--- dataset_info: features: - name: input_ids sequence: int32 - name: attention_mask sequence: int8 - name: labels sequence: int64 splits: - name: train num_bytes: 79618892856 num_examples: 2989146 download_size: 3566147442 dataset_size: 79618892856 configs: - config_name: default data_files: - split: train path: data/train-* ---