Banana Splitz XXL - V1.1
(Banana XXL + FurrySpec | Simple Furry Mix + ReKemonoV9 + Furry Dreams + a few LoRAs )
Minor change in prompt adherence
style update - slightly more painterly and lite 2.5D aesthetic.
Prompt adherence slightly LESS sensitive (allows for more advanced prompting)
Camera angles and poses are about the same.
Banana Splitz XXL - V1.0
(Banana V9 + FurrySpec | Simple Furry Mix + DeepNavyV3+ ReKemonoV7)
Minor update to coloring + lighting
Major style update - smoother, cleaner, less painterly more solid.
Prompt adherence slightly more sensitive.
Offers a high quality 2D - 2.5D look with expressive posing and faces.
Camera angles are more creative when propmting for dynamic angles.
Banana Splitz - V9.0
(Banana V8 + V6)+DeepNavyV3+Luno V6[Renamed ReKemono])
Major update to coloring, lighting, poses, camera angles.
LoRa compatibility is still great
Minor clip difference from 8.0 (adherence is roughly the same)
Banana Splitz - V8.0
(Banana V6+V7a+V3)+DeepNavyV3+Hyphoria+More)
Major clip update. Expanded clip with Hyphoria
which is a model that cleverly merged 10 different models clips
Rule 34 Data set added as usual.
Prompt adherence is crazy strong.
LoRA compatability is also very high.
Backgrounds are wildly beautiful.
Can do humans/Anime with no problem.
Extremely large spectrum of styles can be prompted.
You can combine artist tags easier like in the SD1.4 days
_______________________________________________________________________
Banana Splitz - V7.0 Side B
This is basically V7 but with a more westernized toon style
(Banana V6+V7+V5(ReMerge)+DeepNavyV3+Oops All Toons)
LoRA Used:
KSM - Gives the model a more bit of western toon appearance. Helps with artist tags
Dirtysoda - Another LoRA for helping westernize this model more
Rule 34 Dataset LoRA - Smallest ammount used for activation. This is for better LoRA compatibility and higher usage causes noise. To finetune the R34 Lora within this model, please use actual Rule34 Lora. (READ INSTURCTIONS)
https://civarchive.com/models/1191424/pornmaster-rule-34-prompts-illustrious-and-noob
__________________________________________________________________________________________________
Banana Splitz - V7.0 Side A
(Banana Mix+MolKun-Mashup+Luno Mix)
Banana Splitz XL 6.0 + Banana Splitz 5.0
Other models in recipe
Hyphoria
Nlxl:
Nlxl v0.9 is a checkpoint that produces more stable and detailed results through enhanced Model Merged checkpoints created through model merging, combining the natural image generation capabilities of Illustrious, the extensive dataset of Noob, and the unique strengths of different models depending on the version
_______________________________________________________________________
Merged LoRA info, so you know how much extra of the same lora to use, or if it lora's wont be compatable. Also lets you know whathey do.
KSM - Gives the model a more bit of western toon appearance. Helps with artist tags
Rule 34 Dataset (LoRA)
The developer speaks Chinese only and doesnt really reply correctly to anyone.
but this LoRA adds a large amount of tags that do not exist on danbooru or e621
This creates an effect that allow an awesome solid mix of tags/natural langauge.
It also makes NSFW content so much better. I think it also re-enforces the knowledge of already know tags for pose, expression and anatomy wise.
