chunked/nltk_chunking/model_doc_vit_hybrid/chunk_3.txt

The abstract from the paper is the following:
While the Transformer architecture has become the de-facto standard for natural language processing tasks, its
applications to computer vision remain limited.