score_dir#

ibl_bwb_eval.scoring.ts2.score_dir(pred_dir, gt_dir)[source]#

Score all TS2 prediction files found under pred_dir.

Expects prediction files written by PredictionsWriter (metadata carries label, task, recording_id, seed). pred_dir may point at any subtree (e.g. the shared predictions/ root), files whose task isn’t a ts2-* task are skipped.

Ground truth files are looked up as: {gt_dir}/{task}/{recording_id}/ground_truth.safetensors

Returns a dict keyed by (label, task, recording_id, seed), where each value is a dict mapping metric name to scalar value.

Return type:

dict[tuple[str, str, str, int], dict[str, float]]