card-style is the token for versions 1 and 2. The new version 3 was trained from the ground up using the same dataset and has a new token, cardstyleV3 .
Fuente del modelo
Descripción de la fuente
Retrained from the ground up with the same dataset. Version 3 is the closest yet to the training images style. All sample images use the v2-1_768-nonema-pruned.ckpt checkpoint. Note the token name is changed to cardstyleV3.
Fuentes
1 fuenteArtefactos del modelo
1 artefactoExtractos de fuentes
2 extractosSource context: 1893 downloads · Type TextualInversion · Base model SD 2.1 768
Source context: 1843 downloads · Type TextualInversion · Base model SD 2.1 768