Changed data script so it takes a list of files instead of a directory

wikipedia-bert-128.py

@@ -117,18 +117,17 @@ class WikipediaBERT128(datasets.GeneratorBasedBuilder):
         # dl_manager is a datasets.download.DownloadManager that can be used to download and extract URLs
         # It can accept any type or nested list/dict and will give back the same structure with the url replaced with path to local files.
         # By default the archives will be extracted and a path to a cached folder where they are extracted is returned instead of the archive
-        data_dir = os.path.abspath(os.path.expanduser(dl_manager.manual_dir))
         return [
             datasets.SplitGenerator(
                 name=datasets.Split.TRAIN,
                 # These kwargs will be passed to _generate_examples
                 gen_kwargs={
-                    "filepath": data_dir,
+                    "data_files": self.config.data_files["train"],
                 },
             ),
         ]
 
-    def _generate_examples(self, filepath):
+    def _generate_examples(self, data_files):
         """ Yields examples as (key, example) tuples. """
         # This method handles input defined in _split_generators to yield (key, example) tuples from the dataset.
         # The `key` is here for legacy reason (tfds) and is not important in itself.
@@ -142,10 +141,8 @@ class WikipediaBERT128(datasets.GeneratorBasedBuilder):
             'masked_lm_ids',        # masked_lm_labels=None      : label of masked tokens with padding as 0.
             'next_sentence_labels'  # next_sentence_label=None   : 1 if next sentence, 0 otherwise
         )
-        tfrecords = Path(filepath).glob("*.tfrecord")
-
         highest_id_ = -1
-        for rec in tfrecords:
+        for rec in data_files:
             reader = tfrecord_loader(rec, None, list(TFRECORD_KEYS))
             for id_, d in enumerate(reader, start=highest_id_+1):
                 highest_id_ = id_
@@ -164,3 +161,4 @@ class WikipediaBERT128(datasets.GeneratorBasedBuilder):
                     "labels": labels,
                     "next_sentence_label": d["next_sentence_labels"]
                 }
+