SISAP 2023 LAION2B Challenge: Example using faiss

This repository gives a small working example of solving tasks in the SISAP 23 LAION2B Indexing Challenge.

Requirements

This code was tested using Python 3.8 using Anaconda. See https://github.com/sisap-challenges/sisap23-laion-challenge-faiss-example/blob/master/.github/workflows/ci.yml for an example installation of FAISS in this setup.

Running

Running

$ python search/search.py

takes care of

• downloading datasets/query sets used to benchmark the index to data/
• run 30-NN queries on the index for each query in the query set using a couple of different hyperparameters,
• stores results in result/ as hdf5 files that can be post-processed using https://github.com/sisap-challenges/sisap23-laion-challenge-evaluation. This repository is added as a submodule.

Evaluation

$ python eval/eval.py