Use run_recbole

We enclose the training and evaluation processes in the api of run_recbole(), which is composed of: dataset loading, dataset splitting, model initialization, model training and model evaluation.

If this process can satisfy your requirement, you can recall this api to use RecBole.

You can create a python file (e.g., ), and write the following code into the file.

from recbole.quick_start import run_recbole

run_recbole(dataset=dataset, model=model, config_file_list=config_file_list, config_dict=config_dict)

dataset is the name of the data, such as ‘ml-100k’, model indicates the model name, such as ‘BPR’.

config_file_list indicates the configuration files, config_dict is the parameter dict. The two variables are used to config parameters in our toolkit. If you do not want to use the two variables to config parameters, please ignore them. In addition, we also support to control parameters by the command line.

Please refer to Config Introduction for more details about config settings.

Then execute the following command to run::

python --[param_name]=[param_value]

–[param_name]=[param_value] is the way to control parameters by the command line.