Slop Twerk WAN 2.2 I2V - CivArchive (CivitAI Archive)

Slop Twerk WAN 2.2 I2V - WAN 2.2 I2V High

NSFW

Baked another half-assed slop LoRA. Trained sloppily on 11 of my old AI generated videos that were using various WAN 2.1 twerk LoRAs using PGC's modified GUI of Kvento's Musubi Tuner GUI with a minor fix for WAN 2.2 I2V training that just made the --i2v toggle only apply to latent cache creation instead of training. Trained for 20 epochs but the uploaded epoch is a 50-50 merge between epoch 7 and epoch 3.

Videos were lazily manually tagged as usual like
A black haired woman facing away from the viewer and wearing a short white skirt with no panties twerks her ass causing her butt to shake, jiggle, and ripple.
or
A red haired woman wearing a hat and jean shorts twerks causing her butt cheeks to clap together.
There is no relevant genital data in the dataset so use a genital LoRA if naughty bits are out of view in your initial image.

Prompting something about twerking and ass shaking will work probably. There is some breast jiggle baked in despite the dataset mostly consisting of back views and side views. If the initial image is from the front, the subject may try to turn around but there was limited testing on my part or in other words, I basically only tested it like once and it's that example image of Lancer Artoria in the swimsuit that almost turns around. A lot of the test image prompts were something simple like a description of the character and then

twerks causing her ass to jiggle and shake

Didn't do much testing but if combined with Slop Bounce, probably expect giga janky slop motion.

Training data includes the modified musubi_tuner_gui.py for PGC's GUI with the fixed toggle for properly caching latents for i2v as well as the .json and .toml files that contain the dataset settings and training settings. Also has the workflow images for the model page's initially uploaded videos although the videos should also have the Comfyui workflow included for once.

Description

High Noise and Training Dataset.

FAQ

Comments (42)

shinizayaOct 2, 2025· 13 reactions

CivitAI

If this kinda stuff is your slop I couldn't even comprehend what you could make if you try hard haha

psspsspsspssspssOct 2, 2025· 2 reactions

CivitAI

bring on the slop!

JellaiOct 3, 2025· 1 reaction

CivitAI

How did you make the data videos? Kling?

anonimoose

Author

Oct 3, 2025· 1 reaction

The dataset videos are just my old uploads of WAN 2.1 I2V with twerk LoRAs from Civitai.

gonfinal12Oct 3, 2025· 1 reaction

CivitAI

epic cool

NyxxiNyxOct 4, 2025· 3 reactions

CivitAI

This LoRA slaps hard AF.

Have some buzz my man.

RLS_AnimationOct 5, 2025· 4 reactions

CivitAI

Sir, you are a gentleman and a scholar.

🧐🍷

deltadotexe337Oct 5, 2025· 2 reactions

CivitAI

comes out blurry, cropped, and slow mo

warzonenoobie3625Mar 25, 2026

are u using iv2?

YinsuTarvoOct 6, 2025

CivitAI

amazing work! you are the best! is it possible for you to make one for dancing?

tyrnikOct 9, 2025

CivitAI

Where do you put the training dataset zip? Thanks in advance!

naomixloveOct 9, 2025· 2 reactions

Use the LoRAs only. Training dataset is just info to see what vids were used to make the LoRA.

tyrnikOct 9, 2025

@naomixlove ah good to know. thanks!

9919528Oct 10, 2025· 4 reactions

CivitAI

Incredible Lora, it’s the butter to my bread!

Thankyou ❤️‍🔥

pockeroperatorOct 13, 2025· 7 reactions

CivitAI

This Lora is incredible! Can't stop making bouncing videos

ailu91Oct 17, 2025· 5 reactions

CivitAI

Interestingly, it's also one of the better breast physics LORAs I've seen

supersuikaOct 18, 2025· 1 reaction

CivitAI

“Hello! I really love the realistic hip shake physics in your demo results. However, when I try to generate it myself, the output looks very unnatural and fake, almost like 'two balloons' shaking. Could you please help me understand why this might be happening?

