Trained on same database, but this time, using NoobAI as base model. [EPS]
This LoRA is based on a character from a game that I do not own.
Coco Grimalkin, the other characters, and Purrfect Apawcalypse, and the training images [which were all of Coco's sprites] belong to 90% Studios.
Please do NOT use this LoRA for commercial purposes.
This LoRA will be taken down if the creators wish to do so.
Generates Coco in the PA1 and newer style. This LoRA is my second, and also follows the game's artstyle much more closely than the Olive Higgins one.
Trigger Words:
coco_grimalkin, cat, cute, furry,Base Model:
NAI-XL Epsilon-Pred 1.1 Version (V1.1)
Tested on:
KonanMix NoobAI XL v1.0.
Settings:
V1
Sampling method: Euler a
Schedule type: Automatic
Sampling steps: 20 (30-40 for better quality)
CFG Scale: 7
Resolution [Starting point, change if needed!]: 824x1216
LoRA Weight: 0.9-1.0
V1.1 (New!)
Sampling method: DPM++ 2M SDE / DPM++ 3M SDE / HeunPP2
Schedule type: Karras (Automatic is also fine if you don't want to mess with schedule types)
Sampling steps: 20 or more
CFG Scale: 5.5
Resolution [Starting point, change if needed!]: 824x1216
LoRA Weight: 0.7-1.0
Trained on same database, but this time, using NoobAI as base model. [EPS]
1. 转载模型仅供学习与交流分享,其版权及最终解释权归原作者。
2. 模型原作者如需认领模型,请通过官方渠道联系海艺AI工作人员进行认证。我们致力于保护每一位创作者的权益。 点击去认领
