[https://nvbugs/6278377][fix] Prepend a no-quant baseline entry (- accuracy: 88) under…

[tensorrt-cicd]

Summary

• Root cause: references/gsm8k.yaml had only quantized entries (FP8_BLOCK_SCALES, NVFP4) for stepfun-ai/Step-3.7-Flash, so the BF16/auto lookup from test_auto_dtype had no match and raised ValueError("Not registered specs: ...").
• Fix: Prepend a no-quant baseline entry (- accuracy: 88) under stepfun-ai/Step-3.7-Flash in gsm8k.yaml, matching the existing 88 baseline used for the FP8/NVFP4 variants.
• Automated fix generated by repair-bot

Test plan

• Verify fix on the same GPU type as the original failure
• Check for regressions in related tests

Links

• Bug: https://nvbugs/6278377

Summary by CodeRabbit

• Tests
  • Updated accuracy reference benchmarks for model validation.

[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: 3fa01f2f-f3ba-4b41-97e9-5c4c03a6f070

📥 Commits

Reviewing files that changed from the base of the PR and between 130ae82 and e832f82.

📒 Files selected for processing (1)

• tests/integration/defs/accuracy/references/gsm8k.yaml

---

📝 Walkthrough

Walkthrough

A single line is added to the GSM8K accuracy reference YAML file, inserting a new list entry with accuracy: 88 under the stepfun-ai/Step-3.7-Flash model key. No quantization or spec-decoding fields accompany this entry. All existing entries for this model remain unchanged.

Changes

GSM8K Accuracy Reference Update

Layer / File(s)	Summary
Add Step-3.7-Flash baseline accuracy entry tests/integration/defs/accuracy/references/gsm8k.yaml	Adds a new list item with accuracy: 88 for stepfun-ai/Step-3.7-Flash, without quant_algo or spec_dec_algo fields, representing a baseline (unquantized) accuracy reference.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~2 minutes

Suggested reviewers

• greg-kwasniewski1
• nvchenghaoz
• suyoggupta
• crazydemo

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly and specifically describes the main change: prepending a no-quant baseline entry with accuracy 88 to gsm8k.yaml for the stepfun-ai/Step-3.7-Flash model, which is the core fix for the test failure.
Description check	✅ Passed	The description includes root cause analysis, the specific fix applied, test verification, and bug link, providing clear context for the change. However, it lacks explicit section headers matching the template structure and doesn't include all checklist items.
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.

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

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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

[kaiyux]

included the changes in #15583, closing this PR.