Welcome to DR34MSC4PE's latest WAN LoRA: C0WG1RL + REV3RSE C0WG1RL
9/6/2025: It's finally ready to share - Our WAN 2.2 Reverse Cowgirl Text-to-Video LoRa.
This thing is actually pretty crazy. The usual prompts/triggers still apply: Use r3v3rs3_c0wg1rl for when she's in the reverse cowgirl position. Use c0wg1rl when she's in the c0wg1rl position.
You can also use different camera angles, such as: Close-up detailed view, wide angle view, side angle view from the left/right, low/high angle view looking up/down, etc.
Try adding camera movements as well - "The camera slides slowly to the left, creating a parallax effect", etc.
This lora also works GREAT with a Text-to-Image workflow. Create reverse cowgirl images and add motion with our I2V Lora.
Have fun! It should work with other 2.1/2.2 LoRas.
Please share workflows that work great for you!
Please generate responsibly!
8/27/2025: At long last, v1 of our Cowgirl/Reverse Cowgirl Wan2.2 14b lora is LIVE!
Be sure to use some variation of 'she's positioned in the reverse cowgirl position'.
Sample prompts:
r3v3rs3_c0wg1rl, a nude woman with light brown hair small breasts and a navel piercing straddling a nude man with a visible erect penis penetrating her vagina. The man's hands grip her hips from below as she rides him in the reverse cowgirl position facing forward with a slight devilish smile. His penis is thrusting into her and his testicles are visible below. They are in an urban street setting at night with neon signs in the background including a prominent red and blue "NEON" sign. Cars are parked along the street and buildings with lit windows line the background. The perspective is from a low angle looking up at the woman. The lighting is bright highlighting their bodies against the dimly lit street. The woman's legs are spread wide and the man is lying on his back beneath her.c0wg1rl, a nude woman with light skin long blonde hair and small breasts straddling a man who is lying on his back. The woman's legs are spread wide and she is sitting on top of the man in the cowgirl position with his erect penis penetrating her vagina. She has a navel piercing and a focused expression on her face. The man's lower body and erect penis are visible while his upper body is out of frame. They are in a dimly lit room with a fireplace in the background casting warm light and shadows. Logs are stacked in the fireplace and there is a wooden shelf with various items on the right side of the frame. The perspective is from a slightly low angle looking up at the woman. The lighting emphasizes the contours of their bodies and the texture of their skin.REMEMBER: You can use both r3v3rs3_c0wgirl, or c0wg1rl, depending on her positioning. r3v3rs3 works best and can work for both positions.
This LoRa works GREAT with Kijai's WAN2.2 I2V 14B finetunes found here: https://huggingface.co/Kijai/WanVideo_comfy_fp8_scaled/tree/main/I2V
8/18/2025: WAN 2.2 TI2V v2 is READY!
After a few weeks in the smoker, she's ready to be served. Introducing v2 of our Cowgirl/Reverse Cowgirl Wan2.2 ti2v LoRa.
This LoRa works BEST with Kijai's WAN2.2 TI2V 5B finetune found here: Wan2_2-TI2V-5B-FastWanFullAttn_bf16
T2V and I2V Workflow's have been included - Sage-Attention and Triton are strongly recommended. An install guide for ComfyUI can be found here.
Sample prompts from the dataset have also been included. Feel free to use these as a base for your prompts.
Yes, penises can be a bit... long. We'll hopefully get that improved in a future iteration.
08/12/25: Introducing our Wan 2.2 TI2V 5B Cowgirl Lora.
This Lora is a work in progress. Epoch 280 seems to have the best overall performance during testing, but updates will be made as we get our training & dataset settings dialed in.
For WAN2.2 The trigger has changed, c0wg1rl now denotes the traditional cowgirl position. r3v3rs3_c0wgirl for, you guessed it, reverse cowgirl positions, which this Lora excels at. The WAN2.1 Loras remain unchanged, c0wg1rl is still the trigger.
Use the same types of prompts:
r3v3rs3_c0wg1rl, A nude woman with light skin small breasts and long blonde hair straddling a nude man with a muscular build and light skin in the reverse cowgirl position. The man is lying on his back on a beige couch with the woman facing forward with her eyes closed and mouth slightly open in an expression of pleasure. His erect penis is visibly thrusting in and out of her vagina. The woman is leaning backward and her hands rest on his chest and her legs are spread apart.
The background includes beige curtains and a white flower-shaped floor lamp. The lighting is bright highlighting their bodies and creating a clear view of their skin tones and textures. The perspective is slightly above and to the side focusing on their upper bodies and the action. A selection of the training captions have been provided to give you more of an idea on prompt structure.
Before you comment - yes, movement can be static, or positions can be wonky - please experiment with different phrases in your promts, and different tweaks to your settings.
You may have to add language such as "She bounces up and down", "The man's penis thrusts in and out of her vagina."
I've found the DPM++ Sampler family seems to work pretty well.
Before you ask: Yes, a 14b version is in the works.
Please experiment, and don't forget to share your findings and creations with the class!
===========================================================================
DR34MSC4PE is
c0ur4ge (trainer/qa/inference code) /
ERA5ER (captions/tooling code/data/qa)
Check out other models by DR34MSC4PE
Like our work? Consider donating for coffee: https://ko-fi.com/dr34msc4pe
PLEASE GENERATE RESPONSIBLY!
Description
Trained on the LOW version of WAN 2.2 T2V 14B.
