chunked/nltk_chunking/model_doc_cvt/chunk_1.txt

The Convolutional vision Transformer (CvT) improves the Vision Transformer (ViT) in performance and efficiency by introducing convolutions into ViT to yield the best of both designs.