History and Credits

MinPy is intended to be the NumPy frontend of the MXNet project. Its core members contributed to MXNet’s execution engine. Having completed that part, the team decided to take a step back and rehthink the user experience, before moving to yet another stage of performance optimizations.

The key goal is innovation without compromising performance and usability. This philosophy is shared by the community, but the difference is we see sticking to NumPy interface and imperative programming as higher priority.

Technical inspiration

We have also observed a great deal of innovations from the community. Therefore, we innovate when necessary, and otherwise draw inspiration (and sometimes direct implementations) from the followings:


A great number of people have contributed to the project. Most of them are students from around the world, and work with their spare time. People listed below without specified contributions are generalists who have wider work span. We want also to thank the MXNet development community for timely technical support.

  • Minjie Wang; NYU, project lead (GitHub)
  • Yutian Li: Face++/Stanford (GitHub)
  • Larry Tang: NYU Shanghai/Michigan (GitHub)
  • Haoran Wang: NYU Shanghai/CMU (GitHub)
  • Tianjun Xiao: Microsoft/Tesla (GitHub)
  • Ziheng Jiang: Fudan/NYU Shanghai (GitHub)
  • Alex Gai: NYU Shanghai [Model builder] (GitHub)
  • Sean Welleck: NYU/NYU Shanghai [Reinforcement learning] (GitHub)
  • Xu Zhou: ShanghaiTech [CS231 courseware] (GitHub)
  • Kerui Min: BosonData [RNN example on MNIST] (Linkedin)
  • Murphy Li: NYU Shangai [Tensorboard integration] (GitHub)
  • Professor Zheng Zhang: NYU Shanghai [general advising] (GitHub)