Lora to do lineart / keyframe / keynote style. Trigger: LineartWorks best on simple character prompts rather than complex scenes, since the lora training data focussed on people. Function varies with model, and seems to work better without "score" tags (or if including score tags, include some of the lower numbers not just 7+). Can occasionally go super wonky with more complicated poses or scenes. Trained mostly off images from Symphogear and Madoka, so magical girl tags and bodysuit/battle tags can occasionally have style bleed, varies with prompt and model. Not too compatible with other style loras, for obvious reasons (this is not supposed to have shading, for example), but seemed to work OK with character loras. Trigger: Lineart

lineart / keyframe / keynote / concept style

lineart_keyframe_key_anime_style.safetensors

v1.0

Lora to do lineart / keyframe / keynote style. Trigger: LineartWorks best on simple character prompts rather than complex scenes, since the lora training data focussed on people. Function varies with model, and seems to work better without "score" tags (or if including score tags, include some of the lower numbers not just 7+). Can occasionally go super wonky with more complicated poses or scenes. Trained mostly off images from Symphogear and Madoka, so magical girl tags and bodysuit/battle tags can occasionally have style bleed, varies with prompt and model. Not too compatible with other style loras, for obvious reasons (this is not supposed to have shading, for example), but seemed to work OK with character loras.Trigger: Lineart

Files with SHA256: c3a722d24f167b5053b13af373f5dd909ad0811b9007176410d093a89ca70179