transformer模型的设置

    transformers.pipeline(task: str, 
    					  model: Optional = None, 
    					  config: Optional[Union[str, transformers.configuration_utils.PretrainedConfig]] = None, 
    					  tokenizer: Optional[Union[str, transformers.tokenization_utils.PreTrainedTokenizer]] = None, 
    					  framework: Optional[str] = None, 
    					  revision: Optional[str] = None, 
    					  use_fast: bool = True, **kwargs) → transformers.pipelines.base.Pipeline