You always have the Wine project, and one could contribute to such a project if DirectML is the way one wants to run ML on a system - if they choose to ignore t

You always have the Wine project, and one could contribute to such a project if DirectML is the way one wants to run ML on a system - if they choose to ignore the other APIs to use the GPU.