dmlc / dgl

Python package built to ease deep learning on graph, on top of existing DL frameworks.
http://dgl.ai
Apache License 2.0
13.42k stars 3k forks source link

Invalid DGL files #6472

Open YijianLiu opened 11 months ago

YijianLiu commented 11 months ago

When after using METIS for reddit, I get node_feat.dgl, but I can't load it. graphlist, = dgl.load_graphs(".node_feat.dgl") Check failed: magicNum == kDGLSerializeMagic (15950236669746234431 vs. 15937764024707752255) : Invalid DGL files Stack trace.

paoxiaode commented 11 months ago

can you give the sample code which can reproduce this issue

YijianLiu commented 11 months ago

can you give the sample code which can reproduce this issue

My main questions are about METIS. I am trying to use metis on reddit datatset, and I get reddit.json Reddist dataset has about 240,000, when I use METIS to get 3 parts of graph.dgl, I find that the nodes on one server have up to 150,000, it is greatly more than 80,000 for average node nums, why?

rudongyu commented 11 months ago

@YijianLiu Is this issue a duplicate of https://discuss.dgl.ai/t/about-node-allocation-of-metis/4058?

If so, I will close this one.

github-actions[bot] commented 10 months ago

This issue has been automatically marked as stale due to lack of activity. It will be closed if no further activity occurs. Thank you