refactor: 重构数据管道模块

This commit is contained in:
2026-03-20 21:36:37 +08:00
parent 7045de9c32
commit 3fb2326115
20 changed files with 739 additions and 345 deletions
+5 -3
View File
@@ -1,10 +1,12 @@
from datasets import load_dataset
from modules.utils import process_dataset
from modules.datapipeline import DataPipeline
if __name__ == "__main__":
dataset = load_dataset("Blaze7451/enwiki_structured_content")
process_dataset(
pipeline = DataPipeline()
pipeline.process_dataset(
dataset_dict=dataset,
output_subdir="english-wiki-pretrain",
max_chunk_size=5,
max_chunk_num=5,
)