[OMNIML-4671] synth_support

[ChenhanYu]

Draft PR opened by pensieve-intern for OMNIML-4671.

Stage synth_support of Epic OMNIML-4666. The agent ran from the SPEC on the ticket description; review every change before marking ready.

Always-draft is enforced — the bot never auto-merges.

Summary by CodeRabbit

• Chores
  • Added an example pipeline for Qwen3-8B that provides a synthetic data generation workflow using TensorRT-LLM in a containerized SLURM execution environment for the EAGLE3 speculative decoding flow. The pipeline is enabled by default, includes model location and resource configuration (TP/EP, tasks, GPUs), and uses an NVIDIA TensorRT-LLM container.

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: d2b8a732-a528-4667-8b42-11e7906ec702

📥 Commits

Reviewing files that changed from the base of the PR and between b96cb4c and 277bad1.

📒 Files selected for processing (1)

• tools/launcher/examples/Qwen/Qwen3-8B/step1_synth.yaml

🚧 Files skipped from review as they are similar to previous changes (1)

• tools/launcher/examples/Qwen/Qwen3-8B/step1_synth.yaml

---

📝 Walkthrough

Walkthrough

Adds a new YAML pipeline config Qwen3-8B_EAGLE3_synth defining pipeline flags and pipeline.global_vars.hf_model: /hf-local/Qwen/Qwen3-8B, plus task_0 which runs common/tensorrt_llm/query.sh with TRT-LLM server and query arguments, HF_LOCAL env, and SLURM/container settings.

Changes

EAGLE3 Speculative Decoding Pipeline Configuration

Layer / File(s)

Summary

Qwen3-8B synthetic data generation pipeline tools/launcher/examples/Qwen/Qwen3-8B/step1_synth.yaml

New YAML configuration adding job_name: Qwen3-8B_EAGLE3_synth, top-level pipeline flags (skip: false, allow_to_fail: false), pipeline.global_vars.hf_model: /hf-local/Qwen/Qwen3-8B, and task_0 that runs common/tensorrt_llm/query.sh with TRT-LLM server arguments, prompt input/output paths, HF_LOCAL environment variable, and SLURM execution configuration using nvcr.io/nvidia/tensorrt-llm/release:1.2.0 container.

---

🎯 2 (Simple) | ⏱️ ~10 minutes

🚥 Pre-merge checks | ✅ 5 | ❌ 1

❌ Failed checks (1 inconclusive)

Check name	Status	Explanation	Resolution
Title check	❓ Inconclusive	The title '[OMNIML-4671] synth_support' is vague and generic; it lacks specificity about what synth_support entails or what was actually changed.	Provide a more descriptive title that explains the specific change, e.g., 'Add synthetic data generation pipeline configuration for Qwen3-8B EAGLE3' or similar.

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Security Anti-Patterns	✅ Passed	YAML config only. Scanned 976 Python files: no unsafe torch.load, numpy.load, hardcoded trust_remote_code, eval/exec, or nosec comments. Dependencies use permissive licenses.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch pensieve-intern/OMNIML-4666/synth-support

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@tools/launcher/examples/Qwen/Qwen3-8B/step1_synth.yaml`:
- Line 35: The file ends without a trailing newline (yamlfmt/pre-commit failed)
— open the file containing the container:
nvcr.io/nvidia/tensorrt-llm/release:1.2.0 line (step1_synth.yaml) and add a
single newline character at the end of the file so the file terminates with a
newline; save and re-run the pre-commit/CI checks.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 1d1ea7b1-e933-4ba5-b798-52cf90786db2

📥 Commits

Reviewing files that changed from the base of the PR and between 62401e1 and a5f178d.

📒 Files selected for processing (1)

• tools/launcher/examples/Qwen/Qwen3-8B/step1_synth.yaml

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.78%. Comparing base (62401e1) to head (b96cb4c).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1476      +/-   ##
==========================================
- Coverage   76.78%   76.78%   -0.01%     
==========================================
  Files         473      473              
  Lines       51413    51413              
==========================================
- Hits        39476    39475       -1     
- Misses      11937    11938       +1     

Flag	Coverage Δ
unit	52.55% <ø> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.