Recommend:
vae-ft-mse-840000-ema
use highres fix to improve quality.
打了一个月王国之泪后重操旧业。
新版本算是对2.5d的整合，保留整体二次元画风的同时肢体上比前几个版本要好，脸型也要更多样化一点。
但光影和线条上就和2.5D更像一些，我也纠结了很久究竟该放在哪边，最终还是放在这了。于是干脆起个2.25d的版本名吧，至于孰好孰坏就见仁见智了。
多少有点背离初衷了，一点都不暗了。想暗的话还是自己加noiseoffset吧
After playing Tears of the Kingdom for a month, I resumed my old work. The new version is an integration of 2.5d, which retains the overall anime style while being better than the previous versions on the limbs, but the light and shadow and lines are more like 2.5D, so i simply call it 2.25d version.
起名废玩烂梗系列，事后想想起的不错。
首先暗图效果比较好，dark合适，自己融的时候差不多也是寿司那样什么都塞点，dark 寿司没毛病。
光效比较好，能生成比较暗的图，正常亮度的图片也有不错的光影表现。
融了群友私模和一大堆我也记不住组成的以前的自融模
能确定包含的是自己之前的五仁Five Nuts Mixed 五仁月饼Mix | Stable Diffusion Checkpoint | Civitai 和Unrule | Stable Diffusion Checkpoint | Civitai
以及群友的TmndMix ，绝对的杰作，推荐大家都去试试 https://civarchive.com/models/27259
darker版本里还加了noiseoffset的lora，生图要比brighter版本暗一些，光影表现上要强一点，但缺点就是正常生图也会比较暗。大家一般用brighter版再自己调用noiseoffset的lora也能达到类似效果。
Intro:
named from a dark soul's meme in Chinese community.
merged with my friend's pravite model and some other models merged by myself.
Including Five Nuts Mixed 五仁月饼Mix | Stable Diffusion Checkpoint | Civitai ,
Unrule | Stable Diffusion Checkpoint | Civitai
and my friend's TmndMix https://civarchive.com/models/27259 .it's definitely a masterpiece.
if you like dark sushi ,you should also try tmnd.
noiseoffset lora is merged into the darker version. So the imgs generated by the darker version will be darker ,and have better light and shadow performance. you can use brighter version and noiseoffset lora to get similar effect.
Description
FAQ
Comments (53)
VAE built-in?
no，there is no vae in this model. i always use this one: vae-ft-mse-840000-ema
@Aitasai Hi, could you please tell me where can I get the vae-ft-mse-840000-ema? Is it the https://civitai.com/models/6844/ema-pruned-vae ?
@pilockhulmes just download from hugging face. https://huggingface.co/stabilityai/sd-vae-ft-mse-original/resolve/main/vae-ft-mse-840000-ema-pruned.ckpt
@Aitasai Get it, thanks a lot!
老哥 我在reddit 宣傳你的模型
有個老外找你 他應該會直接找你
Getting really weird results with this one.
No matter the steps or CFG everything looks overcooked. Tried with VAE and with the standard one.
https://files.catbox.moe/1i6rxl.png
serene expression, mesmerizing eyes, portrait, solo, white hair, puffy sleeves, short hair, mole under mouth, long sleeves, juliet sleeves, cleavage cutout, black hairband, parted lips, black thighhighs, black gloves, cleavage, turtleneck, bangs, feather trim, feather-trimmed sleeves, beautiful eyes, black skirt, indoor setting, upper body shot, furrowed eyebrows, (partial table:1.4),smile, pov, date, restaurant,
i tried your prompt, it's normal. i have no idea where is wrong.
by the way, my webui get strange img information from the png you give. is it created by comfyui or something else?
@Aitasai I used ComfyUI, can you share your output too?
@muuuuuu I am getting the same overcooked results in ComfyUI but not in Automatic1111
SD 1.5 ?
im very lost, im not getting anywhere near the results on display or the ones people are showing on gallery, is there a vae or smthn im missing?
vae is vae-ft-mse-840000-ema. But accordding to the reviews, imgs generated without vae just have worse color , total quality is not far from mine.
Mybe highres fix is helpful.
为啥要分brighter和darker呢,又得多下一个啦
本来是想做个光影和暗图专精的，darker其实就比brighter多了个 noise offset。但发布之后有人反应正常图太暗了一点，群友也表示更喜欢不带noise offset的，于是就把不带的发布成了brighter。
只下brighter版再自己手动加noise offset的lora其实就足够了
@Aitasai 噢,我不太懂,请问noise offset的lora是哪种,怎么手动加呢
@inurDaydream 我用的是这个 https://civitai.com/models/13941/epinoiseoffset 就和正常lora一样用，权重可以加高点
手的模型很好很稳定，握持武器的效果也比较好，很好的模型
Can someone explain to me the difference between the pruned and normal versions?
请问各位大德，面部有时候不能画出清楚线条，调高STEP也没有明显改善，这可能是什么问题呢？
你分辨率多少?
@Aitasai 768*768，有试着改过640*480也没改善，一定要512*512然后再送hires吗？
@ThePureWhiteVampire 倒不是一定，但分辨率小人物再占比小点的话就是容易崩面部
可以试一下加入高清修复 迭代次数设置20左右
对，尝试加入高清修复Hires.fix
@L4wrence @gingerjay 有明显改善了！谢谢各位！
Can this be used commercially?
selling this model is not allowed.
images generated by this model are ok.
请教一下，在brighter版本中，我生成的图片饱和度明显偏低，即便是copy 别人的数据也是如此。是否存在额外的设置吗？
可能是vae吧，840000是比较鲜艳的
这个vae在哪里找
@Aitasai 感谢！挂载VAE后问题解决了。
@yangyangyangyang 找通用的VAE就行，作者用到的是vae-ft-mse-840000-ema，网上搜一下就有
@yangyangyangyang 在本網站（civitai），搜尋Ultra Sharp High Contrast Tutorial +vae&upscaler，或直接進這個網址：https://civitai.com/models/34192/ultra-sharp-high-contrast-tutorial-vaeandupscaler
下載後會得到一個壓縮檔，解壓縮以後就會解出vae-ft-mse-840000-ema了。
I keep having photorealistic eyes, not 2D anime eyes. I pair it with grapefruit vae (cannot find the mentioned vae) and 4xUltrasharp hires. Am I doing wrong settings? Can sb share urs?
我一直生出真的眼睛，不是動漫眼。我在用grapefruit vae （找不到簡介中的vae）和 4xUltrasharp 。我用錯了嗎？可以分享更好的設定嗎？
here is the vae i use. https://huggingface.co/stabilityai/sd-vae-ft-mse-original/resolve/main/vae-ft-mse-840000-ema-pruned.ckpt
But i dont think vae caused the problem. You can show your positive and negative prompt and I will have a try.
@Aitasai https://civitai.com/images/518992?period=AllTime&sort=Newest&modelId=24779&postId=150309 you can check my images here, you can see the whole face is out of style
@jeffreykwong1996155 i can't get your result. i guess the negative embeddings may be the promble.
you can try to generate imgs without them.
@lokmjuhb Turn off restore face option next to high res fix.
@cootie thanks a lot! problem solved!
I get overcooked results in ComfyUI but it seems to work fine in Automatic1111. I know RevAnimated had to be fixed to work with comfyUI as well.
this makes sense.
RevAnimated was merged into FiveNutsMix. And FiveNutsMix was merged into DarkSushi.
I follow exactlyI set it exactly according to the generated data, why is the generated character's face blue?
Do you have a plan for making the inpaint model?
All male character have a vagina.
curious about your prompt
@Aitasai best quality, anime quality picture, crisp vivids colors, detailed background, 1boy, very white skin, (((skinny shota demon boy))), teen male demon, small (black horn) on head, shirtless, wolf cut blue hair, malicious face expressions, ultra gore, gore, (pov:1), nsfw, red eyes, ((bathin in a pool of dark red blood)), blood on chest, dark palace, hell atmosphere, depth of field:0.3, 5 fingers,
negative : EasyNegative:05, low quality, text, pussy, girl, woman, vagina, adult, lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry, bad-hands-5, motion lines, censored, censor_bar, adult body, muscle, black skin, pubic hair, poorly drawn face, mutated, extra limb, deformed, breasts, abs, loli, mutated, amputed, (bad_prompt:0.8), fused fingers, boring_e621, 1girl,
啥群啊我也想加。。。
Details
Files
darkSushiMixMix_darkerPruned.safetensors
Mirrors
24779_darkSushiMixMix_darkerPruned.safetensors
darkSushiMixMix_darkerPruned.safetensors
darkSushiMixMix_darkerPruned.safetensors
darkSushiMixMix_darkerPruned.safetensors
darkSushiMixMix_darkerPruned.safetensors
darkSushiMixMix_darkerPruned.safetensors
darkSushiMixMix_darkerPruned.safetensors
darkSushiMixMix_darkerPruned.safetensors
DarkSushiMix-Darker.safetensors
darkSushiMixMix_darkerPruned.safetensors