Where exactly in kmeans.py is 'k' estimated as suggested in 3.2?

[yashjakhotiya]

Hi,

Great work on defining the GCD setting and motivating subsequent papers for GCD!

In the paper in sec 3.2, it is mentioned that the total number of clusters in the unlabelled part of the dataset is found with a black box optimization with Brent's algorithm on the labelled part.

I believe this parameter - finding the total number of clusters - is an integral part of the overall method suggested in the paper and the realistic GCD setting. Also, I believe kmeans.py was used to calculate the results in table 2 and table 3 (from the line 'ours') as suggested in the ordering of scripts in README.md. However, I could not find the k estimation using Brent's algorithm used in kmeans.py.

Could you please let me know if I missed anything in kmeans.py or if you used some other script to calculate tables 2 and 3?

Thank you again for the work!