[https://nvbugs/5625743][fix] Fix output corruption when enable n>1 and streaming #9802
Conversation
|
/bot run |
📝 WalkthroughWalkthroughThe pull request includes a bug fix in the detokenization pipeline to prevent re-decoding of already-processed tokens when handling multiple outputs, and adds a new streaming completions test case that validates batch operations with semantic similarity checks. 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: 1
🧹 Nitpick comments (1)
tests/unittest/llmapi/apps/_test_openai_completions.py (1)
8-8: Consider maintaining namespace per coding guidelines.Per the coding guidelines: "Always maintain the namespace when importing in Python, even if only one class or function from a module is used." Consider changing to
from utils import utiland usingutil.similar()instead.However, this matches the existing import style in the file (lines 12-13), so this is a low-priority suggestion.
As per coding guidelines, prefer namespace imports.
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (2)
tensorrt_llm/executor/result.py(1 hunks)tests/unittest/llmapi/apps/_test_openai_completions.py(2 hunks)
🧰 Additional context used
📓 Path-based instructions (2)
**/*.py
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
**/*.py: The 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 in Python, even if only one class or function from a module is used (e.g., usefrom package.subpackage import fooand thenfoo.SomeClass()instead offrom package.subpackage.foo import SomeClass)
Python filenames should use snake_case (e.g.,some_file.py)
Python class names should use PascalCase (e.g.,class SomeClass)
Python function and method names should use snake_case (e.g.,def my_awesome_function():)
Python local variable names should use snake_case, with prefixkfor variable names that start with a number (e.g.,k_99th_percentile = ...)
Python global variables should use upper snake_case with prefixG(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 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 type and description (e.g.,self.x = 5followed by"""<type>: Description of 'x'""")
Avoid using reflection in Python when functionality can be easily achieved without reflection
When using try-except blocks in Python, limit the except clause to the smallest set of specific errors possible instead of catching all exceptions
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 and use the else block to implement the logic
Files:
tensorrt_llm/executor/result.pytests/unittest/llmapi/apps/_test_openai_completions.py
**/*.{cpp,h,cu,py}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
All TensorRT-LLM Open Source Software code files should contain an NVIDIA copyright header that includes the current year at the top
Files:
tensorrt_llm/executor/result.pytests/unittest/llmapi/apps/_test_openai_completions.py
🧠 Learnings (4)
📚 Learning: 2025-08-18T08:42:02.640Z
Learnt from: samuellees
Repo: NVIDIA/TensorRT-LLM PR: 6974
File: tensorrt_llm/serve/scripts/benchmark_dataset.py:558-566
Timestamp: 2025-08-18T08:42:02.640Z
Learning: In TensorRT-LLM's RandomDataset (tensorrt_llm/serve/scripts/benchmark_dataset.py), when using --random-token-ids option, sequence length accuracy is prioritized over semantic correctness for benchmarking purposes. The encode/decode operations should use skip_special_tokens=True and add_special_tokens=False to ensure exact target token lengths.
Applied to files:
tensorrt_llm/executor/result.py
📚 Learning: 2025-08-14T21:04:50.248Z
Learnt from: thorjohnsen
Repo: NVIDIA/TensorRT-LLM PR: 6910
File: cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp:0-0
Timestamp: 2025-08-14T21:04:50.248Z
Learning: In KV cache onboarding logic during prefill in cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp, when calculating which blocks fall within the attention window, use getTokensPerBlock() to advance token indices rather than block->getUniqueTokens().size(), because the calculation needs to consider the post-prefill state where blocks will be filled to capacity, not their current token count.
Applied to files:
tensorrt_llm/executor/result.py
📚 Learning: 2025-08-15T06:46:54.897Z
Learnt from: eopXD
Repo: NVIDIA/TensorRT-LLM PR: 6767
File: cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp:0-0
Timestamp: 2025-08-15T06:46:54.897Z
Learning: In cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp addToken function, newly allocated blocks are unshared by design. The beam search path in addToken (when sequence.getNumTokens() > windowSize) is currently broken/non-functional with SWA, so the block allocation doesn't follow a shared-then-unshared pattern.
