TusKANNy / seismic

Official software repository of S. Bruch, F. M. Nardini, C. Rulli, and R. Venturini, "Efficient Inverted Indexes for Approximate Retrieval over Learned Sparse Representations". Long Paper @ ACM SIGIR 2024 (Best Paper Runner-up).
MIT License
42 stars 1 forks source link

Questions about loading the index using `mmap`. #2

Open MochiXu opened 2 weeks ago

MochiXu commented 2 weeks ago

Thank you for your outstanding work. When Seismic loads the index file, does it load all the data into memory? Is it possible to use memory mapping to load the index file, which would reduce physical memory usage. I hope to apply this index library to ClickHouse to support sparse vector search operations.

rossanoventurini commented 1 week ago

Very interesting! I’ve sent you an email to better understand your requirements.