⚠️ Important information
All models already include the Lightning LoRAs, except the SVI models, where Lightning is not included.
Do not use additional Lightning LoRAs on models that already have Lightning integrated, or the quality will be degraded.
For SVI models, you can use Lightning LoRAs if you want faster video generation.
2or3 ksampler (for svi) https://civarchive.com/models/2079192?modelVersionId=2668801
v2.1 (for nsfw v2) https://civarchive.com/models/2079192?modelVersionId=2562360
v2.1 with MMAUDIO (for nsfw v2) by @huchukato https://civarchive.com/models/2320999?modelVersionId=2613591
Another wf triple KSampler KSampler https://civarchive.com/models/1866565/wan22-continuous-generation-svi2-pro-or-gguf-or-32-phase-or-upscaleinterpolate-w-subgraphs-and-bus?modelVersionId=2559451
triple KSampler wf setup allows for more motion and helps prevent slow-motion issues. In exchange, your videos will take longer to generate.
For those having issues with my SVI workflow , you can try Kijai's wf here: https://github.com/user-attachments/files/24364598/Wan.-.2.2.SVI.Pro.-.Loop.native.json. Alternatively, you can try fmlf wf https://github.com/wallen0322/ComfyUI-Wan22FMLF/tree/main/example_workflows they will be simpler. There are others on Civitai that work very well too.
Qwen-VL workflow as an alternative to Grok for creating your dynamic NSFW prompts. Thanks @huchukato for his work: https://civarchive.com/models/2320999?modelVersionId=2611094
🟣 SVI Update – NSFW
⚡ Model Presentation (SVI-compatible version)
This update was made because the NSFW V2 models were not fully compatible with SVI LoRAs.
This version was created to work smoothly with SVI, while still functioning without them (though the workflow must be adapted).
Be careful, SVI LoRAs will only work with a workflow specifically designed for them, otherwise, it won't work.
There are two models available for SVI:
✔ Fast Move (FM) – Sexual scenes may differ from the Consistent Face model and will generally be faster.
✔ Consistent Face (CF) – Slightly better image quality, which may be preferable for anime-style videos; sexual scenes differ from Fast Move, but the difference is minimal.
You can also mix models between High and Low LoRAs:
FM (Fast Move) as High + CF (Consistent Face) as Low
CF (Consistent Face) as High + FM (Fast Move) as Low
Both combinations work and give slightly different results, offering more flexibility for your videos.
For this version, the main improvements include:
✔ Fully adapted for SVI LoRAs
✔ Greater flexibility: Lightning and SVI LoRAs must be loaded manually for custom workflows
🟣 SVI LoRAs – Strengths & Weaknesses
⚡ Overview
✅ Strengths
Best solution for making long videos
Excellent transitions between video segments
Reduced degradation compared to other solutions
Strong character coherence: the model retains information from the previous video, helping maintain consistency
⚠️ Weaknesses
Weaker prompt understanding
Weaker camera understanding
Videos are less dynamic
Sometimes slow-motion effect
(can be improved with proper Lightning LoRAs, dynamic prompts or triple ksampler)
🟣 SVI LoRAs – Download Links
⚡ Download
Note: Both LoRAs must be loaded manually in your workflow.
🟣 Suggested Lightning LoRA Combos (Optional)
⚡ Overview
You don’t have to use these Lightning LoRA combos. They are optional and allow you to fine-tune motion and degradation.
You can also use other Lightning LoRAs or assign different combos per video for more control.
🔥 Combo 1 – More Motion (Rapid Video Degradation)
High LoRA →
https://huggingface.co/Kijai/WanVideo_comfy/blob/709844db75d2e15582cf204e9a0b5e12b23a35dd/Lightx2v/lightx2v_I2V_14B_480p_cfg_step_distill_rank128_bf16.safetensors
Weight: 4
or
💜 Combo 2 – Less Image Degradation
or
or
🟢 Combo 3 – Balanced Motion / Moderate Degradation
High LoRA →
https://huggingface.co/Kijai/WanVideo_comfy/blob/709844db75d2e15582cf204e9a0b5e12b23a35dd/Lightx2v/lightx2v_I2V_14B_480p_cfg_step_distill_rank128_bf16.safetensors
Weight: 3
Low LoRA →
https://huggingface.co/Kijai/WanVideo_comfy/blob/709844db75d2e15582cf204e9a0b5e12b23a35dd/Lightx2v/lightx2v_I2V_14B_480p_cfg_step_distill_rank128_bf16.safetensors
Weight: 1.5
🧠 Advanced Usage Tip
You can disable global Lightning LoRAs (in my wf) and assign different combos per video:
Combo 1 for Video 1, Combo 3 for Video 2, Combo 2 for Video 3.
