yttemporal dataset

[00asuka]

Can I download the following yttemporal dataset uploaded on the official site and put it in data/YTTemporal?
gs://merlot/yttemporal180m/yttemporal180m_{i:03d}of 100.jsonl.gz

In YTTemporal_dataset.py

    def _load_metadata(self):
        metadata_dir = './meta_data'
        split_files = {
            'train': 'yttemporal_train.csv',  # _1000000
            'val': 'yttemporal_val.csv',  # there is no test
            'test': 'yttemporal_val.csv'
        }

How should I define split_files?