Conversation
…odel slow parity, bench, invariance, timing, optional A/B worktree, summary; RUNBOOK.md logging, idempotent steps, planner+summary unit tests
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
python -m benchmarks.m5 --out m5- [--ab-branch ...] chains the full M5 gate sequence in order, each step logged with wall time + exit status into /RUNBOOK.md: doctor --json (gate: FAIL aborts), pytest -m slow per parity model (MODEL_ID env; Qwen3-8B/Llama-3.1-8B/Gemma-3-12B defaults, --parity-models/--models-file override), bench --model quality, invariance on quality (TypeSafe cases fetched when missing), timing --reps 5, bench --models-file for the rest, and with --ab-branch a temp worktree + own venv rerunning bench+invariance there (worktree removed after). Step 8 writes SUMMARY.md comparing main vs A/B (agreement/accuracy, flip rate, log-odds drift, time per case, peak memory) from the produced json files. Idempotent via done-markers (--fresh overrides). Planner + summary builder unit-tested with fakes (21 tests, no model loads); BENCHMARKING.md gains a 10-line section. Rebased on PR #41.