Use Sequence for labels features

Hi !
I noticed that the labels features were not wrapped in a `datasets.Sequence` object, but directly in a python list, is there a specific reason for that ? as I used the [run_classification.py](https://github.com/huggingface/transformers/blob/main/examples/pytorch/text-classification/run_classification.py#L460C12-L460C68) HuggingFace example script I ran into the error when testing : raw_datasets["train"].features["label"].dtype == "list" for multi label classification that's what made me notice this.

hallmarks_of_cancer.py

@@ -118,7 +118,7 @@ class HallmarksOfCancerDataset(datasets.GeneratorBasedBuilder):
                 {
                     "document_id": datasets.Value("string"),
                     "text": datasets.Value("string"),
-                    "label": [datasets.ClassLabel(names=_CLASS_NAMES)],
+                    "label": datasets.Sequence(datasets.ClassLabel(names=_CLASS_NAMES)),
                 }
             )