Classification metrics can't handle a mix of binary and continuous-multioutput targets

Classification metrics are designed to evaluate the performance of classification models, which predict categorical labels or classes for given inputs. They are not appropriate for evaluating models that predict continuous or numeric values, which are typically evaluated using regression metrics.

In addition, classification metrics assume that the targets are binary or multi-class categorical variables, and are not designed to handle a mix of binary and continuous-multioutput targets. This means that if your target variable includes both binary and continuous values, you will need to use regression metrics or some other appropriate evaluation method.

Examples of regression metrics include mean squared error (MSE), mean absolute error (MAE), root mean squared error (RMSE), and R-squared. These metrics can be used to evaluate the performance of models that predict continuous or numeric values, and they are able to handle a mix of binary and continuous-multioutput targets.