How to load trials results for hyperparameter_search

Question

I run 100 trials for hyperparameters tuning of a transformers as following:

trainer = Trainer(...)

best_trial = trainer.hyperparameter_search(direction="maximize",
                                                backend='ray',
                                                hp_space=ray_hp_space,
                                                search_alg=hyperopt_search,
                                                n_trials=100,
                                                resources_per_trial={"cpu":4, "gpu":1},
                                                checkpoint_score_attr="training_iteration",
                                                progress_reporter=reporter,
                                                local_dir=training_args.output_dir + "/hyper_results"
                                              )

I did not print "best_trial". I am wondering if possible to load the trials results and anlyze them to get the best result ?