Each combo produces different motion and degradation behavior.
If you want to create several-minute-long videos while maintaining high quality, it is possible, but it will take a very long time. You just need to avoid using Lightning LoRAs and use the full model instead.
🧠 Dynamic Prompts – Better Control & More Motion
You can use dynamic prompts to have better control and this helps to make the video more dynamic.
You just need to give this example prompt to an LLM like ChatGPT. It will be enough for it to describe the image you have and what you want as a video while keeping the same structure of the prompt in the following example.
⚠️ This will be a NSFW prompt; ChatGPT will not accept it.
You can use GROK https://grok.com, which can make NSFW prompt modifications.
For more examples of prompts with different poses, check:
Enhanced FP8 Model.
Give these prompts to GLM with the dynamic prompt structure if you want.
Example 1
(At 0 seconds: Wide shot showing a slightly overweight man casually walking down a city street, camera fixed in front, urban environment with buildings and cars.)
(At 1 second: Suddenly, a massive shark bursts from the pavement ahead, looking terrifying at first, pavement cracking, dust and debris flying, camera from side angle.)
(At 2 seconds: Medium shot from the side, the man stumbles backward in shock, while the shark dramatically slows down and strikes a comically exaggerated sexy pose, revealing large, exaggerated shark breasts, covered by a colorful bikini.)
(At 3 seconds: Close-up on the man’s face, eyes wide in disbelief, as he turns to look at the shark, small cartoon-style hearts floating above his head to emphasize his amazement, camera slightly low-angle.)
(At 4 seconds: Dynamic travelling shot showing the man frozen in the street, the shark maintaining its sexy pose, water splashes and debris still moving realistically, urban chaos around.)
(At 5 seconds: Wide cinematic shot pulling back, showing the man standing in the street, staring at the bikini-wearing shark with hearts above his head, epic perspective highlighting absurdity and humor.)
Example 2 – Anime NSFW
(At 0 seconds: The couple in a cozy bedroom, anime style, soft lighting highlighting their intimate embrace, her back arched slightly as he positions himself.)
(At 1 second: The man’s hips moving rhythmically, the head of his penis sliding effortlessly into her vagina, her body responding with a gentle, fluid motion, anime-style motion lines emphasizing the smooth penetration.)
(At 2 seconds: Her back arching deeply against him to intensify the pleasure, hips swaying with each thrust, breasts bouncing subtly, small hearts floating around them to capture the erotic energy.)
(At 3 seconds: Her face, eyes closed in bliss, a soft moan escaping, hands resting behind her head, anime-style blush on her cheeks, the air filled with a seductive aura.)
(At 4 seconds: The man penetrating her deeply, her body moving in sync with his, the bed sheets slightly rumpled, the room’s warm lighting enhancing the intimate, lustful atmosphere.)
(At 5 seconds: The couple locked in a passionate embrace, the scene exuding vibrant, seductive energy, anime style with smooth lines and soft shadows.)
🟣 Lightning Edition – NSFW I2V V2
⚡ Model Presentation (2 new versions available)
I originally planned to release only one V2, but some people preferred the NSFW V1 over the Fast Move V1 version, so depending on what you’re looking for, one version may suit you better than the other.
For these V2 versions, I tried a new approach:
✔ I made sure that most sexual poses work, while the model is also good for SFW content
✔ More flexible for general use
🔥 NSFW Fast Move V2
Improvements included in this version:
Better prompt understanding
Better camera understanding
Reduced unnecessary back-and-forth movements outside sexual poses
(cannot be completely removed, but strongly reduced)Improved bounce effect on the buttocks
If a man appears, he will no longer automatically attempt to penetrate the woman when she is nude
This version is designed for those who want more dynamic scenes with more movement.
💜 NSFW V2
The difference between this version and NSFW Fast Move V2:
Less camera control
Less camera understanding
But body movements are less pronounced (breasts and buttocks)
Some preferred V1 NSFW to V1 NSFW Fast Move, and this version keeps that spirit
For varied sexual poses, check the previews — there are many.
