Fixes

audiomnist.py

@@ -23,6 +23,8 @@ import os
 
 import datasets
 
+from glob import glob
+
 logger = logging.getLogger(__name__)
 
 _DESCRIPTION = """\
@@ -96,14 +98,12 @@ class GraphemeToPhoneme(datasets.GeneratorBasedBuilder):
     def _split_generator(self, dl_manager, split):
         archive_url = self._get_url(split)
         archive_path = dl_manager.download_and_extract(archive_url)
-        archive = dl_manager.iter_archive(archive_url)
         meta_url = self._get_meta_url()
         meta_file = dl_manager.download(meta_url)
         speaker_map = self._get_speaker_map(meta_file)
         return datasets.SplitGenerator(
             name=split,
             gen_kwargs={
-                "archive": archive,
                 "archive_path": archive_path,
                 "speaker_map": speaker_map,
             },
@@ -128,8 +128,9 @@ class GraphemeToPhoneme(datasets.GeneratorBasedBuilder):
         del result["recordingroom"]
         return result
 
-    def _generate_examples(self, archive, archive_path, speaker_map):
-        for path, sample_file in archive:
+    def _generate_examples(self, archive_path, speaker_map):
+        wav_files = glob(os.path.join(archive_path, 'dataset', '**', '*.wav'))        
+        for path in wav_files:
             match = _RE_FILE_NAME.search(path)
             if not match:
                 logger.warn(
@@ -139,12 +140,13 @@ class GraphemeToPhoneme(datasets.GeneratorBasedBuilder):
             digit, speaker_id = [
                 match.group(key) for key in ["digit", "speaker_id"]
             ]
-            sample = {
-                "digit": digit,
-                "speaker_id": speaker_id,
-                "file_name": os.path.join(archive_path, path),
-                "audio": {"path": path, "bytes": sample_file.read()},
-            }
+            with open(path, 'rb') as wav_file:
+                sample = {
+                    "digit": digit,
+                    "speaker_id": speaker_id,
+                    "file_name": os.path.join(archive_path, path),
+                    "audio": {"path": path, "bytes": wav_file.read()},
+                }
             if speaker_id not in speaker_map:
                 logger.warn(f"Speaker {speaker_id} not found")
             speaker_info = speaker_map[speaker_id]