chunked/nltk_chunking/model_doc_layoutlmv2/chunk_7.txt

Meanwhile, it also integrates a spatial-aware self-attention
mechanism into the Transformer architecture, so that the model can fully understand the relative positional
relationship among different text blocks.