Save rankings feature#49
Merged
Merged
Conversation
Add a `save_rankings: bool = False` flag to `workrb.evaluate()` that persists per-target ranking score arrays for each ranking-task dataset under `<output_folder>/rankings/<model_name>/<task>__<dataset_id>.json`. Each artifact also records `model_name` in its payload so files remain self-describing if moved. To enable this without recomputing the prediction matrix, `RankingTask.evaluate` is split into `compute_prediction_matrix` + `compute_metrics_from_prediction_matrix`; default behavior is unchanged.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Description
Add a
save_rankings: bool = Falseflag toworkrb.evaluate()that persists the full prediction matrix for each ranking-task dataset under<output_folder>/rankings/<model_name>/<task>__<dataset_id>.json. Each artifact carries a self-describing header (schema version, workrb version, model/task/dataset identity, sizes, query/target canary strings).To enable this without recomputing,
RankingTask.evaluateis split intocompute_prediction_matrix+compute_metrics_from_prediction_matrix. Default behaviour ofevaluate()is unchanged.Adds a companion entry point
workrb.evaluate_rankings(rankings_dir, tasks, ...)that replays saved artifacts to compute metrics without a model. A newrankingsmodule handles loading, header validation (schema version is hard reject; structural mismatch is a hard reject; workrb-version drift only warns), and matrix materialization.RankingsArtifactMissingandRankingsArtifactInvalidare exported.Also:
BenchmarkMetadatagainsreplayed_from_workrb_version(None for normal runs), and_get_dataset_ids_to_evaluateis refactored soExecutionMode.ALLconsistently keeps every dataset regardless ofaggregation mode.
Checklist