Building NSG index for single core CPU

ZJULearning / nsg

Navigating Spreading-out Graph For Approximate Nearest Neighbor Search

MIT License

632 stars 150 forks source link

Building NSG index for single core CPU #34

Open K16DIABLO opened 2 years ago

K16DIABLO commented 2 years ago

Hello, I'm interested in your work and I'm trying to reproduce your result.

I want to reproduce the result for DEEP100M dataset on single core. However, _test_nndescent is not working because of segmentation fault. Can you let me know how did you build kNN graphs for this dataset? Also, can you share both kNN graph and nsg index?

Thanks

K16DIABLO commented 2 years ago

I successfully built kNN graph for DEEP100M dataset, but I faced another problem. Current _test_nsgindex raises segmentation fault, so I fixed it by changing the variable type. However, it takes too much time to build the graph even I made it work. I used AWS r6i and c6i to build DEEP100M graph, but it takes more than 2 days. It seems _test_nsgindex is different from that you used to build DEEP100M for your paper.

Thanks