The last step of every retrieval pipeline: you have similarity scores and you want the best few.
Given a 1-D array of scores and an integer k, return the indices of the k highest scores, ordered best first.
Sorting the whole array is O(n log n) and you only need the top k. At 80 million candidates that difference is the difference between a usable system and a timeout.
You have read the opening. The rest is free with a Bezen account, and no card is needed.
Behind the sign-in