hadnling space seaprated files

ml-superb-subset.py

@@ -91,6 +91,8 @@ class MLSuperb(datasets.GeneratorBasedBuilder):
         metadata = {}
         with open(transcript_path, encoding="utf-8") as f:
             reader = csv.reader(f, delimiter="\t", quoting=csv.QUOTE_NONE)
+            if len(reader[0]) == 1:
+                reader = csv.reader(f, delimiter=" ", quoting=csv.QUOTE_NONE)
             for row in reader:
                 id_ = row[0]
                 if not row[0].endswith(".wav"):