bruhhhh Its such a huge an potent lora. Its 3.4gb and is merged is sometimes merged in SUB 0.1 weights. like 0.08, 0.06. Its gonna be doing something powerful
__________________________________________________________________________________________________
Banana Splitz - V6.0
(Banana+Oat+Sudachi Mix)
-Banana Splitz XL 2.0 & 4.0 mix
-Sudachi XL
-Plant Milk - OAT
-Xavier-Shift (Trunk of Furry Junk)
-Mol_Kuen DeepNavy
_______________________________________________________________________
Banana Splitz - V6.0 BETA
(Banana-Hassaku+Walnut Mix)
***NOTE V3.0 AND UP IS FOR EULER A SAMPLER ONLY**
New base merge recipe:
-Banana Splitz XL 3.0 & 5.0 mix
-Sudachi XL
-Plant Milk - Walnut
-Hassaku 3.0
-Mol_Kuen DeepNavy
-Hyphoria
-Rule34 Prompt Lora updated to V2.0
__________________________________________________________________________________________________
Banana Splitz - V5.0 [ READ ME! ]
(Banana-Hemp+Walnut Mix)
***NOTE V3.0 AND UP IS FOR EULER A SAMPLER ONLY***
Try using a CFG Rescale plugin for ComfyUI or A1111 if want to attempt another sampler
New base merge recipe:
-Banana Splitz XL 2.0 & 4.0 mix
-Sudachi XL
-Plant Milk - Hemp ll
-Mol_Kuen DeepNavy
-Rule34 Prompt Lora updated to V2.0
TEXT ENCODER MODIFICATION:
This version will properly acknowledge prompting from those who write longer and wildly detailed prompts. BUT! it will also help/support simple prompts into creative works of art!
Great for beginners and fantastic for pros to push their creative limits!
Positives for the Pros!
highly specific prompting will now have a much higher chance to affect small details.
I suggest even putting weather or time of day (morning, golden hour, night).
Even objects in background like a swing on a tree or a bookshelf indoors.
Style/Artist blending can have surgical precision.
detailed negative prompting can help shape result moderately better too
Positives for everyone!
Simple prompts will have a boost too:
Short or vague prompting will give creative wildcard like behavior.
Can give good solid and creative prompts with simplistic prompting.
Good for beginners to help build upon their prompts making them more verbose and gaining the skill to write more detailed prompts.
Great for beginners to see the changes by adding, changing how they prompt.
Remember! Everyone prompts differently and this model can certainly help you tone up on your own prompting skills!
_______________________________________________________________________
Banana Splitz - V4.0
(Banana-Walnut Version)
New base merge recipe:
-Banana Splitz XL 2.0
-Plant Milk - Walnut
-Sudachi XL
-Rule 34 Dataset V1.0
-Style is mostly the same, but higher quality.
-More detailed backgrounds
-Prompt adherence is average. Needs work.
-Extremely LoRA compatible
-Artist styles absolutely SHINE on here.
-Can one-shot generate a good image 3/4th of the time
_______________________________________________________________________
Banana Splitz - V3.0
(Banana Flax Daquiri Mix)
A 'high-quality' Western-Kemono hybrid with crisp linework, soft but dimensional shading, and a balanced 2.5D aesthetic.
Better and easier prompt adherence
Clean, stylish, thick lines with glossy 2.5D textures
Contains Rule 34 Dataset.
Great eyes! Great Fingers, Awesome poses.
Check out this amazing finetune of Banana Splitz by LS582.
Fifth Mix - Banana Daquiri does exactly what its told, clean, crisp style. This is what ChromaXL should have been IMO.
If you like this model try it out!
_______________________________________________________________________
Banana Splitz - v2.0
(Banana Daquiri Mix)Updates:
-Moderate Style Change
Model now is has moderately higher saturation, slightly lower contrast, more colorful.
Less flat and washed out (if you like that look from V1 I may consider making a V1.X, Plz Comment)
When no style is specified you will get random styles beetween 2.5D cartoon and 2.5D semi-photorealstic
-Massively Improved prompt adherence.
Has improved natural language understanding(needs to be mixed still with tags)
Prompt this model like a combo of Fifth Mix - Banana Daiquiri and NoobAI
-Refined quality and body comp.
Much less random zoom ins and cropping when using "close up" tags
You can get good eyes on raw gens 60% of the time with full body/multi character
You can get good eyes on raw gens 95% of the time with portrait/close up/solo
Different male and female body types
better knowledge of animal dicks and pussies(specifically canine)
-Pure NoobAI cyberfix, train difference injection
Use e621 tags for artist associated with NoobAI
vast knowledge of furry artists and can nail the style very closely 80% of the time. (check last 3 images in example images)
Natural Langauge is better understood. but not perfected.
