subhadarship

yes, that would be useful I guess

num_clusters should be >=2. full documentation coming up soon!

feel free to check out the new documentation: https://subhadarship.github.io/kmeans_pytorch

Thanks for the Q! There is no necessary reason to force the output to be on the CPU. I just did it because it fit my use case. Feel free...

Hi @edgarriba, thanks for the appreciation !! Yes, we can integrate