hi I got 1 or 0 as the test labels.
How do i need to convert them to probability??
this is different than log loss.
What language do you use?
In python in sklearn package you can see 2 rules for almost all estimators:
a) predict returns float (common for all kind of regressors)
b) predict returns binary, predict_proba returns float (common for all kind of classifiers)
In some cases you have to turn the probability option on, but for this you have to dive deeper into documentation (this applies e.g. for svm classiffier called SVC)