-merge --extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx" --detailed-log --debug(experimental)]

Launch build/test pipelines. All previously running jobs will be killed.

--reuse-test (optional)pipeline-id (OPTIONAL) : Allow the new pipeline to reuse build artifacts and skip successful test stages from a specified pipeline or the last pipeline if no pipeline-id is indicated. If the Git commit ID has changed, this option will be always ignored. The DEFAULT behavior of the bot is to reuse build artifacts and successful test results from the last pipeline.

--disable-reuse-test (OPTIONAL) : Explicitly prevent the pipeline from reusing build artifacts and skipping successful test stages from a previous pipeline. Ensure that all builds and tests are run regardless of previous successes.

--disable-fail-fast (OPTIONAL) : Disable fail fast on build/tests/infra failures.

--skip-test (OPTIONAL) : Skip all test stages, but still run build stages, package stages and sanity check stages. Note: Does NOT update GitHub check status.

--stage-list "A10-PyTorch-1, xxx" (OPTIONAL) : Only run the specified test stages. Examples: "A10-PyTorch-1, xxx". Note: Does NOT update GitHub check status.

--gpu-type "A30, H100_PCIe" (OPTIONAL) : Only run the test stages on the specified GPU types. Examples: "A30, H100_PCIe". Note: Does NOT update GitHub check status.

--test-backend "pytorch, cpp" (OPTIONAL) : Skip test stages which don't match the specified backends. Only support [pytorch, cpp, tensorrt, triton]. Examples: "pytorch, cpp" (does not run test stages with tensorrt or triton backend). Note: Does NOT update GitHub pipeline status.

--only-multi-gpu-test (OPTIONAL) : Only run the multi-GPU tests. Note: Does NOT update GitHub check status.

--disable-multi-gpu-test (OPTIONAL) : Disable the multi-GPU tests. Note: Does NOT update GitHub check status.

--add-multi-gpu-test (OPTIONAL) : Force run the multi-GPU tests in addition to running L0 pre-merge pipeline.

--post-merge (OPTIONAL) : Run the L0 post-merge pipeline instead of the ordinary L0 pre-merge pipeline.

--extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx" (OPTIONAL) : Run the ordinary L0 pre-merge pipeline and specified test stages. Examples: --extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx".

--detailed-log (OPTIONAL) : Enable flushing out all logs to the Jenkins console. This will significantly increase the log volume and may slow down the job.

--debug (OPTIONAL) : Experimental feature. Enable access to the CI container for debugging purpose. Note: Specify exactly one stage in the stage-list parameter to access the appropriate container environment. Note: Does NOT update GitHub check status.

For guidance on mapping tests to stage names, see docs/source/reference/ci-overview.md
and the scripts/test_to_stage_mapping.py helper.

kill

kill

Kill all running builds associated with pull request.

skip

skip --comment COMMENT

Skip testing for latest commit on pull request. --comment "Reason for skipping build/test" is required. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.

reuse-pipeline

reuse-pipeline

Reuse a previous pipeline to validate current commit. This action will also kill all currently running builds associated with the pull request. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.

Summary by CodeRabbit

• Tests
  • Updated mixed-precision LoRA test to target the latest model configuration (FP8/BF16 with Llama 3.1 8B), with refreshed prompts and references.
  • Streamlined setup to improve reproducibility and reduce flakiness, including explicit resource cleanup.
  • Enhances coverage of adapter-based workflows and generation quality checks, increasing overall test reliability and confidence.

[amitz-nv]

/bot run

[coderabbitai]

📝 Walkthrough

Walkthrough

Replaced a legacy FP8/BF16 LoRA test with a new LLama‑3.1 FP8/BF16 LoRA test. The new test uses a pre-existing LoRA directory, initializes LLM with updated parameters, runs generation against multiple references, and ensures llm.shutdown() is called in a finally block.

Changes

Cohort / File(s)	Summary
LLM PyTorch tests tests/unittest/llmapi/test_llm_pytorch.py	Removed test_codellama_fp8_with_bf16_lora (QuantConfig/QuantAlgo, temp LoRA setup). Added test_llama_3_1_8b_fp8_with_bf16_lora using LLama‑3.1, FP8/BF16 LoRA with explicit lora_dir, LoraConfig/LoRARequest, LLM init with cuda_graph_config=None, generation with fixed token budget, multiple reference comparisons, and guaranteed llm.shutdown() in finally.

Sequence Diagram(s)

sequenceDiagram
  autonumber
  actor T as Test
  participant L as LLM
  participant R as LoRA Store
  participant G as Generator

  T->>L: Initialize LLM (fp8, bf16 LoRA, cuda_graph_config=None)
  T->>R: Create LoRARequest (lora_dir)
  T->>L: Provide prompts + LoRARequest
  L->>R: Load LoRA adapters
  L->>G: Prepare generation
  G-->>L: Tokens for each prompt
  L-->>T: Generated texts
  T->>T: Compare outputs to reference set(s)
  note right of T: Assert any match per prompt
  T-->>L: finally: shutdown()

Loading

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Pre-merge checks and finishing touches

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	You can run @coderabbitai generate docstrings to improve docstring coverage.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title Check	✅ Passed	The title clearly states that the PR replaces the old test with the new test and includes the NVBugs ID and fix type tag. It succinctly captures the deletion of test_codellama_fp8_with_bf16_lora and addition of test_llama_3_1_8b_fp8_with_bf16_lora. This makes the main change transparent to any reviewer scanning the history.
Description Check	✅ Passed	The PR description provides a clear explanation of the issue stemming from removed QuantConfig support, outlines the solution of using an already‐quantized model, and specifies the new FP8/BF16 LoRA model and adapter. The Test Coverage section lists the new unit test added, and the PR Checklist enumerates all relevant quality checks. The structure aligns with the required Description, Test Coverage, and PR Checklist sections in the template.

