Results for citeseer and acm

[gayanku]

Im having some trouble getting the results for citeseer and acm.

python sdcn.py --name cite gives me

run 1

199Q :acc 0.6008 , nmi 0.3269 , ari 0.3326 , f1 0.5654 199Z :acc 0.6534 , nmi 0.3687 , ari 0.3591 , f1 0.5702 199P :acc 0.6008 , nmi 0.3269 , ari 0.3326 , f1 0.5654

run 2

199Q :acc 0.5939 , nmi 0.3225 , ari 0.3255 , f1 0.5600 199Z :acc 0.6351 , nmi 0.3697 , ari 0.3804 , f1 0.5960 199P :acc 0.5939 , nmi 0.3225 , ari 0.3255 , f1 0.5600

run 3

199Q :acc 0.6011 , nmi 0.3264 , ari 0.3340 , f1 0.5639 199Z :acc 0.6552 , nmi 0.3796 , ari 0.3988 , f1 0.6061 199P :acc 0.6011 , nmi 0.3264 , ari 0.3340 , f1 0.5639

python sdcn.py --name acm

run 1

199Q :acc 0.8628 , nmi 0.5744 , ari 0.6364 , f1 0.8618 199Z :acc 0.8797 , nmi 0.6253 , ari 0.6792 , f1 0.8783 199P :acc 0.8628 , nmi 0.5744 , ari 0.6364 , f1 0.8618

run 2

199Q :acc 0.8688 , nmi 0.5910 , ari 0.6515 , f1 0.8680 199Z :acc 0.8846 , nmi 0.6376 , ari 0.6915 , f1 0.8839 199P :acc 0.8688 , nmi 0.5910 , ari 0.6515 , f1 0.8680

Each time, scores seem to be difference from those in the paper. Is there probably some missing config or setting in the paper or the code?

[gayanku]

Pls ignore. Results are as published. (need to take max, not last epoch).