NumPy has new-style DTypes with additional features and improved consistency. As of NumPy 1.24, these still require use of unstable/experimental API and are not quite production ready. However, projects planning in the mid- or long-term are recommended to use the new API, and we are actively working on finalizing them.
Because of this, no more improvements are expectd here and this repository is now archived.
Examples for new style DTypes can be found in https://github.com/numpy/numpy-user-dtypes
仓库 for the development of user dtypes for numpy.