Is there support for shap_scores.py that for models that don't contain a bias track? When setting up a model with no bias, there is only 1 sequence input, but the shap_scores script assumes that there are three model inputs. This means that usage of this CLI breaks the contribution generation after a model is trained.
Is there support for shap_scores.py that for models that don't contain a bias track? When setting up a model with no bias, there is only 1 sequence input, but the shap_scores script assumes that there are three model inputs. This means that usage of this CLI breaks the contribution generation after a model is trained.