Faster function for top n elements in a dictionaty than heapq.nlargest()

Question

For finding the top n elements of a dictionary we use top_n = heapq.nlargest(). This returns sorted list as an answer. Is there a more optimal way to achieve this