title: Bi-Direction Label-Guided Semantic Enhancement for Cross-Modal Hashing
The CLIP dataset used in this paper is sourced from the paper titled "When CLIP Meets Cross-modal Hashing Retrieval: A New Strong Baseline." You can visit the paper's page to download the corresponding dataset.
Taking NUS-WIDE as an example, our model can be trained and verified by the following command: python main_nus.py