Update openwebtext_split.py

openwebtext_split.py

@@ -19,6 +19,7 @@ import os
 import re
 from itertools import chain
 import tarfile
+import io
 
 import datasets
 
@@ -56,8 +57,9 @@ def custom_iter_archive(path_or_buf, _filter=lambda x: True):
                 continue
             file_obj = stream.extractfile(tarinfo)
             for j, xzinfo in enumerate(file_obj):
-                print(xzinfo.decode())
-                yield xzinfo.name, xzinfo
+                f = io.BytesIO(xzinfo)
+                print(dir(f))
+                yield f.name, f
             stream.members = []
         del stream
     if hasattr(path_or_buf, "read"):