FAQ
Comments (46)
Was this trained with some CG/3D images and videos? The close-up example looks like CG.
Nope, no CG.
Are you generating with full steps and cfg or did you upscaled the videos? your examples are too clean.. if you're using speed lora please tell your settings 😭
I tested this T2V a bit yesterday and found it often made "broken" looking poses that don't make much sense or extra legs without much difficulty. It could also do nice looking stuff with better prompting, but this one does not seem to be plug and play currently.
Could you expand on that a little, any example prompt/workflow you could share where this happened? Just trying to find out if there's something mixed up in the dataset, or if it's just a prompting issue.
Are you training on anime?
I feel like that causes an issue where the model has no "sliding in & out" context.
The model ends up thinking that the penis is attached to the vagina, and the penis is just stretching like rubber during movement.
It needs to see the texture of a real penis move in and out.
Im assuming thats why your lora suffers from the rubber/elastic penis effect anyway
No anime in dataset, but you're right - penises can look like meat tubes. For the majority of the dataset, the penis is fully inside and the head cannot be seen.
I'm currently training and testing a v2 of the i2v Lora with a lot of 'insertions' and 'pull-outs' where the head can be seen. Hopefully that will improve things.
@ERA5ER Definitely worth a shot hey, be interesting to see how it reacts
love
First, this LoRA is incredible.
I have had a little bit of trouble getting them to maintain eye contact with viewer. I think unfortunately looking down seems really baked in.
If anyone has any advice for overcoming this, I'd welcome it.
Couple of things to try -
Add something like "She closes her eyes" or "The woman looks down" to the negative.
Also try including something along the lines of "the woman gazes at the viewer"
Is I2V 14B workflow usable with 5090 ? I am getting OOM errors
It should be? Works fine on a 3090 and 4090. Is your resolution/step count set too high?
Works fine on a 5090 with the provided diffusers. Quality is high enough that I'd just get rid of the light lora's and not mess with fp16 for more details.
@ERA5ER yeah i think it is related to res/step combinations
Was this turned off for online generation just today? I did two other generations over the past few days and it worked fine, but this time it keeps spitting my buzz back at me and saying that the generation failed, without giving a reason.
Not sure! I haven't turned anything off on my end.
Never mind, 20 minutes later it finally generated.
still new at this, how do i find out what workflow was used and other settings? i tried using the default comfyui i2v workflow and i got a very fuzzy video
You can download the videos (at least mine) and the workflow is embedded in them. Drag and drop the MP4 files into your ComfyUI window to get the workfllow.
@ERA5ER thanks a lot <3 is there any way to tell if the vid containt workflow data or not. Ive been blindly downloading a lot of vids and most dont work.
@stinger250 Unfortunately there's no way to easily tell, that I know of, without downloading them and trying. I try to embed the workflow in my stuff, but not everyone does.
I am getting OOM on a 16gb VRAM and 96gb RAM, I dont see resolution change node.
You should try QuantStack it really helps with memory issues.
can this do anal?
Seems it can, as long as your image shows it
I can't get it to do slow penetration, be it cowgirl or reverse cowgirl.. the woman always moves fast.. Tried Slow, Slowly, moving slow on the prompt but it just not work.. someone can give me any tips?
This model (NSFW-22-H-e8) uses both high-noise and low-noise weights. Adding this model in... Increasing the weight by 0.8 causes a performance decrease, 0.85 improves performance, and 1 is the fastest. I don't know if this will help.
The video seems to loop back in the later half? Is there any way to stop this?
Disable the Ping Pong option in the Video Combine node.
@ERA5ER Thank you, though it seems the video length is cut in half as well. Do I increase it in Desired frame count or somewhere else?
@tkgg2219 Yes, pingpong just gets to the end of the video, then plays it backwards. It would of course cut the length in half when disabled.
Yes, increase frame count for longer videos.
Motion is great but with T2V it often seems to make the penis backwards and so the testicles are facing towards the man's face when doing the regular cowgirl position/prompt.
If anyone finds a solution for this, please let us know. This problem persists even if you explicitly prompt to try fixing this issue.
I’m a new user — could someone give me a basic workflow that works on an RTX 2060 6GB? Thank you so much!
Hi, what about start image ? We must put an image in position ? thanks
Yes, you must provide your own starting image. You can find a Wan 2.2 text-to-image workflow and use our t2v Lora to create a still image using the same r3v3rs3_c0wg1rl trigger word and caption. You could also try out our new Qwen 4 Play NSFW Lora, listed above, to create the still images. There are also hundreds of other sdxl or Pony models and Loras that can create the starting image.
@ERA5ER I think he was referring to the pose/framing of the subject in the start image. ie. does there already need to be a nude couple going at it or can we just have a woman standing/sitting zoomed in/out and the LoRA will try to add a guy and adjust the camera angle?
@BobRoberts nope, that would be a different Lora/workflow. Use this Lora when they're doing the sex (in the reverse/cowgirl position)
This won't really help with people going in and out of frame. Try without a Lora for that. I'm sure there are undressing Loras that can help.
I2V always closed eyes after few frame, even adding "looking at viewer" on positive and "closed eyes" on negative prompt
10/10
How can I increase the speed of thrusting? I tried "rapid movement" and it doesn't work
try more fps!
You list the keywords as being c0wg1rl and r3v3rs3_c0wg1rl with a "1", and c0wgirl and r3v3rs3_c0wgirl with an "i". Which is it?