In typical image generation models, there is a very strong bias toward maintaining symmetry in poses.
Source du modèle
Description de la source
In typical image generation models, there is a very strong bias toward maintaining symmetry in poses.
This LoRA model helps generate asymmetrical images by being trained on images where the legs are in asymmetrical poses.
It is specifically designed for scenes where a woman straddles a man (suggestive).
足が非対称な感じにしやすくするLoRA
Sources
1 sourceArtefacts du modèle
1 artefactunevenleg_il_v1.safetensors
safetensors · BF16 · 218 MB · SHA-256 1b3a75bb9792…2f81 · Civitai
TéléchargerExtraits de sources
1 extraitSource context: 383 downloads · Type LORA · Base model Illustrious