ChongYou / subspace-clustering

Toolbox for large scale subspace clustering

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Run it within 1 hour Reproducable?

wassryan opened this issue · comments

When evaluating your Cluster method on my machine(64cpu, 180g memory), following the command of python run_mnist.py, which has 70000 data samples to cluster, the time cost doesn't match your report of 1 hour? I need almost 8 hours to do ElasticNetSubspaceClustering??? Can you help me figure out what's the problem that causes the method running so slowly?

Hi wassryan, thanks for posting the question. Did you encounter the same issue with SSC-OMP? Did you use the options algorithm='spams', active_support=True for ElasticNetSubspaceClustering? If no and yes, then likely the issue is that the spams package does not work very well on your platform.