PornMaster-Anime
客製化lora:我的Telegram ID:@iamddtla0620
我的微信ID:iamddtla
我的line ID:iamddtla
时间宝贵，请直接说明来意，例如:订制LORA.
Customized lora: My Telegram ID: @iamddtla0620
My WeChat ID: iamddtla
My Line ID: iamddtla
Time is precious, please state your intention directly, for example: customized LORA.
2023/08/23
FULL-V2-inpainting-可透过局部重绘生成，子宫脱垂(uterine prolapse)或肛门脱垂(anal prolapse、rectal prolapse)。
--
2023/08/12
V2增加子宫脱垂(uterine prolapse)和肛门脱垂(anal prolapse)、rectal prolapse
增加Gape60的合并，提示词可参考Gape60的培训数据。
1.培训lora请使用FP32或FP16版本
面部局部重绘可使用adetailer。
使用"sd webui cutoff"可避免颜色溢出污染其它元素。
高解析度放大可使用"ultimate SD upscale".
--
FP16版本适合电脑低阶配备。
FP32版本适合电脑中阶配备。
FULL版本适合电脑高阶配备。
"inpainting"版本的模型为局部重绘使用的模型。
2.如果您喜欢我的工作想请我喝咖啡：Paypal
3.推薦使用orangemix.vae
4.提示词可参考danbooru。
5.避免散景（景深、散景、模糊）
负面提示放入:(depth of field, bokeh, blurry:1.4)
--
2023/08/23
FULL-V2-inpainting-can be generated through partial repainting, uterine prolapse or anal prolapse, rectal prolapse.
--
2023/08/12
V2 Increased uterine prolapse and anal prolapse、rectal prolapse.
Added the combination of Gape60, the prompt words can refer to the training data of Gape60.
1.Please use the FP32 or FP16 version for training lora.
adetailer can be used for partial redrawing of the face.
Use "sd webui cutoff" to avoid color overflow contaminating other elements.
High resolution upscaling can be done using "ultimate SD upscale".
--
The FP16 version is suitable for low-end computer equipment.
The FP32 version is suitable for mid-level computer equipment.
The FULL version is suitable for high-end computer equipment.
The "inpainting" version of the model is used for partial repainting.
--
2. If you like my work and want to buy me coffee: Paypal
3.It is recommended to use orangemix.vae
4.For prompt words, please refer to danbooru.
5.How to avoid Bokeh:
Negative prompts to put in: (Depth of Field, Bokeh, blurry:1.4)
---------------------------------------------------------------------------------------------------------------------
|
Type
|
Checkpoint
|
Stats |
856
0
|
Published |
2024-10-18 14:04:41 |
Base Model |
SDXL 1.0
|
Usage Tips |
Clip Skip: 2
|
Trigger Words |
2024/10/18
XL-V1:
1.合并的模型皆为”培训(Checkpoint Trained)”的模型，并免合并混和”(CHECKPOINT MERGE)”的模型。没有任何lora的合并。
2.模型挑选高品质模型、高解析度的模型、或是具有大量培训数据的模型。
3.模型不包含vae.
--
1. The merged models are all "Checkpoint Trained" models, and no "Checkpoint Merge" models are merged. There is no Lora merging.
2. The model selection is high-quality models, high-resolution models, or models with a large amount of training data.
3. The model does not contain VAE.
--
以下为XL-V1所使用到的合并模型，生成图片时可参考这些模型中的提示词:
The following are the merged models used by XL-V1. You can refer to the prompt words in these models when generating pictures:
seaartFurryXL10_seaartFurryXL10
--
加速器:
Lora權重:0.8~1
12~14步。
取样方法(Sampler)=Euler a、
Schedule type: SGM Uniform
CFG:5~6
--
Accelerator:
lora weight:0.8~1
12~14steps
(Sampler)=Euler a、
Schedule type: SGM Uniform
CFG:5~6
