A Sentence is Worth 128 Pseudo Tokens: A Semantic-Aware Contrastive Learning Framework for Sentence Embeddings

Official implementation of A Sentence is Worth 128 Pseudo Tokens: A Semantic-Aware Contrastive Learning Framework for Sentence Embeddings We have just sorted and merged the code, however after a lot of file renaming and moving we are not sure this code still works, we will debug and update this code very soon. The checkpoint files are about 1.7GB which exceeds the limits of our Github LFS quota, we will consider upload to huggingface hub

We will update the README soon