Sign In

AI Infinity Realistic - 更好的手 Better hands

496
3.4k
51
Verified:
SafeTensor
Type
Checkpoint Merge
Stats
3,404
Reviews
Published
Aug 3, 2023
Base Model
SD 1.5
Usage Tips
Clip Skip: 2
Hash
AutoV2
932B83662C
default creator card background decoration
AInfinity's Avatar
AInfinity

模型说明:

我一直希望融合出能够生成更好的手的模型。从我使用Stable Diffusion过去了5个月,我花费了大量的时间精力,熬了无数个夜去尝试研究,生成了几千张图片,融和模型几百次,几乎准备放弃。直到最近融合出我觉得还不错的模型,这个模型可以生成很好的手部,兼具真实感,如果你正在找手部不错并且真实的模型,建议你下载试用。

如果您觉得该模型不错,请记得评分并上传你生成的图片。谢谢!

交流Q群:730690448

English translation from Google Translate:

I've always wanted to merge models that would generate better hands. It has been 5 months since I used Stable Diffusion, I spent a lot of time and energy, stayed up countless nights to try and research, generated thousands of images, merged the model hundreds of times, and was almost ready to give up. Until recently, a model that I think is not bad has been merged. This model can generate very good hands with a sense of reality. If you are looking for a model with good hands and realism, I suggest you download and try it out.

If you think the model is good, please remember to rate it and post your generated images, Thanks!

反向提示词只需要下面这些,不要添加其他提示词 Negative prompts only need the following, don't add other prompts:

nsfw, ng_deepnegative_v1_75t,badhandv4, (worst quality:2), (low quality:2), (normal quality:2), lowres,watermark, monochrome

反向提示词必须使用下面两个嵌入,否则生成的图片手部会不好 Negative prompt must use the following two embeddings, otherwise the generated pictures will have bad hands:

ng_deepnegative_v1_75t badhandv4

请开启高清修复,没有推荐的参数设置,请自由测试,或者参考我的示例图片。

Please turn on high-definition repair,there is no recommended parameter setting, please test freely, or refer to my example images.

如果你想复现我的例图,请使用以下高清修复参数:

If you want to reproduce my example images, please use the following high-definition repair parameters:

图片尺寸 Size: 512x768

重绘幅度 Denoising strength: 0.5

放大倍率 Hires upscale: 2

高清修复采样次数 Hires steps: 30

放大算法 Hires upscaler: ESRGAN_4x

使用本模型请遵守法律法规。

Please abide by the laws and regulations when using this model.