Dense Sparse Retrieval: Using Sparse Language Models for Inference Efficient Dense Retrieval