It is easy to get nearest neighbors for a word using model.get_nearest_neighbors(word) but
I could not find a method that would give me the nearest neighbors for a given vector (which may
be calculated and not identical to any word vector in the model).
How to do that?