Is there a way to calculate the f1 score of the masks in amg.py?

I am trying to use SAM on an existing dataset for a segmentation competition to see how it performs zero shot for cells. the metrics for the competition are the avg f1 score of the all the competition. Is there an existing method?

Start with amg.py and inspect how its masks are produced. Compare those outputs with the existing cell-segmentation competition annotations and determine whether an F1 calculation is already available. Done means the repository provides a clear way to calculate the competition's average mask F1 score.

Written by the indexing model from the issue text.