dhresearch/quality-router-v2-pairwise

Pairwise utility ranker

This file is a seed-0 refit made on 2026-10-03. The original training run wrote predictions and did not save weights. This upload is the refit.

quality_router_v2 version 8 sets learned_downrouting to false. The served card loads no fitted weights. This file is a learned model from the same study, and the served card leaves it unloaded.

Role

Pairwise utility ranker fit on the main export. The phase command resolves to LightGBM when that package imports. The served flow does not call this object. Economy mode on the served card names a learned downrouter and is off.

Data

Main export. Dataset: https://huggingface.co/datasets/dhresearch/outcome-router-v2-main.

Fit

--router pairwise on the main export. Candidate-route flags are absent, matching the shell script.

Seed 0. Software at refit time: python 3.14.2, sklearn 1.9.1, numpy 2.5.3, joblib 1.6.0, lightgbm 4.7.0, scipy 1.18.1.

Agreement

Frozen file data/real_v2/preds_v2_pairwise.jsonl: predicted_route matches on 250 of 250 shared test rows (rate 1.000000). Full-row matches, including rounded probabilities and labels: 250 of 250. Mismatched keys: {}. seed-0 refit. The original run wrote this prediction file and did not save weights. This upload is not the original weight file.

Load

Unpickle with joblib while outcome_router_data and lightgbm are importable.

Class: PairwiseUtilityRanker.