Fix bug in using degree_as_nlabel for GIN dataset (#1451)

In order to make the `label2idx` dictionary, tensor values should be used and not the tensor itself.
Co-authored-by: Jinjing Zhou <VoVAllen@users.noreply.github.com>

python/dgl/data/gindt.py

@@ -220,7 +220,7 @@ class GINDataset(object):
                     # but usually no features means no labels, fine.
                     g.ndata['label'] = g.in_degrees()
                     # extracting unique node labels
-                    nlabel_set = nlabel_set.union(set(g.ndata['label']))
+                    nlabel_set = nlabel_set.union(set([F.as_scalar(nl) for nl in g.ndata['label']]))
 
                 nlabel_set = list(nlabel_set)
                 # in case the labels/degrees are not continuous number