Have a ComfyUI workflow? Get a readable Workflow Doc in your browser.Make a workflow readable.Try it now ↗
本LyCORIS在 丰富的提示词下表现良好 ,在 少量 提示词下表现不佳。添加 光照 相关的tag将会很大程度改善画面,请您完善光照相关描述
Model source
Source description
README
Artist style shiej007-NOOBAI-XL- danbooru
ZH
强烈建议您使用noobai或noobai衍生模型运行此lora
推荐分辨率:1664*896
推荐的权重为:0.9
触发词:s007.
本LyCORIS在 丰富的提示词下表现良好 ,在 少量 提示词下表现不佳。添加 光照 相关的tag将会很大程度改善画面,请您完善光照相关描述
遇到的部分问题:
1.一些概念过拟,当提示词出现手枪时,人物肢体很有可能崩坏。
Sources
1 sourceModel artifacts
1 artifactNoob080Trigger触发词--s007--.safetensors
safetensors · 178 MB · SHA-256 d830f06088e6…1f25 · Civitai
DownloadSource excerpts
1 excerptSource context: 405 downloads · Type LoCon · Base model NoobAI
2.画面偏灰,饱和度偏低,失去了vpre的优势。
3.在描述过少的时候,画面表现平平无奇。
EN
It is highly recommended to run this LoRA using the noobai or noobai-derived models.
Recommended resolution: 1664*896
Recommended weight: 0.9
Trigger word: s007.
This LyCORIS performs well with rich prompts but poorly with fewer prompts. Adding lighting-related tags will significantly improve the image, so please enhance the lighting-related descriptions.
Some encountered issues:
Overfitting of certain concepts, where the appearance of a pistol in the prompts may lead to distorted character limbs.
The images tend to be grayish with low saturation, losing the advantage of vpre.
When descriptions are too sparse, the image results are rather unremarkable.
loratest