-Making Lora/Lyrcoris models use the following.
NoobAI EPS1.1
Illustrious V0.1
WAI-NSFW-Illustrious V11 (Smooth looking with mild risk of sucking)
Fifth Mix Or Banana Splitz (Highly creative / Lower risk of sucking, but makes it useable for these models only)
__________________________________________________________________________________________________
Banana Splitz - v1.0
(The Original Banana Splitz Mix)
ORIGINAL VERSION 1 MERGE DETAILS:
Models used:
Fifth Mix
FurryToonMix
MolKun Mix
NoobAI CyberFix EPS V1.1
Custom Oops! All Furries
(Re-Merge/Pruned All trace of models that use ChromaXL Or ChromaMorgy
Custom Indigo NoobAI Mix(I might release this)
Custom Nova Furry Mix
(Custom Add Diff/Train Diff Mix of 3 versions - Might release this too)
Hassaku V2.1
Interested In any of the custom models used?
Let me know!
Description
The REAL 8.0 is here.
A proper contiuation from V7 Side A
Now with extended CLIP using Hyphoria(which contains 10 model clips merged)
R34 dataset also included to extend tag knowledge.
Promt adherence is so strong. REALLY strong.
Visual aesthetic is a blend of.....
MolKun DeepBlue V3
Bsplitz V7a
Bsplitz V6
Hyphoria(just a tad)
Luno V3
Xavier-Shift V1
FAQ
Comments (18)
JUST TO PREVENT CONFUSION. I HIDDEN COMMENTS THAT REFER TO V7 SIDE B AS V8. THERE IS NOTHING WRONG WITH THEIR OPINIONS.
ITS STRICTLY AVOID CONFUSION TO NEW USERS.
I HATE PROMPTING I HATE PROMPTING I HATE PROMPTING
(also, all my gens come out in this weird 2d esque style that I also hate, I'm trying to strive towards the images in the examples above the gallery & comments)
Anyways, this is the best version yet!
Just a heads up I use the OG A1111.
Forge, Reforge, give different results
ComfyUI, forget it. Youre gonna need 90 nodes to emulate this probably.
(comfy is best for big models like FLUX and WAN, imo)
Everything I make is remixable, if you see art styles you like, feel free to use it!
@LatteLeopard This is why I need to stop using a chromebook.
@GayPurpleSkunk Could you perhaps, give me some aid upon prompting, specifically quality tags in both positive and negative prompting, possibly we could chat about this, since just commenting is quite odd (or not) [God this sounds so corny] thanks mate
@TheGarandMann Of course, feel free to DM me! ^^
I am no longer jolly.
It seems great at generating single characters, but once you start introducing more than 2 characters in the same image, prompting really takes a nosedive and seems to generate nonsense then. Like if i use 3girls or higher, it seems to generate everything from 3 girls to 8 girls, even tho i didn't prompt it. And all of them seems to be really randomly selected based on what you've prompted otherwise and not at al cohesive, like girls growing dicks and men growing tits randomly etc.
EDIT: from what i've been able to test, BREAK seems to be the only thing that helps in this situation. Not perfectly, but it helps. Haven't been forced to use this since the SD1.5 days, but i guess it will have to do...
"3girls" or higher technically are NOT proper danbooru or e621 tag. The correct way to do this is to specify, as "female", "trio", or "female, group sex."
If you look in the gallery below, I have users who have done groups of 3+ with no issues. Even in my cover images I show off a group of 4
Most models struggle with more than one character, but I know this model can definitively do groups of three while maintaining facial details
@LatteLeopard Yes, but they all are basically clones. To make 3 different girls (which is what i was looking for, not only the number of them), BREAK was needed in this case, one for each. This was not needed on other models (not counting yours), not even vpred ones, as long as you specified the number of characters and the name of them, it would be consistent between each generation.
