This article discusses the development of DANCE, a deep learning library and benchmark designed to accelerate advancements in single cell analysis. It offers a comprehensive set of tools for analyzing single-cell data at scale, allowing developers to create their deep-learning models with greater ease and efficiency. Additionally, it can be used as a benchmark for comparing the performance of various computational models for single-cell analysis. DANCE presently includes support for 3 modules, 8 tasks, 32 models, and 21 datasets.