I've got data, in which the target variable is bimodally distributed. The target variable is also bimodally distributed conditional on the predictors.

Is there a possibility, to run a GLM Model on this data?

Especially, is it possible to fit a GLM to a mixed distribution or an empiric distribution?