|
gptkbp:instanceOf
|
gptkb:model
|
|
gptkbp:advantage
|
high sample quality
mode coverage
|
|
gptkbp:application
|
style transfer
super-resolution
inpainting
conditional generation
unconditional generation
|
|
gptkbp:basedOn
|
Markov chain
|
|
gptkbp:contrastsWith
|
gptkb:VAE
gptkb:GAN
|
|
gptkbp:firstPublished
|
2015
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
Diffusion Model
|
|
gptkbp:improves
|
gptkb:Latent_Diffusion_Model
DDIM
Score-based Generative Model
|
|
gptkbp:limitation
|
computationally intensive
slow sampling
|
|
gptkbp:openSource
|
gptkb:Stable_Diffusion
gptkb:OpenAI_Guided_Diffusion
|
|
gptkbp:popularizedBy
|
gptkb:Stable_Diffusion
gptkb:DALL-E_2
gptkb:Imagen
gptkb:Denoising_Diffusion_Probabilistic_Models
|
|
gptkbp:proposedBy
|
Jascha Sohl-Dickstein
|
|
gptkbp:relatedTo
|
deep learning
neural networks
transformers
Bayesian inference
denoising
score-based generative models
probabilistic modeling
stochastic differential equations
|
|
gptkbp:trainer
|
iterative denoising
reverse diffusion process
|
|
gptkbp:usedBy
|
gptkb:Adobe_Firefly
gptkb:Microsoft_Designer
gptkb:Midjourney
gptkb:DreamStudio
gptkb:Stability_AI
gptkb:RunwayML
Canva Magic Media
Google Imagen
OpenAI DALL-E 2
|
|
gptkbp:usedIn
|
image generation
text-to-image synthesis
video generation
audio generation
|
|
gptkbp:bfsParent
|
gptkb:Foundation_Model
|
|
gptkbp:bfsLayer
|
6
|