|
gptkbp:instanceOf
|
gptkb:convolutional_neural_network
|
|
gptkbp:advantage
|
enables high-resolution output
focuses on local structure
reduces number of parameters
|
|
gptkbp:application
|
image-to-image translation
style transfer
super-resolution
|
|
gptkbp:architecture
|
gptkb:convolutional_neural_network
|
|
gptkbp:input
|
image patches
|
|
gptkbp:introduced
|
gptkb:Phillip_Isola
gptkb:Tinghui_Zhou
gptkb:Alexei_A._Efros
gptkb:Jun-Yan_Zhu
|
|
gptkbp:introducedIn
|
2016
pix2pix paper
|
|
gptkbp:mainFunction
|
classifies image patches as real or fake
|
|
gptkbp:notablePublication
|
gptkb:Image-to-Image_Translation_with_Conditional_Adversarial_Networks
https://arxiv.org/abs/1611.07004
|
|
gptkbp:output
|
matrix of probabilities
|
|
gptkbp:platform
|
gptkb:TensorFlow
gptkb:PyTorch
|
|
gptkbp:relatedTo
|
gptkb:CycleGAN
gptkb:pix2pix
|
|
gptkbp:size
|
variable (e.g., 70x70, 16x16)
|
|
gptkbp:usedIn
|
gptkb:Generative_Adversarial_Networks
|
|
gptkbp:bfsParent
|
gptkb:Pix2Pix
|
|
gptkbp:bfsLayer
|
5
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
PatchGAN discriminator
|