docs(BENCH-QWEN38-TENSORFOLD-GAP): link measurement tools - #3361
Open
localai-org-maint-bot wants to merge 2 commits into
Open
localai-org-maint-bot wants to merge 2 commits into
localai-org-maint-bot wants to merge 2 commits into
Conversation
The public result lacks links to the shipped measurement workflow. Scope a CPU-verifiable documentation follow-up for ISSUE-LOCAL-01M3R485ZSPG7VHK1Y8B2207DZ without changing benchmark claims. FOLLOWING_AGENTS_PROTOCOL Following-Agents-Protocol: true AI-Assisted: true Assisted-by: Codex:GPT-6 [Codex]
Readers could not reach the shipped endpoint harness from the public result. Link its instructions and templates, and explain the evidence limits. The benchmark stays blocked and publishes no new measurement. The three focused CPU suites pass all 58 tests. README structure, retained evidence, relative links, and whitespace checks pass. Full preflight has existing unrelated failures; the operator verifies its baseline comparison before handoff. No executable behavior changes. Issue: ISSUE-LOCAL-01M3R485ZSPG7VHK1Y8B2207DZ FOLLOWING_AGENTS_PROTOCOL Following-Agents-Protocol: true AI-Assisted: true Assisted-by: Codex:gpt-6 [exec_command]
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The public Qwen3.8 TensorFold result did not link to its shipped measurement workflow. Add a README news entry and links to the endpoint harness guide, environment templates, and capture runner.
Explain the evidence limits: current comparisons provide neither token parity nor a cross-engine ratio,
--draft offdoes not reconfigure a running vllm.cpp server, and a successful process exit does not prove a valid measurement. The comparison remains blocked by missing artifacts. No code or benchmark values change.Tracks local issue
ISSUE-LOCAL-01M3R485ZSPG7VHK1Y8B2207DZ. The spec precedes the documentation commit. This is a documentation follow-up, not closure of the benchmark campaign.Validation on CPU:
python3 -m pytest -q tests/scripts/test_qwen38_endpoint_bench.py tests/scripts/test_run_qwen38_tensorfold_gap.py tests/scripts/test_validate_qwen38_tensorfold_evidence.py: 58 passed.python3 scripts/check-readme-structure.pyandpython3 scripts/check-agent-record.py: passed.python3 tools/bench/validate_qwen38_tensorfold_evidence.py .agents/evidence/bench-qwen38-tensorfold-gap/latest: passed, retainingBLOCKED_MISSING_ARTIFACTS.Independent review of
700d744982537f5c5205a9a6581c10eed1c1668bpassed with no findings. The reviewer reran all 58 tests and proved the link check detects a broken new target, then restored the file byte-for-byte.The repository-wide preflight is not green. The completed checker sweeps report the same nine failures; the broader mutation sweep remains incomplete. The baseline and changed tree report unrelated release-state, benchmark-index, environment-documentation, oracle-pin, and stale-record failures. The local tool environment also lacks dependencies needed by some broad checks. These are outside this documentation-only change; no checker is weakened. No GPU, model execution, or new performance measurement was used.
FOLLOWING_AGENTS_PROTOCOL
Following-Agents-Protocol: true
AI-Assisted: true
Assisted-by: Codex:GPT-6 [Codex]