-
Notifications
You must be signed in to change notification settings - Fork 1.8k
[None][test] Add post merge test for Seed-OSS-36B-Instruct #8321
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
📝 WalkthroughWalkthroughAdds an optional chat_template_kwargs parameter across evaluators, LM wrappers, and CLI commands, and threads it through to tokenizer.apply_chat_template calls. Updates MMLU and LM Eval paths, introduces corresponding CLI options, and adds integration tests and accuracy references for the Seed-OSS-36B-Instruct model. Changes
Sequence Diagram(s)sequenceDiagram
autonumber
actor User
participant CLI as CLI Command
participant Eval as Evaluator (LmEval/MMLU)
participant LM as LM Wrapper (Text/MM)
participant Tok as Tokenizer
User->>CLI: Run ... --chat_template_kwargs '{"k": "v"}'
CLI->>Eval: instantiate(..., chat_template_kwargs)
Eval->>LM: instantiate(..., chat_template_kwargs)
Note over LM: Stores self.chat_template_kwargs
alt apply_chat_template enabled
LM->>Tok: apply_chat_template(messages, add_generation_prompt, **chat_template_kwargs)
Tok-->>LM: templated_text
else disabled
LM-->>Eval: skip templating
end
LM-->>Eval: proceed with generation
Eval-->>User: results
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Pre-merge checks and finishing touches❌ Failed checks (2 warnings)
✅ Passed checks (1 passed)
✨ Finishing touches
🧪 Generate unit tests (beta)
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. Comment |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Actionable comments posted: 0
🧹 Nitpick comments (4)
tensorrt_llm/evaluate/interface.py (1)
37-39: Use typing.Dict for 3.8 compatibility; pass kwargs safely.Project targets Python 3.8+. Replace PEP 585 generics with typing.Dict to avoid 3.9+ requirement. Logic forwarding chat_template_kwargs looks good.
Apply this diff:
-from typing import Any, Iterable, List, Optional, Union +from typing import Any, Iterable, List, Optional, Union, Dict @@ - system_prompt: Optional[str] = None, - chat_template_kwargs: Optional[dict[str, Any]] = None): + system_prompt: Optional[str] = None, + chat_template_kwargs: Optional[Dict[str, Any]] = None): @@ - return llm.tokenizer.apply_chat_template(messages, - tokenize=False, - add_generation_prompt=True, - **(self.chat_template_kwargs - or {})) + return llm.tokenizer.apply_chat_template( + messages, + tokenize=False, + add_generation_prompt=True, + **(self.chat_template_kwargs or {}), + )As per coding guidelines
Also applies to: 45-46, 69-71
tensorrt_llm/evaluate/mmlu.py (1)
24-24: Adopt Dict for Python 3.8; improve JSON option parsing (lint-safe, better errors).
- Replace dict[...] with typing.Dict for 3.8 compatibility.
- The lambda callback triggers ARG005 and gives poor error messages on bad JSON. Use a small parser function and raise click.BadParameter.
Apply this diff:
+import json @@ -from typing import Any, Iterable, List, Optional, Union +from typing import Any, Iterable, List, Optional, Union, Dict @@ - system_prompt: Optional[str] = None, - chat_template_kwargs: Optional[dict[str, Any]] = None): + system_prompt: Optional[str] = None, + chat_template_kwargs: Optional[Dict[str, Any]] = None): @@ - @click.option( - "--chat_template_kwargs", - type=str, - default=None, - callback=lambda ctx, param, value: json.loads(value) if value else None, - help= - 'Chat template kwargs as JSON string, e.g., \'{"thinking_budget": 0}\'') + def _parse_json_option(ctx, param, value): + if not value: + return None + try: + return json.loads(value) + except json.JSONDecodeError as e: + raise click.BadParameter(f"Invalid JSON: {e.msg}") from e + @click.option( + "--chat_template_kwargs", + type=str, + default=None, + callback=_parse_json_option, + help='Chat template kwargs as JSON string, e.g., \'{"thinking_budget": 0}\'') @@ - system_prompt=system_prompt, - chat_template_kwargs=chat_template_kwargs) + system_prompt=system_prompt, + chat_template_kwargs=chat_template_kwargs)As per coding guidelines
Also applies to: 38-38, 141-147, 303-309, 335-342
tensorrt_llm/evaluate/lm_eval.py (2)
52-61: Use typing.Dict for Python 3.8 compatibility across constructors.Replace PEP 585 dict[...] with typing.Dict to keep 3.8 support.
Apply this diff:
- streaming: bool = False, - chat_template_kwargs: Optional[dict[str, Any]] = None): + streaming: bool = False, + chat_template_kwargs: Optional[Dict[str, Any]] = None): @@ - max_images: int = 999, - chat_template_kwargs: Optional[dict[str, Any]] = None): + max_images: int = 999, + chat_template_kwargs: Optional[Dict[str, Any]] = None): @@ - is_multimodal: bool = False, - chat_template_kwargs: Optional[dict[str, Any]] = None): + is_multimodal: bool = False, + chat_template_kwargs: Optional[Dict[str, Any]] = None):As per coding guidelines
Also applies to: 148-169, 303-332
480-485: Prefer a shared JSON parser over inline lambdas (fixes lint, better UX).Define a single _parse_json_option (returns None or parsed dict; raises click.BadParameter on invalid JSON) and reuse it for all --chat_template_kwargs options to avoid ARG005 and improve errors.
I can provide a consolidated diff to add the helper and update these options if desired.
