Skip to content

rutgers-db/QuantizationBenchmark

Repository files navigation

QuantizationBenchmark

python run.py --build-images --algorithm ProductQuantizationFaiss --force-rebuild

python run.py --dataset sift-128-euclidean --algorithm ProductQuantizationFaiss --data-dir ./data/

Config yaml: Organized by dataset name

Two formats are supported:

  1. Simple format (for algorithms without search-time parameters): All parameters are used for both building and searching. List values are automatically expanded to all combinations.

  2. Build/Search separation format (for algorithms with search-time parameters): Separate build-time and search-time parameters to avoid redundant training. Example: dataset-name: build: param1: [value1, value2] # Build parameters (affect index construction) search: param2: [value3, value4] # Search parameters (only affect query) common: param3: value5 # Used for both build and search

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages