MAYA: Addressing Inconsistencies in Generative Password Guessing through a Unified Benchmark