Remove the ASR template until the datasets>=2.0 API stabilizes

superb_demo.py

@@ -145,7 +145,6 @@ class Superb(datasets.GeneratorBasedBuilder):
             supervised_keys=("file", "text"),
             url="http://www.openslr.org/12",
             data_url="data/LibriSpeech-test-clean.zip",
-            task_templates=[AutomaticSpeechRecognition(audio_file_path_column="file", transcription_column="text")],
         ),
         SuperbConfig(
             name="ks",