rinongal / textual_inversion

MIT License
2.87k stars 278 forks source link

what are the difference between models in ldm and first_stage_models? #160

Open Derry-Xing opened 10 months ago

Derry-Xing commented 10 months ago

Same as the title.

rinongal commented 10 months ago

First stage models are the VAE that encodes the images into the latent space. These bits of code and config are part of the original compviz LDM code.