Just pointing out that even tho prompt adhesion is really good, it comes at the expense of prompt cohesion and quality. I really struggled to come up with a different workflow that has essentially not really changed since the SDXL days, since it insisted in creating body horror otherwise. And not even then is it particularly cohesive, only like 1 in 10 generations are good enough to save, while "normal" illustrious/noobai models have 1 in 4-5 generations.
In fact, the cover image you speak of, you prompted threesome and yet it generated a foursome... so... That's kinda my point.
@richerich I purposely chose the Gen the ended up with four characters for the cover art.
Also they're aren't all clones? Sometimes you see remixes of the cover art pictures, but that's not the case here.
This is really a male focused model to begin with so, I haven't made it more compatible with females until v4.0.
I'm sorry it doesn't prompt like identically to other models you used. If you aren't familiar with the models used in the mix, it might be confusing.
All I can really say is, I'll make sure to test groups more thoroughly. But I dont really think it's a problem at the moment. Especially compared to other models that can barely do two characters.
Gotta say, this checkpoint has been able to produce some of the most insane, extreme detail, perfect images right out of the box. The fidelity and complexity it's capable of is insane. I'm pretty good at figuring out checkpoints and models, but man this makes it easy. Inpainting with it is absolutely wild, too. For a checkpoint with such a large depth, it's impressive. Photorealistic, 3DCG, traditional-realistic, cartoon, it slams them out. I will say it struggles ever so slightly with cartoon, images seem a bit less consistent and high quality.
It has faults, sure. It has a raging tendency to default to POV images the second you add a male character, which seems impervious to negative prompting somehow. Always have to get around that with extra reinforcement in the prompt as opposed to negative prompting. It's also very, very heavily geared towards NSFW, which isn't a downside, but after seeing the kinds of images it could pump out, I didn't want to waste that potential on only gooner slop nonsense, and decided to generate some SFW wallpapers and such. It was a minor struggle, hardly an issue though.
I swear there was one other problem I've encountered using this, but I can't remember it now. Must not have been that big of a problem. If I remember it, I'll edit/add a new comment.
edit: I remembered, it tends to force the standard "two-tone" furry body pattern pretty hard
Anyways, top tier model.
Awesome review, Pretty much nails it.
Unfortunately with merging it becomes a balancing act.
and we both know the #1 usage is gonna be NSFW.
so I crammed in all pieces from all the most explicit models + loras.
Even to the point of getting stuff from Chinese sites that function like this one.
Banana splitz was the first furry model to utilize the rule 34 dataset (which was made into a lora to use a patch, rather than a model).
so solo images are more likely to POV in the sense the character will look at the viewer. if youre getting like random body parts, like a cock at the bottom of the screen to represent the viewer. Thats an extremely unique case, Ive seen it on other models. But I've yet to come across it on my own. But then again, my generation usage is below the average user.
The two-tone coloring. Oh yeah. Its gonna want to follow the pattern of the animal your chosing. Since a black wolf is all black, it will actually create a solo color. but if you try grey, it knows grey wolves have two toned fur. so any color you pick, it will give em a matching secondary color, usually a lighter or darker version of what you chose.
Incredible checkpoint! However, sometimes I have some issues with excess visual noise while generating, especially when inpainting at high resolutions. Does anyone know what could be causing these issues? Some other models like WAI don't have such an issue, but I much prefer this checkpoint and I would love to use it as my main one without having to switch back and forth for inpainting details :D
Maybe I'm not prompting properly? There must be something I'm not doing right. Would love some advice!
im not sure, I use A1111 for its merging tools, as they arent ported to work with any of the Forge/reforge stuff. So I ususally end up testing and making the images on a1111. I unfortunately dont know why inpainting would cause artifacting? But I can confirm it works on A1111 when I've tried.
The only real thing I can think of that would REALLY set my model apart from WAI on a technically aspect would be that WAI uses just Illustrious 0.1 and adds small patches of data with Loras. But I use Illustrious 2.0 + NoobAI merged together as the core, most newer model merges follow this trend.
@LatteLeopard I do use Forge (switched to it a little while ago since A1111 is quite outdated now), so maybe that could be it? 🤔 odd. Thank you for your reply, though! Can't wait to try V9 :D