rm sentence

fleurs.py

@@ -104,7 +104,6 @@ class Fleurs(datasets.GeneratorBasedBuilder):
                 "audio": datasets.Audio(sampling_rate=48_000),
                 "transcription": datasets.Value("string"),
                 "raw_transcription": datasets.Value("string"),
-                "sentence": datasets.Value("string"),
                 "lang_id": datasets.ClassLabel(names=langs),
                 "language": datasets.Value("string"),
                 "lang_group_id": datasets.ClassLabel(
@@ -201,13 +200,11 @@ class Fleurs(datasets.GeneratorBasedBuilder):
             ) = line.strip().split("\t")
 
             lang_group = _FLEURS_LANG_TO_GROUP[lang_id]
-            sentence = tp.normalize(raw_transcription)
 
             data[file_name] = {
                 "id": int(_id),
                 "raw_transcription": raw_transcription,
                 "transcription": transcription,
-                "sentence": sentence,
                 "num_samples": int(num_samples),
                 "lang_id": _FLEURS_LANG.index(lang_id),
                 "language": _FLEURS_LANG_TO_LONG[lang_id],