Another badly trained lora.
Recommended to use 3:2 ratio (832x1216).
Add 'card \(medium\), copyright name' to negative prompts. (Unless desired to create card borders similar to VC)
Model can recognize some consistent characters (Oracle, Alchemist, Pixie, Fortuna, Calamity, Demise, etc...) and some character progressions ('Rookie Hatsune, Ninja Hatsune, Crimson Hatsune, Master Hatsune' however they been summed into single tag by their generic name 'Hatsune \(shinkai no valkyrie\)')
Due to tagging error, some cards are not included in the tags or tagged erroneously. (Minister cards, Astral)
Trained in both Lora (Standard) and Lycoris format
Description
More training epochs.
Slightly different training setup.
FAQ
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.







