这是Anim008的第一个模型作品，欢迎使用。这是一款SDXL Turbo类的融合模型。在源模型基础上混合了写实造型训练图，速度快，国画风格强烈，尤善半小写意的工笔，造型丰富度较多。
本意此模型制作是为了快速抽卡出图还有垫图扭画风，所以没有在细节最终出图质量上打磨，但经几个玩家群测试，经调整也有不错的画质出品，本人跑的范例图不是最佳品质，大家可以多尝试各种配置。
附上当前两个版本的测试对比：
................................................................
关键词(没有其实也行)：
traditional chinese ink painting
建议有效参数：
采样器：DPM++ SDE Karras（色彩丰富，细节清晰，类似国画小写意），Euler a（速度快，墨色较重，细节稍次，更偏工笔）
Size>768；
Steps>7 (速度较快) >8<11 (画质较好)；
CFG Scale=1.5<3 (速度较快)<4 (画质较好)；
................................................................
建议搭配Lora：
https://civarchive.com/models/190242
Lora关键词(Style Tag About Lora)：ink ；
................................................................
测试数据：
在4090显卡上生成768x1280分辨率，可在2秒左右出图。
另外提醒，如需要增加画质，Steps加到40依然有画质提升。但务必控制CFG<4！
................................................................
感谢源模型提供：
古风大模型XL GuFengXL，
https://civarchive.com/models/189523
及Stability Ai的Turbo源模，
https://huggingface.co/stabilityai/sdxl-turbo/tree/main
好了，请一起来感受Turbo的速度吧！
---------------------------------------------------------------------------------
This is Anim008 first model, welcome to use.This a SdXL Turbo Mix Model,
................................................................
Trigger Words(about Model)：
traditional chinese ink painting
Suggested useful configuration:
................................................................
Sampler:
DPM++ SDE Karras
（Rich colors, clear details and line, similar to traditional Chinese painting freehand brushwork），
Euler a
（The ink color is heavier, the details are slightly inferior, but it leans more towards meticulous brushwork）
................................................................
Size>768；
Steps>7 (speed fast) >8<11 (high quality)；
CFG Scale=1.5<3 (speed fast)<4 (high quality)；
................................................................
Suggested adding Lora:
https://civarchive.com/models/190242
Trigger Words(about Lora)：ink ;
................................................................
About the Test Data:
Run a resolution of 768x1280 on my 4090 graphics card, which can output 1 images in about 2 seconds.
If you need to improve the picture quality, you can increase the Steps to 20<40, But please note that CFG must be controlled to be less than 4！
................................................................
Thanks the source model for providing:
GuFengXL
https://civarchive.com/models/189523
And Thanks the Turbo source model From Great Stability Ai，
https://huggingface.co/stabilityai/sdxl-turbo/tree/main
So, Let 's Enjoy Turbo Speed!
Description
古风XL Turbo升级v1.2版，主要提升了造型准确度，画质，
升级对比测试图：
................................................................
关键词(没有其实也行)：
traditional chinese ink painting
建议有效参数：
采样器：DPM++ SDE Karras（色彩丰富，细节清晰，类似国画小写意），
Euler a（墨色较重，细节稍次，更偏工笔），
Size>768；
Steps>7；
CFG Scale=1.5<3；
................................................................
建议搭配Lora：
https://civitai.com/models/190242
Lora关键词(Style Tag About Lora)：ink ；
................................................................
测试数据：
在4090显卡上生成768x1280分辨率，可在2秒左右出图。
另外提醒，如需要增加画质，Steps加到40依然有画质提升。但务必控制CFG<3！
................................................................
感谢源模型提供：
古风大模型XL GuFengXL，
https://civitai.com/models/189523
及Stability Ai的Turbo源模，
https://huggingface.co/stabilityai/sdxl-turbo/tree/main
好了，请一起来感受Turbo的速度吧！
---------------------------------------------------------------------------------
This is Anim008 first model, welcome to use.This a SdXL Turbo Mix Model,
................................................................
Trigger Words(about Model)：
traditional chinese ink painting
Suggested useful configuration:
................................................................
Sampler:
DPM++ SDE Karras
（Rich colors, clear details and line, similar to traditional Chinese painting freehand brushwork），
Euler a
（The ink color is heavier, the details are slightly inferior, but it leans more towards meticulous brushwork）
................................................................
Size>768；
Steps>7；
CFG Scale=1.5<3；
................................................................
Suggested adding Lora:
https://civitai.com/models/190242
Trigger Words(about Lora)：ink ;
................................................................
About the Test Data:
Run a resolution of 768x1280 on my 4090 graphics card, which can output 1 images in about 2 seconds.
If you need to improve the picture quality, you can increase the Steps to 20<40, But please note that CFG must be controlled to be less than 3！
................................................................
Thanks the source model for providing:
GuFengXL
https://civitai.com/models/189523
And Thanks the Turbo source model From Great Stability Ai，
https://huggingface.co/stabilityai/sdxl-turbo/tree/main
So, Let 's Enjoy Turbo Speed!