[None][chore] Guided decoding fixes and benchmarking #6752
Conversation
Signed-off-by: Enwei Zhu <[email protected]>
Signed-off-by: Enwei Zhu <[email protected]>
📝 WalkthroughWalkthroughThe changes introduce and propagate a new optional Changes
Sequence Diagram(s)sequenceDiagram
participant User
participant DatasetLoader
participant InferenceRequest
participant LlmManager
participant Evaluator
User->>DatasetLoader: Provide JSON stream (may include guided_decoding_params)
DatasetLoader->>InferenceRequest: Create with guided_decoding_params (if present)
User->>LlmManager: Submit InferenceRequest
LlmManager->>LlmManager: Set sampling_params.guided_decoding from request.guided_decoding_params
User->>Evaluator: Run evaluate
Evaluator->>Evaluator: Serialize requests (incl. guided_decoding_params) to bench_requests.jsonl
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~15–20 minutes Possibly related PRs
Suggested labels
Suggested reviewers
Note 🔌 MCP (Model Context Protocol) integration is now available in Early Access!Pro users can now connect to remote MCP servers under the Integrations page to get reviews and chat conversations that understand additional development context. ✨ Finishing Touches
🧪 Generate unit tests
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. 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
SupportNeed help? Create a ticket on our support page for assistance with any issues or questions. CodeRabbit Commands (Invoked using PR comments)
Other keywords and placeholders
Documentation and Community
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 5
🔭 Outside diff range comments (1)
tensorrt_llm/bench/utils/data.py (1)
1-1: Add NVIDIA copyright header totensorrt_llm/bench/utils/data.pyPer project guidelines (see CODING_GUIDELINES.md), all non-test source files (
*.py,*.cpp, etc.) must begin with the standard NVIDIA copyright header including the current year.• File to update:
- tensorrt_llm/bench/utils/data.py (currently starts with
import json)Suggested diff (insert at the very top of the file):
+# ----------------------------------------------------------------------------- +# Copyright (c) 2025 NVIDIA Corporation. All rights reserved. +# SPDX-License-Identifier: Apache-2.0 +# ----------------------------------------------------------------------------- + import json
🧹 Nitpick comments (3)
tensorrt_llm/bench/benchmark/throughput.py (1)
385-390: Hard-forcingskip_tokenizer_init=Falsemay override user intent
runtime_config.get_llm_args()can already carry a user-chosenskip_tokenizer_init. Unconditionally resetting it ignores CLI/YAML knobs and makes future debugging harder.Consider only defaulting when the key is absent:
- kwargs['skip_tokenizer_init'] = False + kwargs.setdefault('skip_tokenizer_init', False)tensorrt_llm/evaluate/interface.py (1)
110-126: Bench-request dump is convenient but unboundedRepeating every output 100 × can create very large files for sizeable datasets and is impossible to disable.
Suggest making
num_repeatsa parameter (default = 1) or guarding with an env var.tensorrt_llm/bench/utils/data.py (1)
47-59: Document the new input JSON key in the function docstringThe function now accepts “guided_decoding_params” from the input JSON lines. Please update the docstring to document this optional key and note it is unpacked into a GuidedDecodingParams instance.
I can draft the docstring update with a short example referencing the accepted fields from GuidedDecodingParams if you’d like.
📜 Review details
Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (6)
tensorrt_llm/_torch/pyexecutor/guided_decoder.py(2 hunks)tensorrt_llm/bench/benchmark/throughput.py(1 hunks)tensorrt_llm/bench/benchmark/utils/asynchronous.py(1 hunks)tensorrt_llm/bench/dataclasses/general.py(2 hunks)tensorrt_llm/bench/utils/data.py(5 hunks)tensorrt_llm/evaluate/interface.py(2 hunks)
🧰 Additional context used
📓 Path-based instructions (2)
**/*.py
📄 CodeRabbit Inference Engine (CODING_GUIDELINES.md)
**/*.py: Python code should conform to Python 3.8+.
Indent Python code with 4 spaces. Do not use tabs.
Always maintain the namespace when importing in Python, even if only one class or function from a module is used.
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.
Initialize all externally visible members of a Python class in the constructor.
For interfaces that may be used outside a Python file, prefer docstrings over comments.
Comments in Python 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.
Attributes and variables in Python can be documented inline; attribute docstrings will be rendered under the class docstring.
Avoid using reflection in Python when functionality can be easily achieved without it.
