Sentence BERT base JA mean tokens

A Japanese sentence embedding model based on BERT using mean token pooling for semantic similarity.