JM-CLIP：A Joint Modal Similarity Contrastive Learning Model for Video-Text Retrieval After the paper is published, we will formally share the code of our work.