When using try-except blocks in Python, limit the except to the smallest set of errors possible.
When using try-except blocks to handle multiple possible variable types in Python, keep the body of the try as small as possible, using the else block to implement the logic.
Files:
tensorrt_llm/bench/benchmark/utils/asynchronous.pytensorrt_llm/bench/dataclasses/general.pytensorrt_llm/bench/benchmark/throughput.pytensorrt_llm/_torch/pyexecutor/guided_decoder.pytensorrt_llm/evaluate/interface.pytensorrt_llm/bench/utils/data.py
**/*.{cpp,h,hpp,cc,cxx,cu,py}
📄 CodeRabbit Inference Engine (CODING_GUIDELINES.md)
All TensorRT-LLM Open Source Software code should contain an NVIDIA copyright header that includes the current year. This includes .cpp, .h, .cu, .py, and any other source files which are compiled or interpreted.
Files:
tensorrt_llm/bench/benchmark/utils/asynchronous.pytensorrt_llm/bench/dataclasses/general.pytensorrt_llm/bench/benchmark/throughput.pytensorrt_llm/_torch/pyexecutor/guided_decoder.pytensorrt_llm/evaluate/interface.pytensorrt_llm/bench/utils/data.py
🧠 Learnings (3)
📚 Learning: 2025-08-08T04:10:18.987Z
Learnt from: djns99
PR: NVIDIA/TensorRT-LLM#6728
File: cpp/tensorrt_llm/plugins/mixtureOfExperts/mixtureOfExpertsPlugin.cpp:966-966
Timestamp: 2025-08-08T04:10:18.987Z
Learning: TensorRT plugins currently don't support padding functionality, and TensorRT is not getting new features (in maintenance mode). This means that duplicating parameters like mExpertHiddenSize in function calls, even with TODO comments, can be acceptable as pragmatic solutions within these constraints.
Applied to files:
tensorrt_llm/bench/benchmark/throughput.py
📚 Learning: 2025-08-06T13:58:07.506Z
Learnt from: galagam
PR: NVIDIA/TensorRT-LLM#6487
File: tests/unittest/_torch/auto_deploy/unit/singlegpu/test_ad_trtllm_bench.py:1-12
Timestamp: 2025-08-06T13:58:07.506Z
Learning: In TensorRT-LLM, test files (files under tests/ directories) do not require NVIDIA copyright headers, unlike production source code files. Test files typically start directly with imports, docstrings, or code.
Applied to files:
tensorrt_llm/bench/benchmark/throughput.py
📚 Learning: 2025-07-22T09:22:14.726Z
Learnt from: yechank-nvidia
PR: NVIDIA/TensorRT-LLM#6254
File: tensorrt_llm/_torch/pyexecutor/model_engine.py:1201-1204
Timestamp: 2025-07-22T09:22:14.726Z
Learning: In TensorRT-LLM's multimodal processing pipeline, shared tensor recovery using `from_shared_tensor()` is only needed during the context phase. Generation requests reuse the already-recovered tensor data and only need to call `strip_for_generation()` to remove unnecessary multimodal data while preserving the recovered tensors. This avoids redundant tensor recovery operations during generation.
Applied to files:
tensorrt_llm/_torch/pyexecutor/guided_decoder.py
🔇 Additional comments (4)
tensorrt_llm/bench/dataclasses/general.py (1)
11-28: Addition looks correctOptional
guided_decoding_paramsis properly typed and imported; no further feedback.tensorrt_llm/_torch/pyexecutor/guided_decoder.py (2)
82-91: Logic refinement is soundThe draft path now advances the matcher only when (a) finishing context init chunk or (b) in generation. This prevents unnecessary matcher work on intermediate context chunks – good improvement.
196-198: Relaxed assertion could mask real mis-alignmentAllowing
offset < logits.size(0)keeps CUDA-graph dummy rows safe, but it also hides genuine bookkeeping errors.
Recommend asserting the exact equality whend2t is None(normal path) and only allow<=when dummies are expected:if d2t is None: assert offset == logits.size(0), "Offset/logits mismatch." else: assert offset <= logits.size(0)tensorrt_llm/bench/utils/data.py (1)
88-88: LGTM: Accumulator init for guided decoding paramsInitialization is consistent with other “all_*” accumulators and keeps list alignment intact.
Summary by CodeRabbit
New Features
Bug Fixes
Chores
Description
Test Coverage
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.