CvT

GPTKB entity

Statements (19)
Predicate Object
gptkbp:instanceOf gptkb:convolutional_neural_network
gptkbp:appliesTo computer vision
gptkbp:arXivID 2103.15808
gptkbp:codeAvailableAt https://github.com/microsoft/CvT
gptkbp:combines convolutional neural networks
transformers
gptkbp:contribution incorporates convolutional layers into transformer blocks
improves efficiency and performance over vanilla vision transformers
gptkbp:developedBy gptkb:Microsoft_Research
gptkbp:fullName Convolutional vision Transformer
https://www.w3.org/2000/01/rdf-schema#label CvT
gptkbp:improves gptkb:ViT
gptkbp:introducedIn 2021
gptkbp:notablePublication CvT: Introducing Convolutions to Vision Transformers
gptkbp:usedFor image classification
object detection
semantic segmentation
gptkbp:bfsParent gptkb:Vision_Transformer
gptkbp:bfsLayer 6