# let's check tokenization process
input_ids = tokenizer(text).input_ids
subword_view = [tokenizer.convert_ids_to_tokens(id) for id in input_ids]
np.array(subword_view)