LI TIANCHENG

> The option `print_oe` of `dflow.utils.run_command` does exactly the same thing by virtue of selectors. You only need to add an argument to `dflow_run_command`. Refer to https://github.com/deepmodeling/dflow/blob/48c24cc4f494acb5a12c8d99293f1156c31342ad/src/dflow/utils.py#L657. Besides, I think...

Will this project apply to apple core ml, which will utilize cpu, gpu and ane units in apple silicon.

I tried to solve these warnings but the tests won't pass for windows : https://github.com/OutisLi/deepmd-kit/commit/3976ee936c4a24ce267f25c74da1e610db6d1e0a