Locality Sensitive Hashing in Fourier Frequency Domain For Soft Set Containment Search
This directory contains code necessary for running all the experiments.
Recent versions of Pytorch, numpy, scipy, sklearn and matplotlib are required.
Additional third party softwares used - Dr.Hash, SBERT
You can install all the required packages using the following command:
$ pip install -r requirements.txt
#Datasets and trained models Please download files from https://rebrand.ly/fhash and place in the current folder. This contains the original datasets, dataset splits, trained models and other intermediate data dumps for reproducing tables and plots.
The command lines and scripts used for training models are listed commands.txt.
Command lines specify the exact hyperparameter settings used to train the models.
FinalSubmission-Figs-NeurIPS23-Main.ipynb and FinalSubmission-Figs-NeurIPS23-Supp.ipynb contains code used for generating all the figures presented in the paper (main and appendix respectively).
Notes:
- GPU usage is required for model training
- Hashing is done only on CPU.
- source code files are all in src folder.