March 22, 2025 Update:
Version 2.0 has been released for download on Civit.
The update on this page is to provide EP02, which is not a fine tune.
EP04 Finetune brings major improvements across the board:
✨ Drastically improved physics
🎥 Better handling of long-form video generations
🎯 Increased stability across a wide range of seeds
Whether you're pushing complex scenes or just looking for better consistency in your outputs, this version should give you way more control and realism. Stay tuned 👀
This update includes EP02 which is not a finetune V1.1 WIP.
EP02 Version 1.1 Wan Video:
Additional training as I tinker with Wan. Introduced more jiggling and bouncing breasts; the effect is subtle in some cases, exaggerated in others. V1.1.
Version 1.0 Wan Video WIP:
Test of the Wan Video Lora WIP. Give it a shot and share your posts. Thank you. Trained on 14B T2V however the post used I2V have not made any T2V runs as of yet.
Version 2.0 Enhanced + Cinematics:
While platforms like CivitAI are still evolving their creator support systems (especially for video-focused LoRAs), your support through Ko-fi is what keeps this project alive.
Why Early Access?
Be among the first to experiment with this cutting-edge model.
Your support directly fuels further development, testing, and refinement.
For the cost of a cup of coffee will provide additional support to fine tune your generation through DMS to fans.
This Lora version includes additional training on a curated, high-resolution dataset (images + videos) to push the boundaries of realism. The model now delivers:
✅ Improved physics/dynamic motion (e.g., walking, dancing, bouncing)
✅ Improved long form videos 97+ frames. Best at 129 frames.
✅ Cinematic elements and good coherence at denoise of 0.9.
✅ Enhanced detail in anatomy, particularly for more natural and lifelike results
✅ improved clothing designs.
✅ Sharper outputs thanks to the high-res training data
✅ Improved Camera Dynamics (Pans, Zoom, Orbits)
The model has only been tested using my workflow. Super Simple HV2V - Now with UPSCALER and INTERPOLATION
Version 1.5 Enhanced Details:
Additional training on a curated small dataset (Hence the high Epochs) of high-resolution images and videos to enhance the quality of outputs and physics of bouncing jiggling breasts. I haven't done too much testing on it but so far it seems to generate more realistic bouncing breasts.
Prompting in general is the same; however, just make sure you include more jiggling and bouncing. Refer to the trigger words attached to this version and the info below.
I haven't done many high-frame gens with this, just testing the ability to fine-tune or continue training existing lora's
Original Post:
This is a super simple LoRA model built for Hunyuan Video, designed to render stunning blondes with full, natural curves in high-quality motion. No overcomplication, no gimmicks—just clean, effective results.
No complex tweaks, no headaches—just plug it in and generate. It’s built to work, nothing more, nothing less.
Works with Bouncing Breasts HunyuanVideo lora. Use strength from 0.3 to 0.6 unless you want UBER boobs.
""\_(:3)_/""
Kling not required.
Load the LoRA in your preferred Hunyuan workflow (V2V or I2V). ( Refer to my Super Simple HV2V - Now with UPSCALER and INTERPOLATION with Super Smooth videos.)
Load any short clip of a woman with large breasts. Preferably with lots of bounce. Resolution I've been resizing to is 800 x 1120.
Use the trigger prompt(s):
A beautiful blonde woman with large breastsA beautiful pregnant blonde woman with large breastsAdd your scene description and adjust CFG settings, steps, and denoise as needed.
