SURPRISE ;)
Merry Christmas, everyone.
As a little Christmas gift for you all, I've been working on a new version for the SD 1.5 version of Juggernaut in my spare time for the past few weeks.
I took a small portion of the JuggernautXL dataset and trained it on SD 1.5 (Base Juggernaut Final), while also updating the epiCRealism model in the merge. However, the RPG and Diva models that were added in the 'Aftermath' version have been removed.
Images were created with the following settings:
512*768
Sampler: DPM++ 2M Karras
Steps: 35
CFG: 7
HiRes Fix:
Sampler: 4xNMKD Siax 200k
Steps: 15
Denoise: 0.5
Upscale: 2
Lastly, the recipe for the last Juggernaut (Aftermath) version on SD 1.5:
SD 1.5 Base
Training Set 1 (0.65)
Absolute Reality (0.18)
epiCRealism V3 (0.2)
Humans (0.15)
Skin Enhancer (0.15)
Better Portrait Lighting (0.2)
Training Set 2 (0.55)
RPG 5 (0.38)
Divas (0.15)
And of course the NinjaFix made by @chillpixel
Thanks all those Creators for their hard work :)
Of Course you can run "Juggernaut" on Tensor.Art
Description
Added NinjaFix (more in the Description)
Minor Update to the BaseModel
FAQ
Comments (13)
What is the "NinjaFix"? Is it an embedding or lora we need to download from somewhere?
It´s already in the Model. It´s done by Chillpixel, so i will direct him to your message and hopefully he can answer it better than i can.
At least i can say you dont have to download anything :)
@chillpixel but what is it doing and how can we use it to optimize other models?
@Fenn NinjaFix fine-tunes the model's neural network parameters. I'm still improving it and not yet released code, but happy to apply it for model creators - just ask!
@chillpixel Amazing work!! Hope to use it soon!
@chillpixel so what is it exactly? a model, a lora? what?
The way you're describing it implies heavily that it's a method of merging? Without explaining everything can you at least tell us what this is?
@DarkAgent Without going into details (cause i simply dont know :D ) i would compare it to the Offset Noise LoRa but it "fixes" more than the lighting. I could upload the Original 1.5 Version, so that u guys can compare it to see the difference if thats helpful.
But @chillpixel will answer your question later i guess :) And i am optimistic that u guys will see that fix in some kind of form on CivitAi in the near future ;)
@DarkAgent Thank you for asking! NinjaFix is made up of:
- Ninja: neural network optimization algorithm
- NinjaNoise: custom offset noise model
- NinjaMerge: model auto-merger
Here's how the current version works:
1. Load three models:
- SD v1.5
- NinjaNoise
- Juggernaut
2. Normalize and recondition all models' neural networks.
3. Compare every neuron of Juggernaut and NinjaNoise against SD v1.5.
4. Predict which neuron is best and insert it into a new blank model.
To create NinjaNoise, the original Cross Labs offset noise model and three offset noise LoRAs from the community were normalized and reconditioned, then auto-merged and run through NinjaFix again. The resulting NinjaNoise model is now integrated into NinjaFix.
@chillpixel This looks nice! Do you have github repository so I can play with ninja magic as well ? Can't find it
@chillpixel hi there - would be keen to try this for my models, how best to go about this??
@vitiy82 sorry for taking so long to reply. NinjaFix is a special tool I use to help DreamWalker be better than others. Right now, I can't share the code with everyone. But I use NinjaFix to make free models and LoRAs better in the CivitAI community.
@residentchiefnz Hey! Sorry I've been busy and didn't see this earlier! Great job on your latest release — I'm excited to see Juggernaut included. I'd love to work together with you too. Just send me your next update before you release it, and I'll use NinjaFix to try and make it even better. That's how Kandoo and I have been doing things.
Details
Files
v1-inference (3).yaml
Mirrors
matrixHentaiPlus_v231.yaml
theoneabovethemall_v10.yaml
PornMasterV1-pruned-FP32.yaml
Refined_v11.yaml
PornMaster-Fantasy-FP32.yaml
PornMaster-fantasy-FULL.yaml
v1-inference.yaml
PornMasterV1-pruned.yaml
PornMaster-AsianV1-FULL.yaml
Refined_v11_Dark.yaml
Dream-Factory.yaml
Pop-Factory.yaml
A-Zovya_Photoreal-V1-Ultra.yaml
v1-inference (3).yaml
Based_v1.yaml
v1-inference.yaml
v1-inference.yaml
v1-inference.yaml
v1-inference.yaml
PornMasterV1-FULL.yaml
v1-inference.yaml
v1-inference (1).yaml
v1-inference.yaml
v1-inference.yaml
v1-inference.yaml
v1-inference.yaml
v1-inference.yaml
v1-inference.yaml
v1-inference.yaml
v1-inference.yaml
PornMaster-AsianV1-pruned-FP32.yaml
PornMasterAsianV1FP16.yaml
v1-inference (1).yaml
PornMaster-Fantasy-FP16.yaml
Visionary.yaml
Sweet-Factory.yaml
am i perfection v2.yaml
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.

