# Convolutional Enhancement

CvT-21 is a vision transformer model pre-trained on the ImageNet-1k dataset, improving traditional vision transformers by introducing convolutional operations.