¿What does lightgbm classifier prediction generates in python?

i have a question. When i use lightGBMclassifiers i've realized that the events rate doesnt have a linear relationship with my predictions. So my question is, what kind of predictions generates python using the function predict_proba and whats the difference with the function predict Actually i've read the documentation: https://lightgbm.readthedocs.io/en/latest/pythonapi/lightgbm.LGBMClassifier.html but it isnt clear for me. Thank you for help.