Use e621 tags (no underscore), Artist tag very effective in YiffyMix.
GridList Species/Artist(v64) update!! & LoRAs (SDXL)/samples/wildcards
Recommend Artist tags (NoobXL) & comfyUI workflow.
Example Setting (Use recommended sampler to get correct quality)
Setting SDXL (SDXL-lightning & NoobXL)
Steps = 12~24
Sampler = "DDPM", "Euler A SGMUniform", "Euler SGMUniform"
CFG scale = 3~4
Negative embeddings SDXL = ac-neg1, ac-neg2 (You don't really need this)
Postive LoRA SDXL = SeaArt Quality Tags LoRA (You don't really need this)
Stop at CLIP layers = 2
Setting SD 1.5 + vpred
Steps = 30~40
Sampler = "DPM++ 2M Karras", "DPM++ SDE Karras", "DDIM", "UniPC"
CFG scale = 6~8
Negative embeddings = deformity_v6, bwu [SD-WebUI\embeddings]
Stop at CLIP layers = 1
SD VAE = kl-f8-anime2, Furception [SD-WebUI\models\VAE]
Hires. fix
Hires steps = Steps * Denoising strength
Denoising strength = 0.25
Hires upscaler = 4x-UltraMix_Smooth [SD-WebUI\models\ESRGAN]
ControlNet
ControlNet = softedge_hed, control_v11p_sd15_softedge
ControlNet SDXL = softedge_hed, sdxlsoftedge-dexined, noobai-xl-controlnet
ControlNet Weight = 0.35~0.5
ControlNet Pixel Perfect = true
Wan 2.2 ComfyUI Animeted [workflow]
model = wan2.2_ti2v_5B_fp16
vae = wan2.2_vae
lora (turbo) = lora_wan2.2_ti2v_5B_turbo_lora_rank_64_fp16
lora (furry-nsfw) = wan2.2-5B_furry-nsfw-v2.0-e83
lora (livewallpaper) = wan2.2-5B_livewallpaper-720p
sampler = steps:4, cfg:1, sampler:"eular_ancestral", schedule:"sgm_uniform"
SD WebUI
LoRA Training SDXL
imgs count = 15~50
total steps = epoch * imgs count * folder loop = 3000~4500
network_dim = 64
network_alpha = 128 (SDXL) \ 16 (Noob)
learning_rate = 0.0002~0.0005
unet_lr = 0.0001 #learning_rate/2
text_encoder_lr = 0.00005 #learning_rate/4
lr_scheduler = "cosine_with_restarts"
mixed_precision = "bf16"
optimizer_type = "Adafactor"
optimizer_args = [ "scale_parameter=False", "relative_step=False", "warmup_init=False", ]
YiffyMix v4x V-pred Setting
# YiffyMix v4x A1111 SD-WebUI or Forge V-pred Setting
1.Download config file ".yaml" paste next to the model.
2.Raname ".yaml" same as model name. ( check the ".yaml" has [parameterization: "v"] line )
3.Restart SD-WebUI ( if config fails to load, the model just generate noise )
# YiffyMix v4x ComfyUI or StableSwarmUI V-pred Setting
Add node "ModelSamplingDiscrete", Sampling = "v_prediction"
# ComfyUI V-pred Setting (old version)
1.Copy config ".yaml" to [ComfyUI\models\configs] and refresh ComfyUI
2."Load Checkpoint (With Config)" [Right Click\AddNode\advanced\model_merging]
workflow:[Load Checkpoint (With Config)]-[KSampler]-[VAE Decode]-[Save Image]
# v-pred mode troubleshoot
If you use new version webui-forge and failed detect v-pred model.
Try update webui-forge (click update.bat).
When loading v-pred model, you will see this line in cmd console:
left over keys: dict_keys(['v_pred'])
# v4.x some time will get little fried issue image in this case
Use "Dynamic Prompts with __wildcards__ prompts" and "batch size > 1"
Just generate again with same prompts with parament, it will return normal.
