Change of affiliation after acceptance of a paper?

Different journals may have different rules, however, in general:

Affiliations indicate where the work was done (to be precise, where a (co-)author has been working while doing the work presented in the paper).

If you change institutes while working on a paper (including writing up and revising) you should enter both affiliations. When you are affiliated to multiple institutes while working on a paper (simultaneously), add both as well.

However, in a case like yours, assuming that the paper is now accepted (no more revisions), only the old affiliation should be added (unless your journal of choice suggests something else). No work has been done while you are working in the new position.