Search code, repositories, users, issues, pull requests...

This repository contains the official implementation and data for "SparseCL: Sparse Contrastive Learning for Contradiction Retrieval".

Before running the experiments, ensure that you have the correct environment set up by referring to the requirements.txt file. This file contains all the necessary packages.

If you want to use different models, please change the args: 'model_name_or_path' and 'model_name', e.g., if you want to use UAE as the backbone model, you can replace 'BAAI/bge-base-en-v1.5' with 'WhereIsAI/UAE-Large-V1' and 'our_bge' with 'our_uae'. Remember to adjust the hyperparameters accordingly if you change to different models. Please refer to Table 7 in our paper for our hyperparameter choices.

If you want to use different models, please change the args: 'model_name_or_path' and 'model_name', e.g., if you want to use UAE as the backbone model, you can replace 'BAAI/bge-base-en-v1.5' with 'WhereIsAI/UAE-Large-V1' and 'our_bge' with 'our_uae'. Remember to adjust the hyperparameters accordingly if you change to different models. Please refer to Table 7 in our paper for our hyperparameter choices.

Leave a Comment