There are a lot of parameters on which ML algorithms are implemented. The default parameters for these algorithms are updated from time to time for better implementation of these ML algorithms. So, warnings are shown to inform the end-user that the default parameters will be updated in future releases if they plan to update.
A common warning is a Deprecation warning. This warning implies that the name of the package will be changed in future updates or a class will be shifted to another package.
Thumb rule about warnings is that, read the warning and just understand what it is about.
Warning are not generally related to error in your code, and they are not limited to just the ML algorithms. Warnings can be shown in other libraries like numpy and pandas.