yamle.utils.tuning_utils module#

yamle.utils.tuning_utils.sample_initial_random_configs(config_space, n_samples=10)[source]#

Sample random initial configurations from the config space.

Return type:

List[Dict[str, Any]]

yamle.utils.tuning_utils.best_config_to_command_arguments(best_config, omit=['label', 'no_evaluation', 'no_saving'])[source]#

Convert the best config to command arguments.

Return type:

str

yamle.utils.tuning_utils.plot_different_runs_and_metrics(results_df, save_path)[source]#

Plot each different run separately on the same graph with respect to the epoch and all the logged metrics.

Return type:

None

yamle.utils.tuning_utils.plot_different_runs_and_metric_config_combinations(results_df, save_path)[source]#

Plot a scatter plot for all the different runs and their end metric values with respect toall hyperparameter combinations.

Return type:

None

yamle.utils.tuning_utils.plot_different_metrics_and_trial_id(results_df, save_path)[source]#

Plot a scatter plot where trial id is on the x-axis and the last value of the metric is on the y-axis.

Return type:

None