train4

--- dataset_info: features: - name: code1 dtype: string - name: code2 dtype: string - name: similar dtype: int64 - name: __index_level_0__ dtype: int64 splits: - name: tr