可以通过key来直接查询,得到对应的train、valid和test数据集。
这里的train,valid,test都是Dataset类型,有 features和num_rows两个属性。..., List[str], NoneType] = None,
batched: bool = False,
batch_size: Union[int, NoneType] = 1000...,
remove_columns: Union[str, List[str], NoneType] = None,
keep_in_memory: bool = False,
load_from_cache_file...: bool = True,
cache_file_names: Union[Dict[str, Union[str, NoneType]], NoneType] = None,
writer_batch_size...NoneType] = None, # 使用此参数,可以使用多进程处理
desc: Union[str, NoneType] = None,
) -> 'DatasetDict'
Docstring