Also applies to: 538-544, 588-594, 638-644, 684-689
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (6)
tensorrt_llm/evaluate/interface.py(2 hunks)tensorrt_llm/evaluate/lm_eval.py(15 hunks)tensorrt_llm/evaluate/mmlu.py(6 hunks)tests/integration/defs/accuracy/references/gsm8k.yaml(1 hunks)tests/integration/defs/accuracy/test_llm_api_pytorch.py(1 hunks)tests/integration/test_lists/test-db/l0_b200.yml(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:
tensorrt_llm/evaluate/interface.pytests/integration/defs/accuracy/test_llm_api_pytorch.pytensorrt_llm/evaluate/mmlu.pytensorrt_llm/evaluate/lm_eval.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:
tensorrt_llm/evaluate/interface.pytests/integration/defs/accuracy/test_llm_api_pytorch.pytensorrt_llm/evaluate/mmlu.pytensorrt_llm/evaluate/lm_eval.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:
tensorrt_llm/evaluate/interface.pytests/integration/defs/accuracy/test_llm_api_pytorch.pytensorrt_llm/evaluate/mmlu.pytensorrt_llm/evaluate/lm_eval.py
🧬 Code graph analysis (3)
tensorrt_llm/evaluate/interface.py (1)
tensorrt_llm/evaluate/lm_eval.py (2)
apply_chat_template(66-78)apply_chat_template(197-249)
tests/integration/defs/accuracy/test_llm_api_pytorch.py (5)
tests/integration/defs/accuracy/accuracy_core.py (3)
GSM8K(332-347)evaluate(184-245)evaluate(763-773)tests/integration/defs/conftest.py (1)
llm_models_root(79-93)tensorrt_llm/sampling_params.py (1)
SamplingParams(126-545)tensorrt_llm/llmapi/llm_args.py (1)
KvCacheConfig(1108-1242)tensorrt_llm/llmapi/llm.py (1)
LLM(1084-1100)
tensorrt_llm/evaluate/mmlu.py (1)
tensorrt_llm/evaluate/lm_eval.py (2)
apply_chat_template(66-78)apply_chat_template(197-249)
🪛 Ruff (0.13.3)
tensorrt_llm/evaluate/mmlu.py
306-306: Unused lambda argument: ctx
(ARG005)
306-306: Unused lambda argument: param
(ARG005)
tensorrt_llm/evaluate/lm_eval.py
483-483: Unused lambda argument: ctx
(ARG005)
483-483: Unused lambda argument: param
(ARG005)
541-541: Unused lambda argument: ctx
(ARG005)
541-541: Unused lambda argument: param
(ARG005)
591-591: Unused lambda argument: ctx
(ARG005)
591-591: Unused lambda argument: param
(ARG005)
641-641: Unused lambda argument: ctx
(ARG005)
641-641: Unused lambda argument: param
(ARG005)
687-687: Unused lambda argument: ctx
(ARG005)
687-687: Unused lambda argument: param
(ARG005)
⏰ 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 (3)
tests/integration/defs/accuracy/references/gsm8k.yaml (1)
224-225: Confirm reference vs test setup (dtype/chat template/budget).PR text cites 92.077 (BF16, B200) with chat template or small thinking budget; YAML adds 90.8. Please verify the reference matches the configuration used in the new test (thinking_budget, sampling params, device) and adjust either the YAML or test accordingly.
tests/integration/test_lists/test-db/l0_b200.yml (1)
147-147: LGTM: post-merge placement is appropriate.Long-running GSM8K eval for Seed-OSS is gated under post_merge PyTorch as intended.
tests/integration/defs/accuracy/test_llm_api_pytorch.py (1)
3583-3601: Reassess thinking_budget and max_tokens settings
- Replace thinking_budget=-1 (unlimited) with a small non-negative value (e.g. 0/32/64) to bound “thinking” time.
- Reduce max_tokens=16384 to a more realistic limit (256–512) for GSM8K to cut runtime.
Run the accuracy suite to confirm you still hit the 90.8 target; if not, tune these values accordingly.
|
/bot run --post-merge |
|
PR_Github #21213 [ run ] triggered by Bot |
|
PR_Github #21213 [ run ] completed with state |
7cfff08 to
d52b87b
Compare
|
/bot run --post-merge |
|
PR_Github #21279 [ run ] triggered by Bot |
|
PR_Github #21279 [ run ] completed with state |
d52b87b to
46585cf
Compare
Signed-off-by: Zhen Huang <145532724+zhhuang-nv@users.noreply.github.com>
Signed-off-by: Zhen Huang <145532724+zhhuang-nv@users.noreply.github.com>
aebc0c9 to
6387527
Compare
|
/bot run --post-merge |
|
PR_Github #21346 [ run ] triggered by Bot |
|
PR_Github #21346 [ run ] completed with state |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
Signed-off-by: Zhen Huang <145532724+zhhuang-nv@users.noreply.github.com>
|
/bot reuse-pipeline |
|
PR_Github #21685 [ reuse-pipeline ] triggered by Bot |
|
PR_Github #21685 [ reuse-pipeline ] completed with state |
Signed-off-by: Zhen Huang <145532724+zhhuang-nv@users.noreply.github.com>
Summary by CodeRabbit
New Features
Tests
Chores
Description
GSM8K average accuracy is 92.077 on B200 with BF16 dtype.
We cannot get this score if we do not apply chat template or limit the thinking budget to a small value. Due to the long evaluation time and common model structure, I add this test to post merge stage.
Test Coverage
accuracy/test_llm_api_pytorch.py::TestSeedOss_36B::test_auto_dtype
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 thestage-listparameter 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.mdand the
scripts/test_to_stage_mapping.pyhelper.kill
killKill all running builds associated with pull request.
skip
skip --comment COMMENTSkip 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-pipelineReuse 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.