refactor(utils): 将 pre_tarin_process 和 sft_process 函数合并为 process_dataset

This commit is contained in:
2025-07-16 12:17:58 +08:00
parent f64a1a15ee
commit faf9f81e7c
6 changed files with 21 additions and 50 deletions
+2 -2
View File
@@ -1,9 +1,9 @@
from datasets import load_dataset
from utils import pre_tarin_process
from utils import process_dataset
if __name__ == "__main__":
dataset = load_dataset("BelleGroup/train_3.5M_CN")
# pre_tarin_process(
# process_dataset(
# dataset_dict=dataset,
# output_subdir="belle_sft",
# max_chunk_size=5,