Version Info
v1.x [2D,512~768] old model, unstable, low resolution
v2.x [2D,512~896] e621 dataset + 30~40% SD5 dataset
v3.0 [2D,3D,512~1088] large dataset then v2.x version, can do 2D, 3D
v3.1 [2D,3D,Real,512~1088]
more realistic, but loss some concept (e621 tag count lower 1000)
v3.2 [2D,3D,Real,512~1088]
unstable version, use SNR version FluffyRock, more noise detail
v3.3 [2D,3D,Real,512~1088]
stable version, more detail, more sensitive to prompts
v3.4 [2D,3D,Real,512~1088] ※include Fluffy Rock Quality Tags-LoRA
stable version, more detail, clear result, reduce some noise (like bush,pattern)
v3.5 [2D,3D,Real,512~1088]
unstable version, contrast & detailed enhance
v3.6 [2D,3D,Real,512~1088]
stable version, more sensitive to e621 tags
v3.7 [2D,3D,Real,512~1088]
stable version, reduce noise overactivity, fix blurry eyes and finger error, fix failed background depth
v4.0 [2D,3D,Real,512~1088]
v-pred version, remix from EasyFluff
accuracy anatomy and style with fewer prompt,
little dim blue average color with smooth noise, weakly negative prompt issue
v4.1 [2D,3D,Real,512~1088]
v-pred version, color and low contrast fix, reduced realistic noise.
v4.2 [2D,3D,Real,512~1088]
v-pred version, more contrast and detail, yelow issue.
if you feel too fried use Rescale CFG =0.35
v4.3 [2D,3D,Real,512~1088]
v-pred version, clear artist style, fix most yellow/brown issue
the version doesn't need CFG Rescale
v4.4 [2D,3D,Real,512~1088]
v-pred version, no more yellow/brown issue.
the version doesn't need CFG Rescale
v5.0 [2D,3D,Real,896~1536]
SDXL-lightning version, base on Compassmix XL.
v5.1 [2D,3D,Real,896~1536]
SDXL-lightning version, more e621 dataset, lower human face, better NSFW stuff.
v5.2 [2D,3D,Real,896~1536]
SDXL-lightning version, increase average quality. Little unstable than v51 but more creative.
v6.0 [2D,3D,res:896~1536]
More character better sex pose, limited and uncontrollable style (mostlly anime style).
v6.1 [2D,3D,Real,896~1536]
More realistic detail, reduce anime style, fix flat and boring background.
v6.1a-RE [2D,3D,Real,896~1536]
Same as v61 but adjust average style to simi-realistic.
v6.2 [2D,3D,Real,896~1536]
more effective prompt, keep style also with good realistic detail, saturation down little bit.
v6.3 [2D,3D,Real,896~1536]
Detail(noise) level between v62 and v61, upgrade character/artist accuracy.
v6.4 [2D,3D,Real,896~1536]
Upgrade light quality and average detail.
Note about NoobXL to create stable furry (v6x):
use "furry" tag in propmt can let AI create furry not human.
use "no human" tag in propmt can avoid NoobXL model add human and reduce anthro affect (more original style).
use "anime style" in negative propmt can reduce classic booru style and more realistic.
Some SD prompt trick:
Combine two character:
characterA \(characterB\)
Avoid tag bleeding:
(chain:0)-link fence
(cowboy:0) shot
high (collar:0)
Multi-tag combine, enhance and reduce token use:
from side + side view = from side view
crossed legs + legs up = crossed legs up
Basic Style
Negative Prompt SD1.5
unusual anatomy, mutilated, malformed, watermark,
amputee, mosaic censorship, sketch, monochrome
Negative Prompt SDXL
malformed, worst quality, bad quality, signature, text, url
3D Artwork Style
Prompt v5x: blender \(software\), ray tracing, 3d, unreal engine
Photorealistic Style SDXL
Prompt v5x:
by Mandy Disher, by Wim Wenders, by Robert Rauschenberg, [by Fossa666::0.65],
ultra realistic, photorealism, photograph
Negative prompt v5x: sketch, manga, vector, line art, toony
Prompt v6x: film photography, photorealistic, film grain
Negative prompt v6x: anime style, vibrant, pastel
Photorealistic Style SD1.5
Prompt sd15: [:photorealistic, analog style, realistic, photorealism:0.5]
Negative prompt sd15: [:bwu:0.5]
Recommend Refiner: IndigoFurryMix v110-Realistic Switch at 0.5
Description
Recipes v63-Noob:
partA = ZoinksNoob + (Noob-Eps v11 - ChenkinNoobXL v01) * addDiff:0.55
partB = partA + (YiffyMix v62 - partA) * trainDiff:0.75
YiffyMix v63-Noob = partB + LoRA:[SPO-SDXL_4k-p_10ep_LoRA_webui] * 0.35
FAQ
Comments (23)
Art style in general and prompt adherence are very good. However, there are problems with generating hands correctly
one hand is generated with 6 fingers and other with 5, is very rare generate same number.