You can use the shown prompts and adapt them to your images, but of course, other prompts will work as well.
Don’t hesitate to use other LoRAs for creating specific concepts.
🔧 Recommended Settings
Steps: 2+2
(Jellai recommends 2+3 for even better results, and I agree)Sampler: Euler simple
CFG: 1
This model already includes the Lightning LoRAs, don’t use them, or the quality will be degraded
You need to download both models: H for High and L for Low.
Here is an example made with the base workflow + the First/Last Frame workflow + upscaler (only on the starting images) + inpainting (cum).
https://civarchive.com/images/114733916
🔞 LoRA Examples (with links and recommended weights)
LoRA: Penis Insert WAN 2.2
LoRA weight: 1
Face → Doggy
https://civarchive.com/images/112864483Face → Missionary
https://civarchive.com/images/112864718
https://civarchive.com/images/112864835Face → Doggy (leg aside)
https://civarchive.com/images/112864885Another example
https://civarchive.com/images/112864972
Other examples with different LoRAs
Face → Reverse Cowgirl — LoRA weight: 0.7
https://civarchive.com/images/112865189Face → Cowgirl — LoRA weight: 0.5
Face → Missionary — LoRA weight: 0.3
https://civarchive.com/images/112865381Face → Missionary — LoRA weight: 0.5
Face → Doggy leg aside — LoRA weight: 0.6
https://civarchive.com/images/112865551Face → Doggy — LoRA weight: 0.7
https://civarchive.com/images/112865696Face → Spoon — LoRA weight: 0.7
https://civarchive.com/images/112868124
Recommended LoRA to obtain a very realistic vagina and anus:
https://civarchive.com/models/2217653?modelVersionId=2496754
💡 Tip for Anime Style:
If you’re working in an anime style, feel free to follow the advice of @g1263495582 thanks to him for this.
Try these LoRAs together or separately; they help maintain face consistency:
🔹 Note:
For these LoRAs, use only Low Noise.
For the examples mentioned above, use High and Low Noise as indicated.
LoRA weight: 0.3
Many things can be done with the model, but don’t hesitate to use other LoRAs for specific purposes. And don’t hesitate to lower the LoRA strength to preserve the face as much as possible.
📷 Dynamic prompt example with different camera angles:
(at 0 seconds: wide shot showing the woman standing in the snowy plain, a massive giant dragon emerging behind her, snow cracking and dust rising).
(at 1 second: the woman jumps backward onto the dragon’s back as it bursts fully from the sky, camera tracking the motion from a side angle, debris and snow flying).
(at 2 seconds: medium shot from the side, the woman balances heroically on the dragon’s back as it begins to run forward across the snowy plain, slow-motion on her posture).
(at 3 seconds: close-up on the woman’s determined face, camera slightly low-angle to emphasize her heroic stance, snow and debris flying around).
(at 4 seconds: dynamic travelling shot alongside the dragon, showing the snowy plain, scattered debris and ice fragments flying everywhere as it gains speed).
(at 5 seconds: wide cinematic shot pulling back, showing the dragon taking off with the woman riding on its back, soaring above the snowy plain, epic perspective with snow, wind, and scale emphasizing the drama).
For available camera angles, check further down in the “cam V2” model description.
🔞 Normal Clip vs NSFW Clip:
You can also use the NSFW version for your clip.
It can bring positive effects for sexual scenes, but it can also cause issues, as in this example:
NSFW Clip:
https://civarchive.com/images/112864204Normal Clip (same seed):
https://civarchive.com/images/112864295
Here are the links to NSFW clips (thanks to zoot_allure855 for correcting the BF16 version):
BF16 fixed version:
https://huggingface.co/zootkitty/nsfw_wan_umt5-xxl_bf16_fixed/tree/mainFP8 version:
https://huggingface.co/NSFW-API/NSFW-Wan-UMT5-XXL/tree/main?not-for-all-audiences=true
❓ If you have any questions, don’t hesitate to ask!
Many people ask me for help via private message. You can do that, no problem, but I would appreciate it if you could do it in the comment section, it could help other people. Thank you.
🟣 Lightning Edition – NSFW I2V camera prompt adherence
⚡ Model Presentation (2 versions available)
I decided to release 2 versions. Both can produce different results, as you can see in the previews (the seed was the same).
