[#2511][fix] eagle: qwen2 capture hidden states#10091
[#2511][fix] eagle: qwen2 capture hidden states#10091jieli-matrix merged 1 commit intoNVIDIA:mainfrom
Conversation
📝 WalkthroughWalkthroughThe PR adds speculative execution metadata support to Qwen model components by introducing an optional Changes
Estimated code review effort🎯 2 (Simple) | ⏱️ ~12 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 |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Nitpick comments (1)
tensorrt_llm/_torch/models/modeling_qwen.py (1)
152-152: Consider adding docstrings for the new parameter.While the
spec_metadataparameter is self-explanatory, adding docstrings would improve API documentation and help users understand when and how to use this parameter for speculative decoding.Example for one of the forward methods:
def forward( self, position_ids: torch.IntTensor, hidden_states: torch.Tensor, attn_metadata: AttentionMetadata, residual: Optional[torch.Tensor], mrope_config: Optional[Tuple[torch.Tensor, int]] = None, spec_metadata: Optional[SpecMetadata] = None, **kwargs, ) -> torch.Tensor: + """ + Args: + ... + spec_metadata: Optional metadata for speculative decoding, + used to capture hidden states during forward pass. + """Also applies to: 213-213, 260-260
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
tensorrt_llm/_torch/models/modeling_qwen.py(7 hunks)
🧰 Additional context used
📓 Path-based instructions (2)
**/*.py
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
**/*.py: Code developed for TensorRT-LLM should conform to Python 3.8+
Indent Python code with 4 spaces. Do not use tabs
Always maintain the namespace when importing, even if only one class or function from a module is used (e.g., usefrom package.subpackage import foo; foo.SomeClass()instead offrom package.subpackage.foo import SomeClass)
Python filenames should use snake_case (e.g.,some_file.py)
Python classes should use PascalCase (e.g.,class SomeClass)
Python functions and methods should use snake_case (e.g.,def my_awesome_function():)
Python local variables should use snake_case. Prefix 'k' for variable names that start with a number (e.g.,k_99th_percentile = ...)
Python global variables should use upper snake_case and prefix 'G' (e.g.,G_MY_GLOBAL = ...)
Python constants should use upper snake_case (e.g.,MY_CONSTANT = ...)
Avoid shadowing variables declared in an outer scope in Python code
Initialize all externally visible members of a Python class in the constructor
For Python interfaces that may be used outside a file, prefer docstrings over comments
Python comments should be reserved for code within a function, or interfaces that are local to a file
Use Google style docstrings for Python classes and functions, which can be parsed by Sphinx
Python attributes and variables can be documented inline with attribute docstrings that will be rendered under the class docstring
Avoid using reflection in Python when functionality can be easily achieved without reflection
When using try-except blocks in Python, limit the except to the smallest set of errors possible
When using try-except blocks in Python to handle multiple possible variable types (duck-typing), keep the body of the try as small as possible, using the else block to implement the logic
Files:
tensorrt_llm/_torch/models/modeling_qwen.py
**/*.{h,hpp,cpp,cc,cxx,cu,py}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
All TensorRT-LLM source code files should contain an NVIDIA copyright header with the current year, including .cpp, .h, .cu, .py, and other source files
Files:
tensorrt_llm/_torch/models/modeling_qwen.py
🧬 Code graph analysis (1)
tensorrt_llm/_torch/models/modeling_qwen.py (1)
tensorrt_llm/_torch/speculative/interface.py (1)
SpecMetadata(178-353)
🔇 Additional comments (7)
tensorrt_llm/_torch/models/modeling_qwen.py (7)
21-21: LGTM!The import follows the project's conventions and is correctly placed with other local imports from the parent package.
152-152: LGTM!The optional
spec_metadataparameter maintains backward compatibility while enabling speculative decoding support.
176-178: LGTM!The hidden state capture logic is correctly implemented with proper null checking and matches the
maybe_capture_hidden_statessignature from the SpecMetadata interface.
213-213: LGTM!The
spec_metadataparameter is correctly added to the QwenModel forward signature.
232-233: LGTM!The
spec_metadatais correctly propagated to each decoder layer, enabling layer-wise hidden state capture.
260-260: LGTM!The
spec_metadataparameter is correctly added to the top-level Qwen2ForCausalLM forward signature.
269-269: LGTM!The
spec_metadataparameter is correctly forwarded to the underlying model, completing the parameter threading through the entire call chain.
|
/bot run |
svc-trtllm-gh-bot
added
the
Community want to contribute
|
/bot run |
|
PR_Github #29094 [ run ] triggered by Bot. Commit: |
|
Hi @XiaoXuan42, thx for contribution! I've approved for your workflow run. pip install pre-commit
pre-commit install
git commit -s -m "your fix"plz run pre-commit and |
|
PR_Github #29094 [ run ] completed with state
|
Signed-off-by: SpicyNoodle <522169030@qq.com>
XiaoXuan42
force-pushed
the
eagle_qwen2_capture_hidden_states
branch
from
1f7ac66 to
17c6fb4
Compare
|
/bot run |
1 similar comment
|
/bot run |
|
PR_Github #29350 [ run ] triggered by Bot. Commit: |
|
PR_Github #29350 [ run ] completed with state
|
|
/bot run --disable-fail-fast |
|
PR_Github #29372 [ run ] triggered by Bot. Commit: |
|
PR_Github #29372 [ run ] completed with state
|
|
@XiaoXuan42, only one test failure but not related to your contribution. Since Qwen2 is not the priority model in our test lists, so currently no tests in CI or QA stage cover the model. I'll request related DEV to review for further action, thx for nice contribution again! |
juney-nvidia
requested review from
mikeiovine,
sunnyqgg and
ziyixiong-nv
and removed request for
dongjiyingdjy and
yechank-nvidia
|
/bot run --disable-fail-fast |
|
PR_Github #29929 [ run ] triggered by Bot. Commit: |
|
PR_Github #29929 [ run ] completed with state |
|
/bot reuse-pipeline |
|
PR_Github #29958 [ reuse-pipeline ] triggered by Bot. Commit: |
|
PR_Github #29958 [ reuse-pipeline ] completed with state |
Signed-off-by: SpicyNoodle <522169030@qq.com> Signed-off-by: Daniil Kulko <kulkodaniil@gmail.com>
7 participants
@coderabbitai summary
Description
Hidden states of qwen2 aren't captured when using eagle speculative decoding, which leads to low accept rate for our self-trained draft model. This PR fix this issue based on the code in modeling_qwen3.py and the accept rate becomes normal in our case.
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)
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
/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.
Details
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.