Unlock xxx images using flux, Qwen or Wan 2.2.
ATTENTION
Second version for WAN 2.2 T2V works with the usual models BUT for T2V I recommend some different models and LoRas:
https://huggingface.co/lightx2v/Wan2.2-Lightning/blob/main/Wan2.2-T2V-A14B-4steps-250928-dyno/Wan2.2-T2V-A14B-4steps-250928-dyno-high-lightx2v.safetensors instead of wan 2.2 high
https://huggingface.co/eddy1111111/WAN22.XX_Palingenesis/blob/main/WAN22.XX_Palingenesis_low_t2v.safetensorsinstead of wan 2.2 lowRemoved this recommendationNO speed/low step loras on high noise phase
https://huggingface.co/alibaba-pai/Wan2.2-Fun-Reward-LoRAs/blob/main/Wan2.2-Fun-A14B-InP-low-noise-HPS2.1.safetensors with STR:1.0 on low noise phase
https://huggingface.co/lightx2v/Wan2.2-Lightning/blob/main/Wan2.2-T2V-A14B-4steps-lora-250928/low_noise_model.safetensors with STR:1.0 on low noise phase
???
Profit!
All my videos have workflow embedded, the workflow is just a suggestion. But it works, and works very well.
First version for KREA.
Start using: 20 steps, res_2s, bong_tangent, STR 0.8
After that you can experiment with other arguments. It is not very good and deform hands and feet very badly.
ALL PICTURES ARE PNG WITH WORKFLOW INCLUDED
It is not perfect, sometimes it creates monsters, detached members and strange things, just try another seed or adjust your prompt. Eventually it get it right.
Examples:
Recommended Usage:
Checkpoint: flux1-dev
Strength: 1 (adjust depending on the image)
Sampling Method: DPMPP_2M
Scheduler Type: BETA
CFG Scale: 4
Steps: 35
Feel free to share your creations and feedback so I can improve this model and continue contributing to the creative community!
Description
Increased dataset a lot, all new pictures are realistic so it should show less unintended anime. Still needs work but improved a lot.
FAQ
Comments (41)
Amazing LORA. Why isn't anyone else commenting about this? Thank you!
Thank you for your kind words. I'm glad you liked
The quality and flexibility here is insane for a LORA. Amazing work.
Thank you for your kind words
Damn this is well done and really flexible. How many images did you use to train this?
I'm glad you liked. Thank you for your kind words. I used 1291 pictures with good captions adjusted by hand. It took almost 3 days training for the v3. 1291 pics * 80 epochs =103,280 steps
@alcaitiff Thanks for the info. Whatever you were doing; keep doing it, it seem to be working! Must have taken ages though.
@ConflictX yes, it's hard to make flux obey
Thanks for your work! i am trying to import your workflow but seems like i am missing the Lora Tester node. I think it's a group node you have in your local. Could you share it?
Thank you for the kind words.
A group is only visual, you may be missing some node inside the group. Click on it with the right button and "convert to nodes" this should show what is your real problem. Feel free to contact me on the chat if you need any help.
I added a new version of the workflow on the article here https://civitai.com/articles/13803 it's the file https://civitai.com/api/download/attachments/302127 let me know if it works for you
@alcaitiff It does now, thanks!!
Works very well with other character Lora!!!
Thank you for your review. I'm glad you liked it.
This is a great LoRA! It handles NSFW prompts very well and corrects some of flux's most annoying limitations by providing better anatomy and poses such as ones with spread legs and lifted skirts. It can easily support photo realism or stylized 2D art with slight prompt adjustments. A giant thumbs up for this one!
Thank you very much for your kind words. I'm glad you liked it.
waiting for a V.4 !!!!
Each version takes a lot of time because captions are super important and I must check each one individually. It'll take sometime.
I saw a comparison article, and this is the only one, which actually works.
Great job.
I'm glad it worked for you.
im new can anyone help me the this is i trained my lora with flux gym but whenever i use Mysticc or any other with it i always get different face not my lora face i tried to change the weight n all
very good! best nsfw lora from my testing
I'm glad you liked.
insane quality for flux dev base model, great job keep it up
Thank you very much
This lora at 1.0 strength combined w/ NSFW FLUX LoRA at 0.5 is <chef's kiss>. Thanks for the great work!
Something else I noticed while I browsed your sample images is that your prompting is amazing. Do you use any prompt generation tools to produce such amazing prompts for Flux?
Yes, I made an article that explains my tool and have the link for it.
@alcaitiff I checked your articles page, and saw two articles there. One that compared all of the loras, and one that explains how to use joy caption to provide detailed descriptions of a set of images.
Is that the tool you meant?
Yeah, a CLI for prompt creation. You can ask it anything.
Thanks
paste an image into chatgpt, it will spit out a nice prompt for you.
Great lora bro...you have literally nailed it!!!
Thank you. I'm working on V4. But it will take a while.
@alcaitiff No worries, take your time! If V4 is anything like this one or better, I'm already hyped. You've got real talent, bro!!!
@torc007689 It's up
I don't see the workflow... How can i get them? If i ssave photos are without template
Look in my profile, Mystic wf
How does that work? I thought it was impossible to depict male genitalia.
It was?
Details
Files
MysticXXX-V3.safetensors
Mirrors
NSFW-MysticXXX-V3.safetensors
MysticXXX-V3.safetensors
MysticXXX-V3.safetensors
MysticXXX-V3.safetensors
129_MysticXXX.safetensors
MysticXXX-V3.safetensors
MysticXXX-V3.safetensors
MysticXXX-V3.safetensors
MysticXXX-V3.safetensors
MysticXXX-V3.safetensors
MysticXXX-V3.safetensors
MysticXXX-V3.safetensors
MysticXXX-V3.safetensors
MysticXXX-V3.safetensors
MysticXXX-V3.safetensors
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.