@Duke678 i don't have same issue
@jokervek315 Now that I use styles, it hardly ever happens to me. The only problem I have is with earrings on some heavily influenced characters, such as Android 21. No matter what negative or positive prompt I try to use, it keeps generating them.
@Duke678 well, in this case you can use ControlNet lineart and in preview manually remove ear ring and regenerate it in img2img tab. Or just use photoshop in this case
what image sizes are supported with HighRes fix using v63 & 4x-UltraMix_Smooth?
Depends on HighRes denoising value (SDXL have 768px~1536px stable range):
Denoising in 0.15~0.25, image can upsacler to 2048~4096.
Denoising in 0.3~0.55, maximum upsacler size is 1536, higher resolutions will cause duplicate.
realistic style try 0.15~0.25, anime & flat color style try 0.25~0.35.
@chilon249 thank you. This helps
Insane
I'm trying out v63 and It's an incredible model! Amazing work!
How do I remove earrings from characters who are strongly associated with them? I'm trying to make an Android 21, but 90% of the time it keeps creating earrings or bracelets. I'm trying everything, both positive and negative prompts. no ears accessories, ((no-earrings)), ((no-bracelets)), ((no-jewelry))
Character tag include all about this character feature.
This is why noobXL can easily create mult-characters, but you can't easily removed feature.
A1111/Forge script prompt can trick to avoid it (other way is inpainting).
prompt:
[majin (android:0) 21::0.06],
[elf:(majin (android:0) 21):0.6],
white hair, long big hair, pink body, black sclera
negative prompt:
earrings, ear ring, collar
first 6% steps draw a simple character shape (keep android 21 unique hair style stable)
last 60% steps finished "android 21" detail.
0~60% use "elf" create a similar character templet (need additional description)
### use "(android:0)" to avoid "android" tag affect
You are using negative prompt for "no ears accessories, ((no-earrings)), ((no-bracelets)), ((no-jewelry))"
Meaning that you tell the model that you DON'T want images with no ears accessories, ((no-earrings)), ((no-bracelets)), ((no-jewelry))
Just put "earrings" in negative
There's a extension Negpip, use prompt like (xxx:-1), this might work in your case.
@chilon249 Thanks for the advice, I'll see what I can do. Unfortunately, as a Fooocus user, there's nothing I can do about negpip. I haven't been able to use A1111 with my 2060 since the first Pony model came out. And Comfy takes too long to generate and requires advanced knowledge even for simple inpainting and upscaling. I eagerly await the new version.
Little update: https://civitai.com/images/114277045
I don't understand why, but now it creates the character without bracelets or earrings of any kind. Perhaps the positive no ears, ((no earrings:1.2)), have done their job without HIRES.
New (v63) Recommand artist grid, including 2837 artists from booru and e621, goto find some new style.
I have no idea why, but it seems this model is not able to generate certain characters (unless i explicitly specify what they are), even tho the list says it does. For example, tags like shenzi \(the lion king\) is completely unknown by the model, but other recent furry models can generate her, no problem. Not even if i specify disney does it even come close to looking like her (only mimics the style, not the character). Only when i add the hyena tag does it actually generate the character.
That's mean this character need auxiliary prompts, most unpopular tag has this problem.
"shenzi" has 1k tags in e621, but only 14 tags in danbooru.
In base noobXL model, booru tag is more effective than the e621 tag.
... not to derail the comment, but i see you have a bot problem now. just got a notification that ive gotten 10 reactions in a single day for no reason. and i see the same bots have spammed all the comments in here as well with the same thing.
Great model! I do have one confusion over how it works. I would like to generate artist OC characters that generally have 200+ tags on e621 and are on the character list. I simply write the characters name as it is in the tag and it isn't working even with higher weights. If there are still few tags on danbooru vs on e621, do I need a LoRA or is there a way to have yiffymix recognize them like it does with much more popular copyrighted characters? These are furry characters so they have little or no tags on Danbooru at all.
EDIT: I found the OC characters list, and some artists (like theycallhimcake or Jinu) do not have OCs recognized like others, despite having more tags than many of the ones that are. These characters are in the total tag list however. I'm not sure exactly why this is the case, but I hope to see some more recognized characters in the future because this is a really great model you made.
So, why is it so unstable? Or it is me who did something wrong? The same prompt everytime but the style is always changing. It could be reastic, then semi-realistic and then cartoony. The same prompt and settings.
Amazing model!
I have a question...
Is there any way to optimize the e621/booru tags for better results?
Most tags work well, but there are others that the model does not recognize well...