Example prompt style:
A beautiful blonde woman with large breasts stands nude in a softly lit room, her fair skin glowing. She has long, straight hair cascading over her shoulders, and her blue eyes gaze directly at the camera. Her breasts are full and natural, with pink nipples. She has a slender waist and a flat stomach. Her pose is relaxed, with her arms hanging loosely by her sides. The background features a white wall and a wooden dresser with a green plant on top. The camera angle is eye-level, capturing her from the waist up.A beautiful pregnant blonde woman with large breasts poses topless in a softly lit room, leaning forward slightly. Her fair skin is dotted with freckles, and she has long, wavy hair cascading over her shoulders. She gazes directly at the camera with a gentle smile. The background features a potted plant and a window with blue curtains, creating a serene atmosphere. The camera angle is eye-level, capturing her from the waist up.A beautiful blonde woman with large breasts stands in an office setting, wearing a patterned blouse. She begins by standing still, then she starts to move her hands, gently touching her chest. Her breasts jiggle and bounce as she moves, creating a rhythmic motion that emphasizes their size and shape. The camera captures her from the waist up, focusing on her upper body and the movement of her breasts. The background includes a whiteboard and a chair, indicating an indoor office environment. The lighting is bright, highlighting the details of her attire and the natural movement of her breasts.A beautiful blonde woman with large breasts stands outdoors in a natural setting, surrounded by tall grasses and greenery. She wears a black hat and a sheer black top that accentuates her ample bust. Her breasts are prominently displayed as she moves her arms, causing them to jiggle and bounce rhythmically. The sunlight filters through the foliage, casting dappled shadows on her figure. The composition focuses on her upper body, capturing the movement and texture of her clothing and skin. The background is blurred, emphasizing the subject's presence and the natural environment around her.CFG Scale: 1
CFG values in this range significantly improve animation quality and coherence. Higher values require significantly more VRam and tend to over-constrain the model, leading to stiff or unnatural motion. If generating smaller resolutions like 544 x 720 try CFG 2 and 49 frames.
Negative Prompt:
low quality, bad hands, bad teeth, bad eyes, bad limbs, distortion, talking, speaking
Description
Additional training on high-resolution images and videos to enhance the quality of outputs and physics of bouncing jiggling breasts. I haven't done too much testing on it but so far it seems to generate more realistic bouncing breasts.
Prompting in generate is the same; however, just make sure you include more jiggling and bouncing. Refer to the trigger words below.
I haven't done many high-frame gens with this, just testing the ability to fine-tune or continue training existing lora's
FAQ
Comments (15)
The image quality looks great. Looking forward to trying it. I do wish you would include video examples with the workflow intact, though. Those are super helpful.
The workflow with upscaler in my model page is what I've been using for most of my new content. I promise you can generate the exact same quality content using that workflow seen in my posts aside from the comparisons, in which I use Filmora for post processing.
@SingularUnity That's helpful and thank you, but it's also about prompts and loras that match the examples
@redwoodfog I understand
damn this looks amazing
thank you
Thanks for the wonderful job! I find this LORA very appealing and would like to include it in my workflow. However, I feel that the appearance of the female blonde color increases along with the breast workmanship. I would like to have a LORA where hair color is easy to adjust here. We would love to have a versatile LORA. Please consider this.
Thank you! I'm working on it; a few people have highlighted areas
FANCY WORK!!! Could you share the process of training Lora :)
Thank you for the LoRA! the wildly varying bust size makes it a key part of my workflow (breast expansion). My suggestion for a future version (or a new LoRA) is laying-down-on-her-back boobs of wildly varying sizes (where they flop off to the sides or are supported by the arm-squeeze). This is a rarity in all models I've encountered yet missionary LoRAs are commonplace, so there's a 'gap in the market' so to speak. Cheers!
Thank you, definitely a gap in the market. No pun intended.
I looked into making one for a bit, I needed one that showed this concept from the side rather than from directly above. Gathering the training material proved too difficult for now. Along the way I collected a different pose though - hanging/swinging , so I'll be making a Pendulous Breasts lora. What resource do you guys use to collect material for specific poses?
I gonna cum fucking hard
Just want to show my support and appreciation. Thank you so much for your work, I totally understand the effort put into training and preparing the datasets for these Lora.
Civitai is still crawling when it comes to supporting legitimate creators and you're definitely an artist and creator in my books. Tbf I was wondering when you'd start capitalizing on your work. I will be sending you a commissions request shortly on a concept I've thought, stay blessed in the meantime.
You just made my day good fellow. Especially since I learned the Creator Score is broken and hasn't been updating.