To use the model, add the following from the transformers package:
(1) Tokenizer:
tokenizer = AutoTokenizer.from_pretrained("yikuan8/Clinical-BigBird")
(2) Model type:
#T in [1,2,3,4]
num_classes = 4
model = BigBirdForSequenceClassification.from_pretrained(directory, num_labels=num_classes)