refactor(chinese-cosmopedia): 更新数据集加载方式并调整相关参数

This commit is contained in:
2025-07-15 12:54:31 +08:00
parent bea4f83842
commit e6f5b26a75
2 changed files with 7 additions and 5 deletions
+5 -3
View File
@@ -63,6 +63,7 @@ def process_dataset(
dataset_config: str = None,
max_chunk_num: int = None,
split: str = None,
data_files: List[str] = None,
split_name: str = "train",
column_name: str = "text",
chunk_size: int = 1000000,
@@ -70,9 +71,10 @@ def process_dataset(
):
dataset_dict = load_dataset(
data_dir=dataset_name,
data_files=dataset_config,
split=split
path=dataset_name,
name=dataset_config,
split=split,
data_files=data_files
)
train_dataset = dataset_dict[split_name]