Fast Move Version: provides more movement, and the movements will be faster, with better prompt understanding and camera handling
(you can see it in the 4th preview "fast move high" where the man slaps the woman)Natural Motion Version: offers more natural breast movements depending on the situation and produces slower scenes.
👉 Check both and choose the one that works best for you.
This NSFW edition is, of course, focused on sexual poses.
You should achieve very good results.
To create specific concepts, feel free to use specific LoRAs — they work very well with this model.
🔧 Recommended Settings
Steps: 2+2
(Jellai recommends 2+3 for even better results, and I agree)Sampler: Euler simple
CFG: 1
This model already includes the Lightning LoRAs, don’t use them, or the quality will be degraded
You can also use this T5, which may improve understanding:
https://huggingface.co/NSFW-API/NSFW-Wan-UMT5-XXL/tree/main?not-for-all-audiences=true
📝 Usage Tips
Start with a clean, high-resolution image for better results.
The model will not change the face; if it does, increase the resolution.
Also adjust your prompt if you are not getting what you want.
Wan understands some terms well, but not all.
For prompts, check the video previews: adapt them to your image.
Other prompts will work too.
🎥 Training
I trained 2 LoRAs:
The first one using several videos and images of different sexual positions.
The second one to bring more dynamic motion.
Respect to everyone who creates this type of LoRA — it requires a lot of work.
🙏 Credits
This model wouldn’t exist without the incredible work of these creators:
alcaitiff : https://civarchive.com/models/1295758/nsfw-fluxorwan-22orqwen-mystic-xxx?modelVersionId=2300332
Sweet_Pixeline : https://civarchive.com/models/1844313/penis-play-wan-22
anonimoose : https://civarchive.com/models/2008663/slop-twerk-wan-22-i2v
dtwr434 https://civarchive.com/models/1331682?modelVersionId=2098405
A special thanks to Alcaitiff and CubeyAI, two very kind and humble people.
🔔 Important
Please don’t support my work with buzzes here, I don’t need it.
If you want to support someone, support the creators listed above — they truly deserve it.
💬 Feedback
Feel free to give me feedback, positive or negative, to help improve future updates.
Update WAN 2.2 V2 CAM I2V – NEW feature Camera & Prompt Improvements
This custom version of WAN 2.2 I2V has been updated to deliver better prompt comprehension and improved handling of camera angles and cinematic movements. It provides more accurate scene interpretation, smoother transitions, and enhanced control over dynamic.
Key Features:
Excellent understanding of prompts and scene composition.
Supports various camera angles and movements, including zoom, dolly, pan, tilt, orbit, tracking, and handheld shots.
Ideal for cinematic storytelling, animated sequences, and creative video-to-image projects.
Flexible multi-step prompts, standard 4 steps (2+2), can be increased for higher fidelity.
Recommended sampler: Euler simple.
You can, of course, use your usual prompts; this is just one example among many.
Example Prompt with Different Camera Angles
(at 0 seconds: wide frontal shot of a man standing in front of an open fridge, cinematic lighting, subtle ambient kitchen reflections, the fridge contents visible, camera static).
(at 1 second: medium shot from the front as he opens the fridge fully, reaches for a can, slight zoom-in to emphasize the action, cinematic framing).
(at 2 seconds: camera shifts to a side medium shot, tracking him as he lifts the can to his mouth, fluid movement, maintaining lighting and reflections).
(at 3 seconds: camera starts a smooth 360-degree orbit around the man, following him as he drinks from the can, motion fluid, background slightly blurred for cinematic effect).
(at 4 seconds: close-up on his face and upper body while drinking, orbit continues subtly, fridge reflections accentuating realism, cinematic polish).
(at 5 seconds: final wide shot as he lowers the can, camera completes orbit to original angle, showcasing the kitchen space, lighting, and dynamic movement).
Available Camera Movements
Zoom / Dolly
zoom in
zoom out
camera zooms in on subject
camera zooms out gradually
dolly in
dolly out
camera dollies in slowly
camera dollies out steadily
crash zoom
Pan
pan left
pan right
camera pans across the scene
gentle pan left
sweeping pan right
Tilt
tilt up
tilt down
camera tilts up to reveal…
camera tilts down from…
Orbital / Tracking / Arc / Rotation
orbit around subject
360° orbit
camera circles around
tracking shot
camera tracks alongside subject
arc shot
curved camera movement
Other Movements & Styles
static camera / static shot
handheld shot
camera roll
Note:
LoRAs work perfectly with this model, offering full compatibility and consistent results across styles and concepts.
