chunked/nltk_chunking/model_doc_layoutlmv2/chunk_78.txt

In case you want all wordpieces of a word to be labeled, you can
initialize the tokenizer with only_label_first_subword set to False.