An Illustrious LoRA designed to achieve a unique style that mixes the characteristic style of Anime illustrations with a Western Cartoon-Comic aesthetic.
Compatible with Anime-based and Cartoon-based checkpoints (Cartoon-based checkpoints recomended for better results)
KaliMichi's Style - Cel Shading v1.6 HotFix:
Enhanced hand and foot generation, alongside lighting refinements. This update addresses style-specific rendering flaws from version 1.0. I employed a focused fine-tuning approach on handsand feet to yield highly consistent results.
Note: The style and level of detail will be refined in later versions to truly capture that unique style. Additionally, the final results may vary depending on the base checkpoint you use.
Please note that some Anime based Models can force their Anime aesthetic and mitigate the LoRA's style, in that case some images might result in a more Animeish look.
Recommended Settings:
LoRA Weight: 1-2 (Weights lower than 0.9 can lead to suboptimal results for the style.)
Sampler: Euler a
Steps: 30-40
CFG Scale: 3-5
Note: Because of the training data, please make sure to specify the character's clothing in detail to avoid accidentally generating inappropriate content.
Description
KaliMichi's Style - Cel-Shading AM v1.0
My second step into the world of Anima! This is the initial release of the KaliMichi's Style - Cel Shading LoRA for Anima (also available for Illustrious).
v1.0 | Initial Release
The first version release of KaliMichi - Cel Shading is designed to blend the aesthetics of Japanese anime with the charm and expressiveness of Western cartoon styles, creating a unique mixed-rendering look in a Cel-Shading aesthetic.
⚠️ Known Issues & Recommended Settings
some issues with anatomy in more complex compositions that might require the support of certain concept LoRAs.
The results might not feel very pronounced; I am working on addressing this in future versions to better capture the signature style of the series: KaliMichi's Style.
What's Next?
I'm sorry for any inconvenience these issues may cause. I'll do my best to address them and improve the LoRA in future updates, please let me know if you find one or more so I can address it as soon as possible.
Thank you for your support and for giving my work a chance! Your support motivates me to keep learning, refining this LoRA, and bringing you bigger and better projects in the future. 💖