VQGAN

GPTKB entity

Statements (51)
Predicate Object
gptkbp:instanceOf gptkb:model
gptkbp:affiliation gptkb:Heidelberg_University
gptkbp:architecture gptkb:convolutional_neural_network
Vector Quantization
gptkbp:category gptkb:artificial_intelligence
gptkb:machine_learning
computer vision
gptkbp:citation 1000+
Esser, P., Rombach, R., & Ommer, B. (2021). Taming Transformers for High-Resolution Image Synthesis. arXiv preprint arXiv:2012.09841.
gptkbp:component gptkb:transformation
decoder
encoder
GAN discriminator
vector quantizer
gptkbp:developedBy gptkb:Björn_Ommer
gptkb:Patrick_Esser
gptkb:Robin_Rombach
gptkbp:feature efficient training
combines VQ-VAE and GAN
discrete latent space
high-resolution synthesis
supports transformer-based modeling
gptkbp:format gptkb:Tensor
gptkbp:fullName Vector Quantized Generative Adversarial Network
https://www.w3.org/2000/01/rdf-schema#label VQGAN
gptkbp:influenced gptkb:Midjourney
gptkb:Stable_Diffusion
DALL-E mini
gptkbp:input gptkb:illustrator
gptkbp:language gptkb:Python
gptkbp:license gptkb:MIT_License
gptkbp:notableFor text-to-image synthesis
VQGAN+CLIP art generation
gptkbp:notablePublication Taming Transformers for High-Resolution Image Synthesis
https://arxiv.org/abs/2012.09841
gptkbp:openSource yes
gptkbp:output gptkb:illustrator
gptkbp:publishedIn 2021
gptkbp:relatedTo gptkb:CLIP
gptkb:GAN
gptkb:VQ-VAE
gptkbp:repository https://github.com/CompVis/taming-transformers
gptkbp:trainer gptkb:COCO
gptkb:ImageNet
gptkb:OpenImages
image datasets
gptkbp:usedFor image generation
image synthesis
gptkbp:bfsParent gptkb:AI_artists
gptkb:VQGAN+CLIP
gptkbp:bfsLayer 7