mindformers.trainer.TrainingArguments

class mindformers.trainer.TrainingArguments(output_dir: str = None, use_parallel: bool = None, profile: bool = None, only_save_strategy: bool = None, sink_mode: bool = None, sink_size: bool = None, batch_size: int = None, per_device_train_batch_size: int = None, per_device_eval_batch_size: int = None, learning_rate: float = None, weight_decay: float = None, adam_beta1: float = None, adam_beta2: float = None, adam_epsilon: float = None, max_grad_norm: float = None, num_train_epochs: float = None, lr_scheduler_type: Union[mindformers.trainer.utils.LRType, str] = None, optim: Union[mindformers.trainer.utils.OptimizerType, str] = None, warmup_steps: int = 0, save_strategy: Union[mindformers.trainer.utils.SaveIntervalStrategy, str] = None, integrated_save: bool = None, save_steps: int = None, save_seconds: int = None, save_total_limit: Optional[int] = None, seed: int = None, data_seed: Optional[int] = None, resume_from_checkpoint: Optional[str] = None, auto_trans_ckpt: bool = None, resume_training: bool = None, do_eval: bool = None, eval_step_interval: int = None, eval_epoch_interval: int = None)

TrainingArguments is the subset of the arguments we use in our default config which is relate to the training in MindSpore.

convert_args_to_mindformers_config(task_config: mindformers.tools.register.config.MindFormerConfig = None)

convert training arguments to mindformer config type for adapting hugging-face.

property get_device_id

get device id for training.

property get_device_num

get device num for training.

property get_rank_id

get rank id for training.