One card per model, as published next to each model in the
cfb_model_artifacts release: model type, training 'xgboost' version,
feature list, label, training date and the rule-era encoding contract.
Format
A named list with one element per model (ep_model,
wp_spread, wp_naive, fg_model, fd_model, two_pt_model,
xpass_model, cfb_cp_model, cfb_cp_model_air_yards, qbr_model,
wp_ot_reach), each a list parsed from the card's JSON.
Source
The cfb_model_artifacts release of
https://github.com/sportsdataverse/sportsdataverse-data.