I have two types of comments:
In the header explaining what a method does, parameters, return value... Later doxygen will generate from these comments a nice documentation.
The other type of comments explains code passages: What I am doing here, sometimes why I'm doing something that way and not the obvious way... These comments help me to understand my own code when I have to read it again some days, months or even years later.
But try to avoid writing how you are doing something - most of the time reading the code is better.