logger#
Source code: tianshou/highlevel/logger.py
- class LoggerFactory[source]#
Bases:
ToStringMixin
,ABC
- abstract create_logger(log_dir: str, experiment_name: str, run_id: str | None, config_dict: dict | None = None) BaseLogger [source]#
Creates the logger.
- Parameters:
log_dir – path to the directory in which log data is to be stored
experiment_name – the name of the job, which may contain os.path.delimiter
run_id – a unique name, which, depending on the logging framework, may be used to identify the logger
config_dict – a dictionary with data that is to be logged
- Returns:
the logger
- abstract get_logger_class() type[BaseLogger] [source]#
Returns the class of the logger that is to be created.
- class LoggerFactoryDefault(logger_type: Literal['tensorboard', 'wandb', 'pandas'] = 'tensorboard', wand_entity: str | None = None, wandb_project: str | None = None, group: str | None = None, job_type: str | None = None, save_interval: int = 1)[source]#
Bases:
LoggerFactory
- create_logger(log_dir: str, experiment_name: str, run_id: str | None, config_dict: dict | None = None) BaseLogger [source]#
Creates the logger.
- Parameters:
log_dir – path to the directory in which log data is to be stored
experiment_name – the name of the job, which may contain os.path.delimiter
run_id – a unique name, which, depending on the logging framework, may be used to identify the logger
config_dict – a dictionary with data that is to be logged
- Returns:
the logger
- get_logger_class() type[BaseLogger] [source]#
Returns the class of the logger that is to be created.