- Fixing split generation

dv-presidential-speech.py

@@ -113,7 +113,7 @@ class DhivehiPresidentialSpeech(datasets.GeneratorBasedBuilder):
                 gen_kwargs={
                     'local_extracted_archive': local_extracted_archive[split],
                     'audio_files': dl_manager.iter_archive(audio_path[split]),
-                    'metadata_path': dl_manager.download_and_extract(metadata_path[split]),
+                    'metadata_path': metadata_path[split],
                     'path_to_clips': f'{path_to_clips}-{split}/waves',
                 },
             ) for split in split_type