Which loss function to use for multi-objective regression?

Question

Machine learning

Good day, I have a task to write a regression model with multiple outputs (4) using the keras framework
Tried to compile a model with mse loss function, but gives an error:

ValueError: A target array with shape (50, 1) was passed for an output of shape (None, 4) while using as loss `mean_squared_error`. This loss expects targets to have the same shape as the output.

I searched on the Internet, but could not find suitable information.
Actually, the question is: which loss function to use?