I've done plenty of test runs but CIVITAI takes extra long to verify my posts now. Please check back soon!
May 23 2025 Update — I’m Back 🎉
Hi everyone! After a short hiatus and getting familiar with Civitai’s new guidelines, I’m finally back—and I’m dropping two fresh checkpoints for the Twerk LoRa.
V2.0 Free through Civitai
Exclusive Access: EP100 for $1.5 https://singularunity.checkoutpage.com/sutwrkep100
V2.0: Subtle but mostly Stable
What’s new? V2.0 tones things down from the punchy V1.0. It keeps the natural jiggle and muscle dynamics you loved, but with cleaner geometry and far fewer surprises in general.
Why you’ll like it: If you value predictability and more subtle results.
Exclusive Access: Epoch 100 — “Wild & Dynamic”
What’s different? Think of Epoch 100 as the adrenaline-fueled cousin. It can push motion intensity and deformations further than ever—but that can come with extra instability.
Pro tip: Bring patience (and maybe a few extra GPU cycles). With the right prompt-tuning and seeds, you can squeeze out some truly jaw-dropping animations.
A Quick Word on the New Rules
Civilta's policy changes took me by surprise at first with nearly 90% of my NSFW posts were suddenly hidden, but after digging in, I see where they’re headed. And tbh I think in the long run it will be a net positive.
April 2, 2025 Update
Hello everyone! I’m excited to introduce my Twerking LoRa model, specifically refined to capture natural, physics-driven twerking movements. Building upon previous feedback and learnings from previous models such as MotionCraft, this version focuses on the subtle jiggling, bouncing, and muscle deformation that bring twerk performance to life. As always, I rely on your results and feedback to keep improving—so please share your outputs!
Note: This model works great with the MotionCraft Lora!
Image Selection Guide
Realistic Female References
The LoRa has been trained primarily on realistic human figures. For best results, use images that clearly display hips and thighs.Good Lighting
Even lighting highlights soft tissue and muscle engagement. Dim or uneven lighting can reduce visible motion details.Minimal Background Distractions
A simple or uncluttered backdrop helps the model focus on the twerk movements.Body-Hugging Outfits
Nude, Shorts, leggings, or tight skirts are excellent choices—loose but form-fitting clothing reveals the bounce and elasticity of the thighs and glutes.
Trigger Prompts
Incorporate the word twerk or twerking alongside other descriptive phrases to elicit the best motion responses. Here are some examples:
“Twerking motion with natural jiggling and bouncing”
“Dynamic twerk emphasizing elastic thigh and glute deformation”
“Rhythmic twerking emphasizing subtle physics-driven movements”
“Steady twerk routine with visible muscle contractions and releases”
“Twerking performance showcasing realistic skin stretch and recoil”
Example Prompts
Below are five example prompts resembling the style you requested. They all include the key term twerk or twerking, with richly detailed descriptions emphasizing the body’s rhythmic, elastic movements.
1. A woman wearing a bright green sports bra and matching biker shorts stands upright, hands on her hips. As she begins twerking, her glutes contract in a fluid, rhythmic motion, producing gentle jiggling in her thighs and lower back. Each bounce highlights the elastic recoil of her muscles and skin.2. A woman is twerking, moving her hips and thighs in a rhythmic manner. A smooth, continuous, and cinematic camera pan shows her body moving in a fluid and smooth motion, with her muscles contracting and relaxing. The subtle physics-driven soft tissue dynamics are visible as her skin stretches and rebounds in response to her movements.3. A woman performs a captivating twerk routine, showcasing her natural jiggling, bouncing, and the elastic deformation of her thighs, glutes, hips, and lower back. With her blonde hair styled in braids and a striking tattoo on her upper arm, she bends over and lifts her legs to reveal her butt in gray tights. The scene captures the fluidity and grace of her movements, set against the backdrop of a cozy bedroom with a bed, a window with blinds, and a desk with a chair.4.A woman in a white ribbed crop top and olive green shorts stands with her hands on her hips, showcasing her toned back and fit physique. As she begins to twerk, her natural jiggling, bouncing, and the elastic deformation of her thighs, glutes, hips, and lower back regions are highlighted, capturing the subtle physics-driven movements that define her dynamic performance.Description
V2.0
What’s new? V2.0 tones things down from the punchy V1.0. It keeps the natural jiggle and muscle dynamics you loved, but with cleaner geometry and far fewer surprises in general.
Why you’ll like it: If you value predictability and more subtle results.
