Several options for achieving good coverage of hyperparameter space more efficiently than systematic grid search are discussed in scikit-learn documentations. Can pick or experiment and pick to enable us to explore large parameter search space within 'reasonable' time. Should lead to better-performing models.
Sampling-based HPO can be a good alternative. E.g., https://github.com/optuna/optuna or one of sampling-based option provided by scikit-learn.
Several options for achieving good coverage of hyperparameter space more efficiently than systematic grid search are discussed in scikit-learn documentations. Can pick or experiment and pick to enable us to explore large parameter search space within 'reasonable' time. Should lead to better-performing models.
Sampling-based HPO can be a good alternative. E.g., https://github.com/optuna/optuna or one of sampling-based option provided by scikit-learn.