Update dataset.py

dataset.py

@@ -32,14 +32,8 @@ class CustomDataset(datasets.GeneratorBasedBuilder):
     def _split_generators(self, dl_manager):
         audio_path = dl_manager.download_and_extract(_DATA_URL+"/audio.zip")
         csv_path = dl_manager.download_and_extract(_DATA_URL+"/comma.csv")
-        # Use the correct paths for "audio.zip" and "comma.csv" based on the provided file structure
-        # audio_path = dl_manager.extract(os.path.join(archive_path, "audio.zip"))
-        # csv_path = os.path.join(archive_path, "comma.csv")
+        audio_path = os.path.join(audio_path, "audio")
 
-        audio_path = dl_manager.extract(os.path.join(_DATA_URL, "audio.zip"))
-        csv_path = os.path.join(_DATA_URL, "comma.csv")
-
-        
         return [
             datasets.SplitGenerator(
                 name=datasets.Split.TRAIN,
@@ -54,7 +48,7 @@ class CustomDataset(datasets.GeneratorBasedBuilder):
             for row in csv_reader:
                 original_sentence_id, sentence, locale = row.values()
                 audio_file = f"{original_sentence_id}.mp3"
-                audio_file_path = audio_path + "/" + audio_file
+                audio_file_path = os.path.join(audio_path, audio_file)
 
                 yield key, {
                     "audio": audio_file_path,