你问我配方？我不到啊。我都融了一个多月了。
这个模型不能跑NSFW！这个模型不能跑NSFW！这个模型不能跑NSFW！
重要的事情说三遍！
这个模型主打的就是一个干净画面，适合头像和矢量图。
也可以融其它模型，我推荐kenshi和Pasel系列模型。
当然，对于中文是否有支持，取决于你是否安装了某些扩展，至于是什么拓展，我也不知道。
建议CFG区间在9-13.5之间，你要是不听建议那也没办法。
效果图的tag没写胸部大小描述词，这个模型也没有特化。
我个人非常讨厌拿私人模型进行商业用途，因为模型的底图的溯源版权难以追踪，这对于画师来说是不公平的事情，所以，我会在这里呼吁：”当你在使用这个模型时，生成的内容仅仅作为公共用途，并且永远不要盈利“
至于你下载之后是否会进行其它用途，（比如收费替人跑图，或是贩卖模型（当然我非常反感贩卖模型，请最好不要这样做）），那便是你自己的行为，我无权出警，我觉得良好的社区需要每一个人自行约束来维护。
如果你要将这个模型与其它作品混合，再进行其余的收费，我只会善意地提醒你，当然你也可以把你的作品发到评论区。
This model cannot run NSFW! This model cannot run NSFW! This model cannot run NSFW!
Important things are to be repeated for 3 times!
This model focuses on a clean picture, suitable for head portrait and vector image.
Other models can also be integrated. I recommend Kenshi and Pasel series models.
Of course, whether you have support for Chinese depends on whether you have installed some extensions. I don't know what extensions are.
It is recommended that the CFG range be between 9-13.5. If you don't listen to the suggestions, there is no way.
The tag of the sample picture does not contain the chest size description, and the model is not specialized.
Personally, I hate to use private models for commercial purposes, because it is not fair for painters to track the traceable copyright of the base map of the model. Therefore, I will appeal here: "When you use this model, the generated content is only for public use, and never for profit“
As for whether you will use it for other purposes after downloading it (such as charging for people to run maps, or selling models (of course, I hate selling models very much, please don't do so)), that is your own behavior. I have no right to call the police. A good community needs to be maintained by everyone.
If you want to mix this model with others and then charge for the rest, I will only kindly remind you. Of course, you can also send your works to the comment area.
Description
FAQ
Comments (9)
都说话啊 怎么都不说话 是不想说话吗？
Really a good and different model !
Can you add a smaller 2gb pruned fp16 no-vae safetensor version ?
If this can be of any help:
- https://github.com/Akegarasu/sd-webui-model-converter
"convert to precisions: fp32, fp16, bf16 - pruning model: no-ema, ema-only - checkpoint ext convert: ckpt, safetensors"
- https://github.com/arenatemp/stable-diffusion-webui-model-toolkit
"Cleaning/pruning models - Converting to/from safetensors - Extracting/replacing model components"
Hello, my friend, I forgot the mixing formula at that time. If necessary, you can try to use the "non-interpolation" method for model fusion, and choose float16. If that doesn't work, I'm afraid I can't help you,
By the way, I use the model fusion provided by webui instead of adjusting the difference merge. I usually use this method to fuse models, and rarely use fp16. I'm afraid it is difficult to help you.
However, if I make progress in the future, I will update this website and mark the fp16 version as soon as possible.
Oh, it seems that the method I just mentioned is the version of fp16. Let me see how this model reacts before uploading the version of fp16.
Ok, do what you can. Thanks anyway.
Are you using the A1111 webgui ?
@ritcher1 nope,i use the Chinese version,Because the version of A1111 is updated a little frequently, I can't adapt to it, and there are some hidden bugs. A friend of mine has been using it all the time. He thinks that every run is the same as a medieval sailor, either sinking or returning with a full load. The Chinese version is currently 1.7 and is also being updated.
@HarukaAkagi So, any chance to see the smaller fp16 version ? This model is really good.
@ritcher1 maybe next week
@HarukaAkagi And pruned .ckpt version too, plz
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.