Repository navigation
Conversation
📝 WalkthroughWalkthroughCompletionRequest.to_sampling_params now supplies a renamed boolean parameter Changes
Sequence Diagram(s)(omitted — changes are parameter rename and per-token value extraction; no new control-flow warranting a sequence diagram) Estimated code review effort🎯 2 (Simple) | ⏱️ ~8 minutes Suggested reviewers
📜 Recent review detailsConfiguration used: .coderabbit.yaml 📒 Files selected for processing (2)
🚧 Files skipped from review as they are similar to previous changes (2)
⏰ 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)
✨ 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/Issue comments)Type Other keywords and placeholders
Status, Documentation and Community
|
d4c4d9a to
f835571
Compare
|
/bot run |
jhaotingc
changed the title
|
PR_Github #15034 [ run ] triggered by Bot |
|
/bot kill |
![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
🔭 Outside diff range comments (1)
tensorrt_llm/serve/postprocess_handlers.py (1)
60-73: Fix create_logprobs crash, normalize return_logprobs & add NVIDIA headerThe current
create_logprobswill fail when passed a flatList[float]or a mapping without the expected.logprobstructure, andChatPostprocArgs.from_requestcan receive an integer/None forreturn_logprobs. Also, the file needs the standard NVIDIA header.Please apply the following changes:
• tensorrt_llm/serve/postprocess_handlers.py – update
create_logprobssignature and body to handle:
- legacy floats
- dicts mapping token IDs to floats or objects with
.logprob- indexable sequences
• tensorrt_llm/serve/postprocess_handlers.py – inChatPostprocArgs.from_request, castrequest.logprobsto bool
• tensorrt_llm/serve/postprocess_handlers.py – insert the standard NVIDIA copyright header at the topProposed diffs:
--- a/tensorrt_llm/serve/postprocess_handlers.py +++ b/tensorrt_llm/serve/postprocess_handlers.py @@ –1,0 +1,8 @@ +# ============================================================================= +# Copyright (c) 2025, NVIDIA CORPORATION. +# Licensed under the Apache License, Version 2.0 (the “License”); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# http://www.apache.org/licenses/LICENSE-2.0 +# ============================================================================= @@ -60,7 +68,7 @@ def create_logprobs(token_ids: List[int], - logprobs: List[float]) -> ChatCompletionLogProbs: + logprobs: List[object]) -> ChatCompletionLogProbs: assert len(token_ids) == len(logprobs), \ "token_ids and logprobs have different lengths" content: List[ChatCompletionLogProbsContent] = [] - for token_id, logprob in zip(token_ids, logprobs): + for token_id, step_logprob in zip(token_ids, logprobs): token = tokenizer.decode(token_id) # returning multiple logprobs is not supported + # support float, dict[int,Obj], and indexable sequences try: if isinstance(step_logprob, (int, float)): lp = float(step_logprob) @@ -78,13 +86,23 @@ def create_logprobs(token_ids: List[int], else: # sequence or mapping cand = step_logprob[token_id] - lp = float(getattr(cand, "logprob", cand)) + lp = float(getattr(cand, "logprob", cand)) except Exception: lp = None - if lp is None or not isinstance(lp, (float, int)): - lp = -9999.0 - lp = max(lp, -9999.0) + if not isinstance(lp, (int, float)): + lp = -9999.0 + lp = max(lp, -9999.0) first_logprob = ChatCompletionLogProbsContent( token=token, logprob=lp, bytes=list(token.encode("utf-8", errors="replace"))) content.append(first_logprob)--- a/tensorrt_llm/serve/postprocess_handlers.py +++ b/tensorrt_llm/serve/postprocess_handlers.py @@ class ChatPostprocArgs(PostprocArgs): - return cls( + return cls( echo=request.echo, role="assistant" if request.add_generation_prompt else request.messages[-1]["role"], model=request.model, num_choices=request.n, tools=request.tools, tool_choice=request.tool_choice, stream_options=request.stream_options, - return_logprobs=request.logprobs, + return_logprobs=bool(request.logprobs), last_message_content=None, reasoning_parser=request.reasoning_parser, reasoning_parser_dict={}, )
📜 Review details
Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (2)
tensorrt_llm/serve/openai_protocol.py(1 hunks)tensorrt_llm/serve/postprocess_handlers.py(1 hunks)
🧰 Additional context used
📓 Path-based instructions (2)
**/*.py
📄 CodeRabbit Inference Engine (CODING_GUIDELINES.md)
**/*.py: Python code must target Python 3.8+
Python indentation: 4 spaces, no tabs
Maintain module namespace in imports (from package.subpackage import foo; then use foo.SomeClass())
Python file names use snake_case
Python class names use PascalCase
Python functions/methods and local variables use snake_case; variables starting with a number get k_ prefix (e.g., k_99th_percentile)
Global variables use G_ prefixed UPPER_SNAKE_CASE (e.g., G_MY_GLOBAL)
Constants use UPPER_SNAKE_CASE in Python
Avoid shadowing variables from outer scopes in Python
Initialize all externally visible members of a Python class in init
Prefer docstrings for interfaces used outside a file; comments for local code
Use Google-style docstrings for classes and functions (Sphinx-parsable)
Document attributes/variables inline with short docstrings
Avoid reflection when simple alternatives exist (e.g., prefer explicit parameters over dict(**locals()))
In try/except, catch the narrowest exceptions possible
For duck-typing with try/except, keep try body minimal and put logic in else
Files:
tensorrt_llm/serve/postprocess_handlers.pytensorrt_llm/serve/openai_protocol.py
**/*.{cpp,cxx,cc,cu,h,hpp,hxx,hh,cuh,py}
📄 CodeRabbit Inference Engine (CODING_GUIDELINES.md)
Prepend NVIDIA copyright header (current year) to all source files
Files:
tensorrt_llm/serve/postprocess_handlers.pytensorrt_llm/serve/openai_protocol.py
⏰ 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
| # TODO: migrate to use logprobs and prompt_logprobs | ||
| _return_log_probs=bool(self.logprobs), | ||
| logprobs=bool(self.logprobs) | ||
| ) |
There was a problem hiding this comment.
