该模型主要表现为原神剧情PV中的平涂风格，基于AnimagineXLv3.0训练。
推荐参数参考例图，在某些情况下，对手指的表现似乎比原版更好（
无需触发词，因为是训练画风，因此模型大小会比单角色要大一点，这是在正常范围内的
The model primarily embodies the flat style seen in the Genshin Impact storyline PV, trained on AnimagineXLv3.0.
Recommended parameters are based on reference images. In certain cases, finger representation appears to be even better than the original version.
There's no need for a trigger word, as it's trained on art style. Therefore, the model size might be slightly larger than a single character, which falls within the normal range.
Description
FAQ
Comments (2)
Can i have your dataset and train settings? Your lora performance is GREAT!
The dataset was obtained by capturing the original PV video of Genshin Impact. I did not remove the metadata of the LORA training parameters, which can usually be viewed in the training parameter section on the LORA options page in sd-webui, or by using the plugin https://github.com/kohya-ss/sd-webui-additional-networks.git to view the training parameters.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.