NOTICE:LET ME KNOW before you put this model on commercial usage.
My twitter account:@eagelaxis :) Contact me if needed.
Discord Account:Eagelaxis#7818
Version Choosing Advice:V2f,V3,Coda and V3.5 are recommended,especially CODA for first-time users.
Hard to tell how many models used to merge.
Check the example images to recognize this model's art style
For more example images, just take a look at https://pixai.art
More attention on shades and backgrounds compared with former models(Andromeda-Mix | Stable Diffusion Checkpoint | Civitai)
Hands-fix is still waiting to be improved.
Highres-fix(upscaler) is strongly recommended(using the SwinIR_4x,R-ESRGAN 4x+anime6B by myself) in order to not make blurry images.
(Sorry for the misunderstanding caused by myself."Bad_prompt_v2:1.4" is a kind of embedding recognized as collection of normal negative prompts such as"lowres,bad anatomy".)
Recommend: Clip skip 2 Sampler:DPM++2M Karras Steps:20+
CFG scale:4-8 Vae:Pastel-Waifu-Diffusion.vae.pt(The vae used by Pastel-mix si just good enough)
Highres.fix:SwinIR_4x Hires steps:10+ Denoising strength:0.4+ Upscale by: 1.5+
Loras along with embeddings on hands-fix are strongly recommended.
V4.5 should be coming soon.
Looking forward to your reviews!
Description
A late update of WhaleFall series.
Mainly focused on more flat anime styles compared with CODA.
(Actually you can see whalefall as a branch of Version3,while Coda with Version3.5.)
Changed shades and lighting to make the whole background kinda like oil painting texture.
Maybe not suitable for some loras.
Would be nice if you could share either your generated art or some advice :)
(Lack of inspiration these days.)
Some little compliants (u don't have to translate this):
之前有不少国内的朋友想让我把模型传到诸如liblibai、吐司之类的国内网站上,说实话一开始是有这个想法的,但是后面一看到那堆霸王条款 想想还是算了。不要求你跟小破C站一样尊重作者,但是转载的话起码得问下我吧。也没必要拿什么“想分钱了”“训练素材违法”之类的绑架哥们儿,到现在为止我没通过上传模型挣过一分钱,我也只是个平平无奇的融模狱友而已。
国内的小伙伴可以在novelai贴吧里面找到我(虽然一万年没回过消息了)
FAQ
Comments (73)
为什么我使用whalefall版本会出错?就是无法载入这个模型,出现错误提示后回到之前的模型。秋叶整合包和星空大佬的A卡整合包都这样。
是这个版本还是V1
我记得之前检查过clip错误什么的。。我也用的是秋叶整合包
可能需要更新一下版本()
我在用1.3.1
本来想上webui看一下的
不过这个点宿舍也断电了()
很难帮上什么实际的忙
还是很抱歉Orz
@Eagelaxis 是V1,这个我版本我还没试过,我再试试看!
@Eagelaxis 确认,这个版本可以正常使用,没有出错
Looks great so far. Thanks.
The VAE is pt, but there is also a safetensor Version of it available.
Thanks for the tip :)
would find a safetensors version instead soon
good
怎么屏蔽掉xiodf3130这个人啊,发着不相关的模型图片恶心人
两种方法
1.在setting中的hidden user里面输这个人的名字
2.对着他发的图的右上角点,扩充栏里面应该有“hide content from this user”这个选项
3.可以顺便点一手举报
4.C站不出个举报封号的功能太无语了
每次都有这些恶心人的b
@Eagelaxis 我在其他模型看到他的时候就已经在设置中添加他的id,也对他的图右上角进行了点击屏蔽,结果今天在你模型这刷新了下,又看到了,屏蔽他没有用,举报也没用
@Gumu 我也遇到了一样的情况。不知为啥屏蔽不起作用了。插个眼等解决方案。
可以隐藏掉,我是看不到
同感.....这人一直在火的模型底下发NIJI图片霸屏,还很多人点赞,真的硬蹭...点踩也没用
@Eagelaxis 谢谢,亲测有用
@lewis181 你说的不是同一个人,不过都挺恶心人的
那就是一个SB,觉得自己很搞笑一样,就TM一狗der。
就是你把别人单位员工食堂的饭卡破解了,每天去专点贵菜,导致食堂半年损失100万?你是真的牛批,带着你们工作室所有人天天去吃?我说你以前三人一杯星巴克现在怎么这么大方了,半年吃一百万的白食你是真的牛批,别人都说比特币病毒是你搞的,我以前不信,现在看你还会改造饭卡,我现在有点怀疑了...
Beginner here, what's the difference between V4 and whalefall2 ?
I'd like to recommend V4 and Coda for every beginner on the Cetus-Mix series.
Compared with the other 2 series,V4 has the best adaptability for loras and acceptant painting styles.Coda could focus on impressive lighting and shades while WhaleFall focused on anime and oil painting styles.
There's also lots of shown images in the reviewing area. You can take them as usable advice on choosing different versions :)
@Eagelaxis Ok, thank you !
@Eagelaxis What versions do you recommend for advanced users?
Hi how can i find pastel waifu diff vae?
Strangely, a lot of the models have a VAE under "files" but none of them are pastel.
我是十分理解作者不上传国内平台的 并且支持,讲真的国内的环境不配,霸王条款没有一点对作者的尊重 更别说一堆无脑反ai小鬼和一堆网络垃圾人喜欢搅屎泼粪😅
小画尸太多破防很正常
@yaqiaoc233 那种不分青红皂白瞎打拳的就是很讨厌
我看了一下,那些在小红书叫的最凶的画的还没我12岁的侄子画的好
+1反正下载不需走梯子流量,就算走也付得起了
时代的洪流终会淹没这些“反ai小鬼”
The art and strokes here are amazing, and the hands are quite consistent compared to counterfeit or anythingv5
if you don't have a good gpu like me, you can try this on Google Colab (for free, no ban, no limited token). Also a useful guide here Youtube.
VAE link not working
Here you can download it
· VASVASVAS/vae at main
Can someone explain when i use Lora, it changes a little bit, its not like the checkpoint, and how to fix it.
Im a Begginer
when using Lora it's like using the same car but giving it a different color, window tint, spoiler etc.
when using a Checkpoint it's like completely changing the car.
when combining both you can have a different car with a spoiler, different window tint etc.
@greeney thank you :)
is this the VAE?
https://huggingface.co/samle/sd-webui-models/blob/23997884993f591d36bed7c645892d3d0828017f/pastel-waifu-diffusion.vae.pt
Definitely.
In fact, Pastel-Waifu-Diffusion.vae is exactly the same as kl-f8-anime2 (that is, thay have the same hash), which is originally released by hakurei, the author of WaifuDiffusion.
The original source can be found at his huggingface webpage of WaifuDiffusion-v1.4:
thanks!!!!
@emhhbw 请问这个放哪里呢,放vae那里没有显示诶,后缀是ckpt
i got the error when trying to use waterfall mix, anyone know how to fix?:
RuntimeError: "log_vml_cpu" not implemented for 'Half'go in your stable diffusion folder, locate either webui-user.bat (for window) or webui-user.sh (for mac) and remove "#" before "export COMMANDLINE_ARGS" and add this: --use-cpu all --precision full --no-half --skip-torch-cuda-test
The result should look like this: export COMMANDLINE_ARGS="--use-cpu all --precision full --no-half --skip-torch-cuda-test"
Note: This will make stable diffusion runs on CPU only with is very slow.
Nice to meet you. Would it be possible to use the images generated from this model in a YouTube video? They are very cute, and I would love to use them.
I have heard many great things about this model. My question is- what are the differences between versions? Why are v2, v3, coda recommended when there are more recent versions like v4 and whalefall?
edit: ok, I've seen your reply to a comment below. can you please add that to your description, as it looks like the description is out of date
Did you find out the differences?
So coda versions have a lot more porn in them (and they tend to be more realistic than 'anime'). Personally, I use the mainstream versions (v3, v4... etc) because it has a more 'general' understanding of things rather than a specific situation (as in porn) and also because they are stylistic as hell and I love them.
Have you done anything specific to increase the quality of hands?
The description seems to indicate that it is an intended fix for the future. So eventually.
https://tusiart.com/ https://www.liblib.art/
These two Chinese websites are stealing models from Civitai.
Bro, who cares. people here steals artists works anyway lmao 😂😂😂
@raxsauroboros no one here "steals" any more than everyone else outside of this place (this includes all artists that don't use AI); this is a pointless and redundant point to bring up
Also, part of the userbase here is already chinese or other east asian, lol
@kintilian45 Are we on the same website, lol? There's a ton of new lora every day that use someone else's art style or character without their consent. If that isn't stealing, then I don't see any problem with someone "stealing" model from Civitai for their websites.
what do you expect? it is not possible to access civitai without some extra effort there.
Its alright, i didnt get the best results, however I wasnt putting in much effort and the images were still good.
Hi, could I convert your model to huggingface/diffusers format?
请问我能不能把您的模型转换成抱脸diffusers格式?我想在diffusers管道运行模型
Hi! Excuse me, is there any settings for the checkpoint, such as for the prompts and negative prompts, and what sampling to use? i've read the description but i dont really understand it. and what is hands-fix? is it an embeddings or anything because i've searched it but found nothing, som im kinda confused. Hope theres someone answering this. Thank you in advance!
什么时候更新 sdxl 版
GOOD!!!
Cannot find recommended VAE (24/3/20)
you're missing the VAE this model is tuned for, which is why the colors are muted. You used to be able to get the VAE here, but it doesn't seem to be on the page anymore and the other links are broken. Trawling other comments, I found this; https://huggingface.co/samle/sd-webui-models/blob/23997884993f591d36bed7c645892d3d0828017f/pastel-waifu-diffusion.vae.pt
@hellozzj Thanks.
does anyone know where I can find .yaml file for the checkpoint (if thats needed)? i tried to look around, but couldnt find anything.
请问我可以模型做推文吗
Best of the Best
A pretty cute neat anime model. Moderately flexible, responsive and versatile. The only thing is that it is excessively fixed for female characters, and sometimes there are problems with hands, even with embeddings.
But overall it's very pretty. Thank you.
Which version(s) of CetusMix do you like best?
The performance in drawing male characters is disastrous.
Cetus-MIX is EXACT model than Meina-MIX??? They give me same image outputs..
it is normal this???
Do you know of any lora that would capture CetusMix quality of background for illustrious SDXL or Pony? I'v been scrolling for very long time on loras for background/landscape/foliage/forrest for Illustrious and nothing, everything is so washed, undetailed.
Details
Files
cetusMix_Whalefall2.safetensors
Mirrors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
Cetus-Mix.safetensors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
CetusMix Whalefall 2.safetensors
Cetus-Mix.safetensors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
cetusMix_Whalefall2.safetensors
SDNzaGtq8X-CetusMix_WhaleFall2.safetensors
cetusMix_Whalefall2.safetensors
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.




