spookey14

I see in the code base that `FedMLRunner` accepts a `ClientTrainer` as an optional argument. However, even after defining custom trainer I can see that the code does not pass...