osmarks e57931d47f Multithread query server ...

While profiling suggests that most operations are cheap and IO-bound rather than CPU-bound, the GEMM for deduplication is pretty slow. As such, use multiple threads for higher throughput.

2025-01-31 13:47:47 +00:00

..

common.rs

release version

2025-01-24 09:24:28 +00:00

dump_processor.rs

release version

2025-01-24 09:24:28 +00:00

generate_index_shard.rs

RobustVamana algorithm for big index run

2025-01-16 21:10:12 +00:00

get_embedding.py

release version

2025-01-24 09:24:28 +00:00

main.rs

release version

2025-01-24 09:24:28 +00:00

ocr.rs

improve observability and fix up Reddit dump for full-scale run

2024-11-02 19:38:05 +00:00

query_disk_index.rs

Multithread query server

2025-01-31 13:47:47 +00:00

reddit_dump.rs

release WIP DiskANN index orchestration code

2025-01-01 14:40:24 +00:00

score_model.rs

release version

2025-01-24 09:24:28 +00:00

video_reader.rs

File metadata storage.

2024-06-26 20:02:12 +01:00