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
+2 -2
View File
@@ -6,9 +6,9 @@ if __name__ == "__main__":
item_size = max_chunk_num * chunk_size
process_dataset(
dataset_name="chinese-cosmopedia",
dataset_name="opencsg/chinese-cosmopedia",
output_subdir="chinese-wiki",
split=f"train[:{item_size}]",
data_files=[f"0000{i}.parquet" for i in range(5)],
max_chunk_num=max_chunk_num,
chunk_size=chunk_size,
)