I'm using an anime/2D-style base image. Does this LoRA specifically require real-life or 3D-style images to work properly? Alternatively, should I adjust the LoRA strength? I'm currently using a strength of 1. Any advice would be greatly appreciated! Thank you!”

anonimoose

Author

Oct 18, 2025

Hard to say anything concrete without an example video with a workflow from you, but this LoRA is probably more in tune with semi-realistic and 3D-style images simply due to the nature of the training dataset being that. There also aren't that many purely anime style videos in the Gallery either and I didn't really do much testing with anime styles let alone testing this at all so who knows. You could try adjusting the LoRA strength as you definitely should be adjusting it depending on what other LoRAs you are using considering how they may clash and exaggerate specific movements or degrade image appearance. Here are three generations though with strength at 1.0 for both High and Low with two using one of your Yoruichi gens and a random Lucina uploaded to Litterbox because Catbox broke so they'll only stay up for three days from now. Lastly, I wouldn't say this LoRA is particularly realistic physics-wise considering it's trained on AI slop while being sloppily trained. If anything, the outputs are probably more in line with being "unnatural and fake" than otherwise. Also, before I forget, all the example generations use Lightx2v LoRAs for faster generation. I didn't actually test this at any point with normal WAN 2.2.

https://litter.catbox.moe/fvb3f0ydsg6wdhxi.mp4
https://litter.catbox.moe/rina8qik8w511ig0.mp4
https://litter.catbox.moe/x11wvqfmdz5oky08.mp4

supersuikaOct 19, 2025· 1 reaction

@anonimoose Alright, figured it out! Another LoRA was messing with the results. I just adjusted its strength and that did the trick.

Ceazy_BoyOct 19, 2025· 5 reactions

CivitAI

Absolute cinema

57709Oct 25, 2025· 1 reaction

CivitAI

Perfection

SwaggerBoiOct 25, 2025· 5 reactions

CivitAI

Amazing how good it is.

bakaboy1234Oct 29, 2025· 1 reaction

CivitAI

what model did you use for the ref images that you animated

anonimoose

Author

Oct 31, 2025

iamddtla's Pornmaster-Pro noob-V3.
https://civitai.com/models/1045588?modelVersionId=1767015

a354286237512Nov 4, 2025· 2 reactions

CivitAI

What if physical characteristics change? For example, a short, thin person in a picture might instantly become a muscular woman with a large bottom in the generated video, simply inheriting her clothing.

warzonenoobie3625Mar 25, 2026

gota prompt it or make a 1st video and them in the right postion before u put it in

tadirel864375Nov 5, 2025· 8 reactions

CivitAI

What strength should I use High/Low for best results?

clzpetn804Nov 6, 2025· 1 reaction

CivitAI

Your LoRAs are amazing! proving that in a lot of cases, less is more when training LoRA

cr2022Nov 12, 2025· 3 reactions

CivitAI

Thank you for including the dataset I wish more creators where like you. The first dataset for Wan2.2 I can analyze and see how it is done. This will help me to understand how to create animated poses for Wan2.2 and train my own loras. There is so little information out there it seems how to properly train a Wan2.2 Lora and captioning the dataset. That a lot of people who want to train loras are not doing it because they did not know how to. The more people who learn how to train Loras and share them for free with the community the better it will be. Now I just need to find a Wan2.2 dataset similar to yours that shows me how to train characters.

DankMemeSpecialistNov 24, 2025· 4 reactions

CivitAI

my video generations weirdly lose saturation when I use this lora :( everything else about it is amazing

hiddentrapintere9103Dec 5, 2025· 3 reactions

CivitAI

Hi, can anyone tell me what model was used to generate the initial image? I like the look of it

AdorkableJan 25, 2026

Hi, have you found it yet? I really like this style of images too, just like you 😊

xDegenerateDec 19, 2025· 2 reactions

CivitAI

Working like a charm! Thanks!

Futanari Warning https://files.catbox.moe/mzbbi2.mp4

soulblazer333Dec 21, 2025· 10 reactions

CivitAI

"Half-assed" my ass. Great work.

johnmmartinez34608Jan 4, 2026

CivitAI

nice jiggle effect, will continue to use

bapybaraJan 29, 2026

CivitAI

skyrim modders take notes

deltadotexe337Jan 29, 2026· 2 reactions

CivitAI

motion comes out very blurry with i2v. anyone else experiencing this?

TitanxFilmsFeb 2, 2026· 2 reactions

same, no idea how to fix as I'm just starting out

shmaqkomcaj8460Feb 8, 2026

高噪声和低噪声要同时使用，分别接入两个模型才能正常运行

FrutiManoFeb 28, 2026

CivitAI

Will you make versions for LTX 2? Your LoRAs are the best!

B0rgheseMar 15, 2026

CivitAI

.safetensor High lora version, pls

JmDarkoApr 11, 2026· 5 reactions

CivitAI

Is there a way to stop the man's ass/pectorals to jiggle too? lmao

LORA

Wan Video 2.2 I2V-A14B

by anonimoose

Download (Beta) View on CivitAI

Details

Downloads

20,261

Platform

CivitAI

Platform Status

Available

Created

10/2/2025

Updated

6/18/2026

Deleted

Files

slop_twerk_HighNoise_merged3_7_v2.safetensors

Size:

292.63 MB

SHA256:

83fe88476f551aa0acdd119c6993de30910139b1635dc08c7f986fc3459d7cc5