Update for version 2.7.1

It seems that in datasets 2.7.1 the AutomaticSpeechRecognition task template has changed and the parameter audio_file_path_column="file" has changed to audio_column="audio" as can be seen in [here](https://github.com/huggingface/datasets/blob/99680a7b5946d41873d84e16b30a2587b1f935fa/src/datasets/tasks/automatic_speech_recognition.py#L14) . By default the audio_column is "audio" and that exactly our feature so it should work with the default value.

parlament_parla.py

@@ -59,7 +59,7 @@ _SPLITS = {datasets.Split.TRAIN: "train", datasets.Split.VALIDATION: "dev", data
 class ParlamentParla(datasets.GeneratorBasedBuilder):
     """ParlamentParla."""
 
-    VERSION = datasets.Version("2.0.0")
+    VERSION = datasets.Version("2.7.1")
 
     BUILDER_CONFIGS = [
         datasets.BuilderConfig(name="clean", version=VERSION, description="211 hours of clean quality segments."),
@@ -84,7 +84,7 @@ class ParlamentParla(datasets.GeneratorBasedBuilder):
             license=_LICENSE,
             citation=_CITATION,
             task_templates=[
-                AutomaticSpeechRecognition(audio_file_path_column="path", transcription_column="sentence")
+                AutomaticSpeechRecognition(transcription_column="sentence")
             ],
         )