Repository navigation
Prepare v0.1.0 preview release (CRF, every prediction flagged) - #43
Merged
Merged
Conversation
The release image is built with the baselines extra so it can serve CRF releases. v0.1.0 serves crf-batches-01-02 with the uncalibrated review policy (every prediction flagged). The rollback drill r1 (rules) -> v0.1.0 (CRF) -> r1 passed on local Docker. Release notes state that no acceptance gate was evaluated.
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
release.ymlbuilds withEXTRAS="serve baselines", so tagged images can serve the CRF. The defaultserveimage is rules-only and could not load a CRF bundle. The Dockerfile comment lists the three variants.crf-batches-01-02(bundle weights SHA-256d1f8cb5a…07056) with the uncalibratedreview-policy-0, so every prediction is flagged Review required.docs/release_notes/v0.1.0.mdstates that this is a preview and no acceptance gate was evaluated (only 4 hand-labelled test records exist; the test split is unread). It lists what was measured (agreement and the hand-labelled dev slice) and why the CRF and not DeBERTa.Verified locally (Docker, 2026-11-06)
docker build --build-arg EXTRAS="serve baselines"→ 1.06 GB image;deploy/smoke.shpassed.release createforr1(rules, existing image) andv0.1.0(CRF).deploy/rollback_drill.sh r1 v0.1.0passed in 15 s: ready check against the manifest, extraction, idempotent review save, retrieval and export at every step; reviews kept across rollback.app verifypassed;/readyzreportscrf-batches-01-02, the weights digest, preprocessing1andreview-policy-0.After merge: tag
v0.1.0;release.ymlbuilds, smoke-tests and pushes the image to GHCR and creates the GitHub release.Checks
ruff check,ruff format --check,pytest -m "not slow"pass;release.ymlparses.