Hi @<1581454875005292544:profile|SuccessfulOtter28> ! You could take a look at how the HPO was built using optuna: None .
Basically: you should create a new class which inherits from SearchStrategy . This class should convert clearml hyper_parameters to some parameters the Ray Tune understands, then create a Tuner and run the Ray Tune hyper paramter optimization.
The function Tuner will optimize params for should be a function which creates a new clearml task, this task being a clone of the task you want to optimize. Then the values of the objectives you want to optimize for are fetched in this function and evaluated (from the cloned task). In the optuna.py , the function that does all of this is objective .

@<1523701435869433856:profile|SmugDolphin23> Quick question, does the SearchStrategy use a Bayesian Optimisation?