While there will certainly be a mistake in the version published with the proceedings (i.e., on the USB stick or other media given out at the conference), it is likely that it can still be corrected in the final archival version. If your conference is associated with a professional society that maintains archival versions (e.g., IEEE, ACM, AAAI), the same mechanisms for handling errata on a journal article post-publication can generally be applied to errata on a conference article post-publication. I know whereof I speak, because I had much the same thing happen to me with one of my early conference articles as well; it was in an ACM conference, and though dealing with the errata was a pain, I got it through and the version you download today should be correct.

While there will certainly be a mistake in the version published with the proceedings (i.e., on the USB stick or other media given out at the conference), it is likely that it can still be corrected in the final archival version. If your conference is associated with a professional society that maintains archival versions (e.g., IEEE, ACM, AAAI), check and see if the same mechanisms for handling errata on a journal article post-publication can be applied to errata on a conference article post-publication. I had much the same thing happen to me with one of my early conference articles as well; it was in an ACM conference, and though dealing with the errata was a pain, I got it through and the version you download today should be correct.

Correction: per the OP's comment apparently the IEEE won't do it, which seems problematic and is news to me. In that case, the best thing to do is probably to just

1. Make sure it's correct in the journal version (and include a footnote in that paper that explicitly notes the error in the prior conference publication), and
2. Post a note alongside your self-archived pre-print giving the errata as well.