Well of course it's up to your ears, but from my experience:
Take not a too short release otherwise you easily get holes in the verse, since usually the flow continues right after the compressor kicks in. Attack just as long as you need to keep the punch of the consonants. Ratio? well no tip for that, depends really on the dynamics of the take.
Furthermore my way is to already compress just a little bit while recording to get a good overall level into the computer.
The real compression and eqing I do later on, the latter depends on the frequencies of the voice and the instruments and the place where the voice is supposed to sit in the mix.