Using group_by() in dplyr for certain number of rows

In dplyr, is there a way to use group_by() not with a variable but group by a specific number of rows? I am using SQLite with dplyr and I don't think SQLite has windowing function.