Epoch 100 — “Wild & Dynamic”
What’s different? Think of Epoch 100 as the adrenaline-fueled cousin. It pushes motion intensity and deformations further than ever—but that can come with extra instability.
Pro tip: Bring patience (and maybe a few extra GPU cycles). With the right prompt-tuning, you can squeeze out some truly jaw-dropping animations.
FAQ
Comments (28)
I'm confused. You say the exclusive access version is wilder and crazier, but in all the side-by-side examples but one, it's the most toned down, realistic version. The others shake a lot more wildly. Was this a mixup in the examples? The first set definitely has the exclusive one looking wilder, but I just see the opposite elsewhere.
I'm asking because I want something more grounded, and less exaggerated, but if the exclusive one generally is crazier, I don't actually want it. Regardless, version 2 is a good improvement, and it's great to see your work get even better.
Really? I don't believe there is any mistake. EP100 has been trained more and does tend to produce more dynamic movements this includes camera dynamics of the scene at the cost of stability, that being said when it gets it right the generation also has more fidelity overall. What I mean by unstable is that EP100 may require more generations at different seeds until you find a generation you like. That being said it's all about the use case, EP100 is the superior model if you have the time and resources; however, overall V2.0 is the most adaptable and balanced.
I should also add the the side by side comparison where done using all the same parameters only changing the LoRa.
You got 90% of your posts removed and think it's a net positive for the site??
this update looks way better
something wrong? i can't use WAN
Is On site Wan i2v disabled?
Don't know, NSFW stuff is restricted as I understand
@SingularUnity Why is it restricted? Because, mostly ALL addons are gone when i try to have more than one model to this.
@traygod_dev could be the base model I'm using 720p I'm not entirely sure. Though I know that I haven't been able to generate NSFW video content on the site. I hardly every generate anything on the platform as it is.
@SingularUnity im also using 720p image 2 video
It's pretty funny when you use it with other Loras for things like sex positions. If the strength is high enough, suddenly a penis becomes a goopy, fluid object XD
Still, great work!!
Didn't intend it to be used with penises but glad you're enjoying it.
I thought WAN2.1 lora's had to be cross_attn? But in this lora I see ffn and self_attn which I thought was more for image generation or SDXL?
It could totally be me being naive, but this LORA doesn't work for me in with WAN2.1 14B I2V base model from hugging face. But the FusionX_LORA does.
Don't pay attention to:
This appears to be a full model, not a LoRA.Not at all, when training a LoRa you can define what modules to focus on depending on your needs. For this particular purpose cross-attn was not needed. The base is the 14B I2V model for this LoRa.
In terms of it not working for you on hugging face I'm not sure why not.
""\__(:3)__/""
But fnn is specific to spatial refinement and feature transformations and is applicable to any training scenario.
Cross attention is the alignment between visual content and prompting.
Self attention is the temporal coherence and motion refinement between the scenes and prompts.
This model focused on a few parameters but since you asked nicely modulation and time embedding are two big ones. The former is responsible for subtle visual dynamics and adds fine grained adjustments, the latter aids in synchronizing and accurately pacing the generation across time.
I'll eventually get to typing up some guide to detailed training and finetuning to achieve specific objectives.
@SingularUnity Thank you so much for the response. You hold a large amount of hard earned knowledge and I appreciate you giving me a breakdown of some of my misplaced assumptions. It actually helps me a ton.
It's been difficult trying to parse through "it works on my machine!" in this new AI space because dependencies, hardware, and file naming/type/conventions are alllllll over the place (lora formatting too).
I'm on the non-comfy/non-automatic train. I'm implementing models and loras through just python/pytorch to see how everything works before getting dependent on workflows. If you're already, or ever think about going this direction, I'd love to be a pen-pal.
Thanks again for the reply. You've helped me already ask better questions. Have a great rest of your week!
Did they remove the lora off the on-site generator?
I think this one gives problems in combination with the causVid lora. I get weird output, colors are shifting etc
I got this working on t2v, just had to amp the strength up. Good lora.
to what?
Please figure out an isolated glute(cheek) twerk lora for all of us <3
do you need to prompt the asshole in order for it to appear? im getting smooth asses
2.2 please
Be nice if there were some black women twerking, considering it started with us. Just saying.
what lora loader should we use? she we place it in low or high noise?
Details
Files
SU_Twrk_EP55.safetensors
Mirrors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
tw-1824672.safetensors
tw-vid.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors
SU_Twrk_EP55.safetensors