💡 Verification agent
🧩 Analysis chain
Duplication of logprobs flags; also AI summary inconsistency
You’re now setting both _return_log_probs and logprobs based on bool(self.logprobs). If both are consumed downstream, this is fine short-term, but it’s easy for them to drift semantically. Please confirm the backend reads the new logprobs flag and plan deprecating _return_log_probs.
Note: The AI summary says CompletionRequest gained logprobs=..., but the actual change is in ChatCompletionRequest. Flagging for clarity.
Additionally, the file is missing the standard NVIDIA copyright header required by the coding guidelines.
Run this to verify SamplingParams has a logprobs field and to spot downstream usage:
Optional type alignment (outside this hunk): the Chat API’s logprobs input is effectively boolean in your usage. Consider allowing both bool/int for compatibility.
# openai_protocol.py (suggested outside this hunk)
# Change ChatCompletionRequest field typing to accept both:
logprobs: Optional[Union[bool, int]] = None🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "Check SamplingParams definition includes 'logprobs'..."
rg -n "class SamplingParams" -A 20 tensorrt_llm | sed -n '1,120p' | rg -n "logprobs\s*:"
echo
echo "Callsites constructing SamplingParams with 'logprobs' param..."
rg -n "SamplingParams\\(.+logprobs\\s*=" -A 2
echo
echo "Downstream usage of SamplingParams.logprobs..."
rg -n "SamplingParams(?![\\s\\S]*class).*" -n
rg -n "\\.logprobs\\b" tensorrt_llm | rg -v "openai_protocol\\.py"Length of output: 233
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "Searching for class SamplingParams across the repo..."
rg -n "class SamplingParams" -A 20 .
echo
echo "Searching for any references to SamplingParams..."
rg -n "SamplingParams" -A 2 .
echo
echo "Searching for usage of .logprobs outside openai_protocol.py..."
rg -n "\.logprobs\b" . | rg -v "openai_protocol\.py"Length of output: 271470
Align ChatCompletionRequest logprobs flag and update summary
- Verify that the new
logprobsfield onChatCompletionRequest(not_return_log_probs) is actually forwarded into_sampling_paramsand respected by the backend. Once confirmed, plan to deprecate and remove_return_log_probs. - Correct the AI summary so it refers to
ChatCompletionRequest.logprobs(the change is in the chat API, notCompletionRequest). - Add the standard NVIDIA copyright header to the top of
tensorrt_llm/serve/openai_protocol.py.
🤖 Prompt for AI Agents
In tensorrt_llm/serve/openai_protocol.py around lines 590 to 593, update to
forward the new ChatCompletionRequest.logprobs flag (not the legacy
_return_log_probs) into to_sampling_params so the backend receives and respects
ChatCompletionRequest.logprobs; keep backward-compatible handling if needed but
mark _return_log_probs deprecated and plan its removal. Also correct the AI
summary text to reference ChatCompletionRequest.logprobs (chat API) instead of
CompletionRequest, and add the standard NVIDIA copyright header at the top of
this file.
|
PR_Github #15035 [ kill ] triggered by Bot |
|
PR_Github #15034 [ run ] completed with state |
|
PR_Github #15035 [ kill ] completed with state |
Signed-off-by: Jhao-Ting Chen <jhaotingc@nvidia.com>
f835571 to
628774d
Compare
|
/bot run --disable-fail-fast |
|
PR_Github #15036 [ run ] triggered by Bot |
|
PR_Github #15036 [ run ] completed with state |
| first_logprob = ChatCompletionLogProbsContent( | ||
| token=token, | ||
| logprob=max(logprob, -9999.0), | ||
| logprob=max(logprob[token_id].logprob, -9999.0), |
There was a problem hiding this comment.
This will break TRT flow IIRC.
There was a problem hiding this comment.
I didn't test, but that's the data format I got after I sent "logprobs": true to /v1/chat/completions, for both torch and trt sampler.
What's the way to sent the curl and have you checked if 6097 fix the issue (correctly get logprob from pytorch flow).
If you've verified the function works on 6097, we don't need this fix. Thanks!!
|
@jhaotingc can you pls give the launch server command to reproduce the bug? I can check if it is a known issue. |
|
@LinPoly sure. Basically any model has this. Torch/TRT sample has this error too. |
|
@jhaotingc Hi Jhao-ting, Do you still want to land this PR to ToT or you want to close it directly? Thanks |
|
@jhaotingc this pull request has had no activity for more than 120 days and currently has merge conflicts. Is it still active? Please resolve the conflicts or post an update. Otherwise, it will be closed after another 60 days of inactivity. |
logprobs is not correctly passed in from request.
Turned out to be request argument not passed in sampling param.
Thanks @Pernekhan for discovering this.
Summary by CodeRabbit
New Features
Bug Fixes
Refactor
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.
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.