aka x_loss function

  • An evaluation function that measures how much the model is wrong.
    • It needs to be a continuous value.
    • If it is a discrete value, it becomes problematic because there is no change when it is slightly modified.