花间一壶酒,独酌无相亲。举杯邀明月,对影成三人。
——唐 李白
大家好,我是chosen鲷,这次带来的是花间酒lora。
训练这个模型的初衷是因为想要一个,如同古诗般优美意境的风格,所以搜集了许多国风水彩的插画,作为训练集训练了这个模型,男性女性均可,国风建筑也可以,总体有一种水彩晕染的意味,可以作为一种风格挂件使用。
lora兼容性蛮好的,但是由于构图上人物占比较小,容易崩脸,建议打开ADetailer,建议权重0.8。触发词:Hanama wine
推荐的tag:
Hanama wine
这里写想要的内容
<lora:Hanama wine-000018:0.8>,
Negative prompt: EasyNegative,badhandsv5-neg,Subtitles,word,
Steps: 32, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 2553923380, Size: 512x768, Model hash: e3edb8a26f, Model: ghostmix_v20Bakedvae, Denoising strength: 0.5, Clip skip: 2, ADetailer model: face_yolov8n.pt, ADetailer confidence: 0.3, ADetailer dilate erode: 4, ADetailer mask blur: 4, ADetailer denoising strength: 0.4, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer version: 23.11.0, Hires upscale: 2, Hires steps: 12, Hires upscaler: R-ESRGAN 4x+, Lora hashes: "Hanama wine-000018: 7d3593e99f39", TI hashes: "EasyNegative: 66a7279a88dd, badhandsv5-neg: aa7651be154c", Version: v1.6.0-2-g4afaaf8a
大模型我推荐用ghost或者chosen_Chinese_styleV2以达到最佳效果(下方推荐),其他的常见常见二次元模型也均可,推荐使用偏二次元的模型。
感谢大家的下载使用,如果喜欢的话欢迎关注我,我的bilibili pixiv ID均为chosen鲷,欢迎大家关注我的频道https://space.bilibili.com/3867754?spm_id_from=333.1007.0.0,https://www.pixiv.net/users/40911695 谢谢。
The following content is from Google Translate
Hello everyone, I am chosen鲷, and this time I bring you Hanama wine lora.
The original intention of training this model was to create a style with a beautiful poetic atmosphere, so I collected many Chinese-style watercolor illustrations as the training set for this model. Both male and female characters, as well as Chinese-style architecture, can be generated with a watercolor-like effect, which can be used as a decorative style.
Lora has good compatibility, but due to the small proportion of characters in the composition, it is prone to facial distortion. It is recommended to open ADetailer with a weight of 0.8. Trigger word: Hanama wine
Recommended tags:
Hanama wine
Write desired content here
<lora:Hanama wine-000018:0.8>, Negative prompt: EasyNegative, badhandsv5-neg, Subtitles, word, Steps: 32, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 2553923380, Size: 512x768, Model hash: e3edb8a26f, Model: ghostmix_v20Bakedvae, Denoising strength: 0.5, Clip skip: 2, ADetailer model: face_yolov8n.pt, ADetailer confidence: 0.3, ADetailer dilate erode: 4, ADetailer mask blur: 4, ADetailer denoising strength: 0.4, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer version: 23.11.0, Hires upscale: 2, Hires steps: 12, Hires upscaler: R-ESRGAN 4x+, Lora hashes: "Hanama wine-000018: 7d3593e99f39", TI hashes: "EasyNegative: 66a7279a88dd, badhandsv5-neg: aa7651be154c", Version: v1.6.0-2-g4afaaf8a
For the large model, I recommend using ghost or chosen_Chinese_styleV2 for the best results (recommended below). Other common anime-style models can also be used, preferably those with an anime-like style.
Thank you all for downloading and using it. If you like it, please follow me. My bilibili and pixiv IDs are both chosen鲷. You are welcome to follow my channels: https://space.bilibili.com/3867754?spm_id_from=333.1007.0.0, https://www.pixiv.net/users/40911695 Thank you.