fixing indentation and adding zipped version

ml-superb-subset.py

@@ -102,18 +102,19 @@ class MLSuperb(datasets.GeneratorBasedBuilder):
 
                     yield id_, {
                         "audio": wavs_path + row[0],
-                        "sentence": " ".join(row[2:]),
+                        "sentence": "".join(row[2:]),
                         "id": id_,
                     }
             else:
                 for row in reader:
+                    # print(row)
                     id_ = row[0]
                     if not row[0].endswith(".wav"):
                         row[0] += ".wav"
                     metadata[row[0]] = row[-1]
 
-                yield id_, {
-                        "audio": wavs_path + row[0],
-                        "sentence": row[-1],
-                        "id": id_,
-                    }
+                    yield id_, {
+                            "audio": wavs_path + row[0],
+                            "sentence": row[-1],
+                            "id": id_,
+                        }

ml_superb_subset.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec205f088507141b315496fc09c958397f3a04b6c3fca009cc600103ec3f1d4c
+size 1934801121