Model
E2E-INVAE
E2E-INVAE is a variational autoencoder with a semantic latent space. Its latent variables carry more interpretable structural information, which benefits downstream generation tasks.
Timeline (last 90 days)
Oct 14
MiniDog adopts E2E-INVAE because its semantic latent space carries more interpretable structural information than a conventional VAE.
Unverified50%