Update openwebtext_split.py

openwebtext_split.py

@@ -57,7 +57,7 @@ def custom_iter_archive(path_or_buf, _filter=lambda x: True):
             if not file_path.endswith('xz'):
                 continue
             file_obj = stream.extractfile(tarinfo)
-            decompressed = tarfile.open(fileobj=file_obj, mode='r:*')
+            decompressed = tarfile.open(fileobj=file_obj, mode='r:xz')
             for j, xzinfo in enumerate(decompressed):
                 if not xzinfo.name.endswith('txt'):
                     continue
@@ -105,5 +105,5 @@ class Openwebtext(datasets.GeneratorBasedBuilder):
     def _generate_examples(self, files):
         """Yields examples."""
         for idx, f in enumerate(files):
-            print(type(f))
+            print(f.readlines())
             yield idx, {"text": re.sub("\n\n\n+", "\n\n", f.read()).strip()}