rxn.onmt_utils.train_command.OnmtTrainCommand

class rxn.onmt_utils.train_command.OnmtTrainCommand(command_type, no_gpu, data_weights, **kwargs)

Bases: object

Class to build the onmt_command for training models, continuing the training, or finetuning.

Parameters

• command_type (RxnCommand) –

• no_gpu (bool) –

• data_weights (Tuple[int, ...]) –

• kwargs (Any) –

__init__(command_type, no_gpu, data_weights, **kwargs)

Parameters

• command_type (RxnCommand) –

• no_gpu (bool) –

• data_weights (Tuple[int, ...]) –

• kwargs (Any) –

Methods

__init__(command_type, no_gpu, data_weights, ...)	param command_type
cmd()	Return the "raw" command for executing onmt_train.
continue_training(batch_size, data, dropout, ...)	param batch_size
execute_from_config_cmd(config_file)	Return the command for executing onmt_train with values read from the config.
finetune(batch_size, data, dropout, ...[, ...])	param batch_size
save_to_config_cmd(config_file)	Return the command for saving the config to a file.
train(batch_size, data, dropout, heads, ...)	param batch_size

cmd()

Return the “raw” command for executing onmt_train.

Return type

List[str]

static execute_from_config_cmd(config_file)

Return the command for executing onmt_train with values read from the config.

Parameters

config_file (Union[str, PathLike]) –

Return type

List[str]

save_to_config_cmd(config_file)

Return the command for saving the config to a file.

Parameters

config_file (Union[str, PathLike]) –

Return type

List[str]