chunked/nltk_chunking/model_doc_tapex/chunk_7.txt

However, it is
still a challenge to apply pre-training on structured tabular data due to the absence of large-scale high-quality tabular data.