ValueError: provided clabel is inconsistent with node partition
Hi!
I'm using minimize_nested_blockmodel_dl (with deg_corr=True)
to construct hierarchical clustering.
However, the exception ValueError: provided clabel is inconsistent with node partition
often occurs for large graphs (mostly more than half million vertices).
Even in a same graph, sometimes it works, sometimes it fails.
I tried several graphs from The University of Florida Sparse Matrix Collection. The exception occurred in below graphs and others (link to download *.gt.gz files)
stat96v3 (|V|=1,113,780 |E|=3,317,717)
- [
ch8-8-b5 (|V|=564,480 |E|=3,386,848)
] (https://www.dropbox.com/s/0vtg7ceoz8135xc/ch8-8-b5.gt.gz?dl=1) delaunay_n20 (|V|=1,048,576 |E|=3,145,686)
naca0015 (|V|=1,039,183 |E|=3,114,818)
chevron4 (|V|=711,450 |E|=2,836,285)
largebasis (|V|=440,020 |E|=2,560,040)
Is there a size limitation for minimize_nested_blockmodel_dl
?
Thank you! Oh-Hyun