in the past the train operator in Data Intelligence allowed to use other hardware configurations called "resource plan" together with the customizing of individual python pacakges (the PIP packages).
I am missing this feature in the new train operator. Only with tensorflow I can use a e.g GPU configuration.
How can I utilize a larger CPU / Infrastructure with the new train operator and use my own PIP packages ?
I would think that is very important to speed up the performance of DI when the "normal" environment is not fast enough.
Any help appreciated