split 1 and 2 in E5Data.py

[SiyuanHuangSJTU]

Hello, thanks for your gerat work! I have noticed that in E5Data.py, there exists:
if dataset in [ "allnli_split2", "quora_duplicates_split1", "quora_duplicates_split2", ]:

and

# combine split1 and split2 new_data_map = {} for dataset in data_map: new_dataset = dataset.replace("_split1", "").replace("_split2", "") if new_dataset not in new_data_map: new_data_map[new_dataset] = [] new_data_map[new_dataset] += data_map[dataset] data_map = new_data_map

I noticed that they were not used. What are actually split 1 and 2 used for? Thanks.

[huangfu170]

mark