Applied to files:
tensorrt_llm/executor/result.py
📚 Learning: 2025-08-09T20:57:04.084Z
Learnt from: sklevtsov-nvidia
Repo: NVIDIA/TensorRT-LLM PR: 3294
File: cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_tma_warp_specialized_input.cu:118-127
Timestamp: 2025-08-09T20:57:04.084Z
Learning: In the CUTLASS MoE finalize fusion implementation (cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_tma_warp_specialized_input.cu), when setting `fused_finalize_epilogue.stride_final_output` with shape `(hidden_size, num_output_tokens, 1)`, the `num_rows_in_final_output` should be set to `num_output_tokens` (not `hidden_size`) because of a swap+transpose operation that maps rows of the output tensor to `hidden_size` and columns to `num_output_tokens`.
Applied to files:
tensorrt_llm/executor/result.py
🪛 Ruff (0.14.7)
tests/unittest/llmapi/apps/_test_openai_completions.py
230-230: Comment contains ambiguous × (MULTIPLICATION SIGN). Did you mean x (LATIN SMALL LETTER X)?
(RUF003)
⏰ 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 (2)
tensorrt_llm/executor/result.py (1)
634-636: LGTM! Critical fix for streaming with n > 1.The update to
_last_token_ids_lenafter detokenization correctly prevents re-decoding of already-processed tokens in subsequent streaming responses when multiple outputs (n > 1) are generated. This ensures each beam output's token tracking remains consistent across streaming chunks.tests/unittest/llmapi/apps/_test_openai_completions.py (1)
208-239: Good test coverage for the streaming + n>1 fix.The test properly validates the fix by testing batch completions with n=3 in streaming mode, which directly exercises the code path fixed in
result.py. The similarity check with threshold=0.8 is reasonable given potential non-deterministic variations.However, verify that the
similar()function fromutils.utilexists and confirm its behavior is appropriate for this validation.
|
PR_Github #27352 [ run ] triggered by Bot. Commit: |
|
PR_Github #27352 [ run ] completed with state |
SimengLiu-nv
force-pushed
the
fix-option-n-streaming
branch
from
997d8d7 to
3f94aff
Compare
|
/bot run |
|
PR_Github #27432 [ run ] triggered by Bot. Commit: |
|
PR_Github #27432 [ run ] completed with state |
tensorrt_llm/executor/result.py
Outdated
|
|
||
| # Update _last_token_ids_len after detokenization to prevent | ||
| # re-decoding the same tokens in subsequent responses when n > 1. | ||
| beam_output._last_token_ids_len = len(beam_output.token_ids) |
SimengLiu-nv
force-pushed
the
fix-option-n-streaming
branch
from
3f94aff to
fa294e6
Compare
|
/bot run |
|
PR_Github #27748 [ run ] triggered by Bot. Commit: |
|
PR_Github #27748 [ run ] completed with state |
SimengLiu-nv
force-pushed
the
fix-option-n-streaming
branch
from
fa294e6 to
bf8fc92
Compare
|
/bot run |
|
PR_Github #27902 [ run ] triggered by Bot. Commit: |
|
PR_Github #27902 [ run ] completed with state |
SimengLiu-nv
force-pushed
the
fix-option-n-streaming
branch
from
bf8fc92 to
112988a
Compare
|
/bot run --disable-fail-fast |
|
PR_Github #30614 [ run ] triggered by Bot. Commit: |
|
PR_Github #30614 [ run ] completed with state
|
…nd streaming. Added tests accordingly. Signed-off-by: SimengLiu-nv <[email protected]>
Signed-off-by: SimengLiu-nv <[email protected]>
Signed-off-by: SimengLiu-nv <[email protected]>
SimengLiu-nv
force-pushed
the
fix-option-n-streaming
branch
from
112988a to
59261df
Compare
|
/bot run --disable-fail-fast |
|
PR_Github #30786 [ run ] triggered by Bot. Commit: |
|
PR_Github #30786 [ run ] completed with state
|
| beam_output.token_ids) != prev_token_lens.get( | ||
| id(beam_output), 0) | ||
| if not output_received_new_tokens: | ||
| continue |
There was a problem hiding this comment.
This could lead to wrong outputs when using beam search.
If one beam exits early, it may in the next iteration swap its tokens with another non-exited beam without adding a new token. As the number of tokens did not change, the swapped tokens will not be decoded, which could result in a wrong output.
You may adjust this line to
if not output_received_new_tokens and not self.sampling_params.use_beam_search:
continue
4 participants
…nd streaming. Added tests accordingly.
Summary by CodeRabbit
Bug Fixes
Tests
✏️ Tip: You can customize this high-level summary in your review settings.
Description
Test Coverage
tests/unittest/llmapi/apps/_test_openai_completions.py::test_batch_completions_with_option_n_streaming
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.