recbole.trainer.hyper_tuning¶

exception recbole.trainer.hyper_tuning.ExhaustiveSearchError[source]¶: Bases: Exception

class recbole.trainer.hyper_tuning.HyperTuning(objective_function, space=None, params_file=None, params_dict=None, fixed_config_file_list=None, display_file=None, algo='exhaustive', max_evals=100, early_stop=10)[source]¶

Bases: object

HyperTuning Class is used to manage the parameter tuning process of recommender system models. Given objective funciton, parameters range and optimization algorithm, using HyperTuning can find the best result among these parameters

Note

HyperTuning is based on the hyperopt (https://github.com/hyperopt/hyperopt)

Thanks to sbrodeur for the exhaustive search code. https://github.com/hyperopt/hyperopt/issues/200

export_result(output_file=None)[source]¶

Write the searched parameters and corresponding results to the file

Parameters:: output_file (str) – the output file

static params2str(params)[source]¶

convert dict to str

Parameters:: params (dict) – parameters dict
Returns:: parameters string
Return type:: str

plot_hyper()[source]¶

run()[source]¶: begin to search the best parameters

trial(params)[source]¶

Given a set of parameters, return results and optimization status

Parameters:: params (dict) – the parameter dictionary

recbole.trainer.hyper_tuning.exhaustive_search(new_ids, domain, trials, seed, nbMaxSucessiveFailures=1000)[source]¶: This is for exhaustive search in HyperTuning.