Classification metrics can't handle a mix of continuous and multiclass targets

This error message typically occurs when you are trying to use a classification metric that is only applicable for multiclass targets (i.e. discrete labels) on a dataset that has both continuous and multiclass targets.

For example, if you have a dataset where some of the targets are categorical variables (e.g. "red", "blue", "green"), and some are continuous variables (e.g. the height of a person), you cannot use a multiclass classification metric like accuracy or F1 score to evaluate the performance of a machine learning model on this dataset.

Instead, you should use an appropriate evaluation metric that can handle both continuous and categorical variables, such as mean squared error (MSE) or mean absolute error (MAE) for the continuous variables and accuracy or AUC-ROC for the categorical variables. Alternatively, you could consider transforming the continuous variables into categorical variables, for example by binning them into discrete categories, so that you can use a multiclass classification metric.