Update openwebtext_split.py

openwebtext_split.py

@@ -19,7 +19,6 @@ import os
 import re
 from itertools import chain
 import tarfile
-import lzma
 
 import datasets
 
@@ -56,7 +55,8 @@ def custom_iter_archive(path_or_buf, _filter=lambda x: True):
             if not file_path.endswith('xz'):
                 continue
             file_obj = stream.extractfile(tarinfo)
-            with lzma.LZMAFile(file_obj, 'rb') as inner_stream:
+            print(type(file_obj))
+            with tarfile.open(file_obj, 'rb') as inner_stream:
                 for j, xzinfo in enumerate(inner_stream):
                     print(type(xzinfo))
                     yield xzinfo.name, xzinfo