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

---

📜 Recent review details

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 80517b7 and 29fe5ef.

📒 Files selected for processing (1)

• tests/unittest/llmapi/test_llm_pytorch.py (1 hunks)

🧰 Additional context used

📓 Path-based instructions (3)

**/*.{h,hpp,hh,hxx,cpp,cxx,cc,cu,cuh,py}

📄 CodeRabbit inference engine (CODING_GUIDELINES.md)

Use only spaces, no tabs; indent with 4 spaces.

Files:

• tests/unittest/llmapi/test_llm_pytorch.py

**/*.py

📄 CodeRabbit inference engine (CODING_GUIDELINES.md)

**/*.py: Python code must target Python 3.8+.
Indent Python code with 4 spaces; do not use tabs.
Maintain module namespace when importing; prefer 'from package.subpackage import foo' then 'foo.SomeClass()' instead of importing the class directly.
Python filenames should be snake_case (e.g., some_file.py).
Python classes use PascalCase names.
Functions and methods use snake_case names.
Local variables use snake_case; prefix 'k' for variables that start with a number (e.g., k_99th_percentile).
Global variables use upper SNAKE_CASE prefixed with 'G' (e.g., G_MY_GLOBAL).
Constants use upper SNAKE_CASE (e.g., MY_CONSTANT).
Avoid shadowing variables from an outer scope.
Initialize all externally visible members of a class in the constructor.
Prefer docstrings for interfaces that may be used outside a file; comments for in-function or file-local interfaces.
Use Google-style docstrings for classes and functions (Sphinx-parsable).
Document attributes and variables inline so they render under the class/function docstring.
Avoid reflection when a simpler, explicit approach suffices (e.g., avoid dict(**locals()) patterns).
In try/except, catch the most specific exceptions possible.
For duck-typing try/except, keep the try body minimal and use else for the main logic.

Files:

• tests/unittest/llmapi/test_llm_pytorch.py

**/*.{cpp,cxx,cc,h,hpp,hh,hxx,cu,cuh,py}

📄 CodeRabbit inference engine (CODING_GUIDELINES.md)

Prepend the NVIDIA Apache-2.0 copyright header with current year to the top of all source files (e.g., .cpp, .h, .cu, .py).

Files:

• tests/unittest/llmapi/test_llm_pytorch.py

🧬 Code graph analysis (1)

tests/unittest/llmapi/test_llm_pytorch.py (6)

tests/integration/defs/conftest.py (1)

• llm_models_root (78-92)

tensorrt_llm/lora_helper.py (1)

• LoraConfig (83-102)

tensorrt_llm/executor/request.py (1)

• LoRARequest (24-53)

tensorrt_llm/llmapi/llm.py (3)

• LLM (1084-1100)
• generate (249-327)
• shutdown (741-748)

tensorrt_llm/sampling_params.py (1)

• SamplingParams (126-545)

tests/integration/defs/common.py (1)

• similar (536-538)

🪛 Ruff (0.13.3)

tests/unittest/llmapi/test_llm_pytorch.py

509-509: String contains ambiguous ？ (FULLWIDTH QUESTION MARK). Did you mean ? (QUESTION MARK)?

(RUF001)

---

510-510: String contains ambiguous ？ (FULLWIDTH QUESTION MARK). Did you mean ? (QUESTION MARK)?

(RUF001)

---

513-513: String contains ambiguous （ (FULLWIDTH LEFT PARENTHESIS). Did you mean ( (LEFT PARENTHESIS)?

(RUF001)

---

513-513: String contains ambiguous ） (FULLWIDTH RIGHT PARENTHESIS). Did you mean ) (RIGHT PARENTHESIS)?

(RUF001)

---

513-513: String contains ambiguous ， (FULLWIDTH COMMA). Did you mean , (COMMA)?

(RUF001)

---

513-513: String contains ambiguous （ (FULLWIDTH LEFT PARENTHESIS). Did you mean ( (LEFT PARENTHESIS)?

(RUF001)

⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)

• GitHub Check: Pre-commit Check

🔇 Additional comments (1)

tests/unittest/llmapi/test_llm_pytorch.py (1)

505-540: LGTM! Well-structured test with proper cleanup.

The test correctly:

• Uses try-finally to guarantee llm.shutdown() is called even if generation fails
• Tests both with and without LoRA on identical prompts to validate the adapter's effect
• Places assertions after the finally block, ensuring cleanup happens before verification
• Follows the established pattern of other LoRA tests in this file

The test structure aligns with the PR objective of using a pre-quantized FP8 model with a BF16 LoRA adapter.

Note: The Ruff warnings (RUF001) about fullwidth Chinese punctuation characters (lines 509-510, 513) are false positives—the Chinese text intentionally uses fullwidth punctuation, which is standard for Chinese writing.

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

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

[tensorrt-cicd]

PR_Github #20804 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #20804 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #15728 completed with status: 'FAILURE'

[shaharmor98]

LGTM

[amitz-nv]

/bot run

[tensorrt-cicd]

PR_Github #20888 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #20888 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #15801 completed with status: 'FAILURE'

[amitz-nv]

/bot run

[tensorrt-cicd]

PR_Github #20901 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #20901 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #15812 completed with status: 'FAILURE'