BertTokenClassifier

What is the significance of ignore_label=[‘O’] in the following code?

from bert_sklearn import BertTokenClassifier
model = BertTokenClassifier(bert_model=‘bert-base-cased’,
max_seq_length=178,
epochs=3,
gradient_accumulation_steps=4,
learning_rate=5e-5,
train_batch_size=16,
eval_batch_size=16,
validation_fraction=0.,
label_list=label_list,
ignore_label=[‘O’])