[https://nvbugs/5590408][fix] Fallback to greedy sampling in two-model overlap scheduler #9321
Conversation
…l overlap scheduler Signed-off-by: ziyixiong-nv <219238287+ziyixiong-nv@users.noreply.github.com>
|
/bot run |
📝 WalkthroughWalkthroughModified conditional logic in the Changes
Estimated code review effort🎯 2 (Simple) | ⏱️ ~10 minutes
Pre-merge checks and finishing touches❌ Failed checks (1 inconclusive)
✅ Passed checks (2 passed)
✨ Finishing touches
🧪 Generate unit tests (beta)
Tip 📝 Customizable high-level summaries are now available in beta!You can now customize how CodeRabbit generates the high-level summary in your pull requests — including its content, structure, tone, and formatting.
Example instruction:
Note: This feature is currently in beta for Pro-tier users, and pricing will be announced later. 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 (2)
tensorrt_llm/_torch/speculative/model_drafter.py (2)
623-627: Consider conditionally buildingpy_draft_logitsfor efficiency.The
py_draft_logitslist is built unconditionally in the inner loop, but it's only used whendisable_overlap_scheduleris True (line 631). When the overlap scheduler is enabled, this computation is wasted.Apply this diff to optimize:
target_model_req.py_draft_tokens = [] - py_draft_logits = [] for token_idx in range(self.max_draft_len): target_model_req.py_draft_tokens.append( draft_tokens_host[token_idx][req_idx]) - py_draft_logits.append(draft_logits[token_idx][req_idx]) # The overlap scheduler doesn't support rejection sampling yet, so we don't update the py_draft_logits to get it fallback to greedy sampling. if self.disable_overlap_scheduler: + py_draft_logits = [] + for token_idx in range(self.max_draft_len): + py_draft_logits.append(draft_logits[token_idx][req_idx]) target_model_req.py_draft_logits = torch.stack(py_draft_logits) - else: - logger.warning( - "Overlap scheduler doesn't support rejection sampling yet. Fallback to greedy sampling." - )
629-629: Add a TODO comment to track the incomplete implementation.The PR description states this is a workaround because "the GPU-version validation function is not implemented." Adding a TODO comment will help track this technical debt and guide future implementation.
Apply this diff:
- # The overlap scheduler doesn't support rejection sampling yet, so we don't update the py_draft_logits to get it fallback to greedy sampling. + # TODO(https://nvbugs/5590408): The overlap scheduler doesn't support rejection sampling yet + # because the GPU-version validation function is not implemented. We don't update the + # py_draft_logits to fallback to greedy sampling. if self.disable_overlap_scheduler:
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
tensorrt_llm/_torch/speculative/model_drafter.py(1 hunks)
🧰 Additional context used
🧠 Learnings (2)
📚 Learning: 2025-08-19T12:45:11.997Z
Learnt from: amitz-nv
Repo: NVIDIA/TensorRT-LLM PR: 7033
File: tensorrt_llm/_torch/pyexecutor/model_engine.py:0-0
Timestamp: 2025-08-19T12:45:11.997Z
Learning: In tensorrt_llm/_torch/pyexecutor/model_engine.py, DoRA (Delta Orthogonal Rank Adaptation) functionality was removed from the PyTorch flow to eliminate issues with inverted DoRA detection logic. The original is_dora condition was checking if scaling_vec_pointer == 0, which was potentially incorrect.
Applied to files:
tensorrt_llm/_torch/speculative/model_drafter.py
📚 Learning: 2025-11-07T09:18:04.997Z
Learnt from: Funatiq
Repo: NVIDIA/TensorRT-LLM PR: 8587
File: tensorrt_llm/_torch/pyexecutor/llm_request.py:129-139
Timestamp: 2025-11-07T09:18:04.997Z
Learning: In `LogitsStorage.get()` method in `tensorrt_llm/_torch/pyexecutor/llm_request.py`, when `exclude_last=True`, there is an invariant that at least 2 chunks must have been appended to `_logits_indices`. The parameter is designed to drop the entire last chunk (not just the last token), which is expected behavior for the overlap scheduler that generates one extra token in a separate chunk.
Applied to files:
tensorrt_llm/_torch/speculative/model_drafter.py
🧬 Code graph analysis (1)
tensorrt_llm/_torch/speculative/model_drafter.py (1)
tensorrt_llm/logger.py (1)
warning(132-133)
⏰ 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
|
PR_Github #25120 [ run ] triggered by Bot. Commit: |
|
PR_Github #25120 [ run ] completed with state |
Signed-off-by: ziyixiong-nv <219238287+ziyixiong-nv@users.noreply.github.com>
|
/bot run |
|
PR_Github #25169 [ run ] triggered by Bot. Commit: |
|
PR_Github #25169 [ run ] completed with state |
|
/bot run |
|
PR_Github #25206 [ run ] triggered by Bot. Commit: |
|
PR_Github #25206 [ run ] completed with state |
4 participants
…l overlap scheduler
Summary by CodeRabbit
✏️ Tip: You can customize this high-level summary in your review settings.
Description
The current two-model overlap scheduler doesn't support rejection sampling because the GPU-version validation function is not implemented yet, so we need to fallback to greedy sampling.
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.