# SUPPLEMENTARY ADVICE
You can use negative prompts, but be careful: this will double the generation time.
Only use them if you really want to prevent something from appearing in your video.
In that case, enable the corresponding node; otherwise, keep it disabled.
⚠️ Important
The model must be used with CFG set to 1, so negative prompts do not work by default.
However, there is a simple way to enable them.
How to enable negative prompts:
Open the Manager
Search for kjnode and install it
In your workflow, add the WAN Nag node
To use it correctly:
Connect this node after the LoRA Loader
Feed it with the negative prompt
Then connect it to the first kSampler (High)
👉 Only use this option when necessary, to avoid unnecessarily increasing generation time.
And here is the negative prompt for unwanted movements:motion artifacts, animation artifacts, movement blur, motion distortion, dynamic distortion, shifting shapes, unstable render, instability, wobbling effect, jittering effect, vibrating render, inaccurate details, visual noise, distorted surfaces, rendering errors, warped shapes, exaggerated butt movement, jiggle, overanimated hips, unnatural butt motion, hyper bounce, extreme curves, distorted hips, unnatural pose, unrealistic anatomy, deformed body, disproportionate body, floating limbs, blurry textures, clipping, stretching, low detail, messy background, artifacts, butt bounce, moving hips, swinging hips, shaking butt, wiggling butt, moving lower body, moving pelvis, jiggling buttocks, bouncing butt, unstable stance, unnatural hip motion, exaggerated hip movement, hip sway, hip rotation, bottom motion, pelvis motion, wobbling hips, fidgeting lower body, dancing hips, pelvic movement, motion blur, unnatural movement
And here is the negative prompt from the official ComfyUI workflow:色调艳丽，过曝，静态，细节模糊不清，字幕，风格，作品，画作，画面，静止，整体发灰，最差质量，低质量，JPEG压缩残留，丑陋的，残缺的，多余的手指，画得不好的手部，画得不好的脸部，畸形的，毁容的，形态畸形的肢体，手指融合，静止不动的画面，杂乱的背景，三条腿，背景人很多，倒着走
Note: If you use the default workflow without the node, this negative prompt will not work.
UPTDATE: Lightning Edition – T2V
I’m not really experienced with T2V myself, but a colleague who works with it a lot tested this version and confirmed that it performs very well. From the few tests I managed to do on my side, I got the same impression, although I haven’t compared it directly with the base version yet.
The settings are the same as in the image-to-video version: 2+2 steps or more if you prefer, CFG at 1, and Euler simple (though other samplers also work great).
I don’t plan to make an NSFW version for the T2V release,
but for the I2V version, I’m already quite happy with the first results.
Update WAN 2.2 V1.1 I2V
Updated version of the original Lightning merge — same settings 2+2 steps or more, featuring more movement and smoother flow (depending on the prompt).
The model already works very well in NSFW. Just use the right LoRAs, and the movement will improve.
WAN 2.2 V1 I2V
This checkpoint is based on the original WAN 2.2, with the Lightning WAN 2.2 and Lightning WAN 2.1 LoRAs already integrated. This improves image quality, makes motion smoother and more dynamic, and removes the slow-motion effect that can occur with the Lightning models.
A common setup is to use 2 steps on the high model and 2 steps on the low model, though other settings may work as well. Do not apply the Lightning LoRAs manually — they’re already included in this checkpoint.
My workflow
https://civarchive.com/models/2079192/wan-22-i2v-native-enhanced-lightning-edition
Description
FAQ
Comments (68)
I went through all camera angles and movements v2 can show and made a video out of it with examples: https://youtu.be/fGAPouLK_ng
Well, I’ve only watched the beginning, and it already looks really high quality. I plan to watch it entirely tomorrow. Thank you already for this, and I’ll of course give you feedback. Could I put it in the description?
amazing video, i wonder if it would be possible to share the prompt constructor node. looks very promising.
I just watched your video completely, and I’d like to honestly say that it’s really of very high quality. It’s complete, well-structured, well edited, and you express yourself very well. That must have taken a lot of work. I’m waiting for your permission to put it in the description. Thank you for your work.
I'm using an I2V NSFW version, but why is she always shaking her body, waist, hips? She does it all the time, even when the situation totally doesn't call for it. Is there any way to stop this?
It's all I can get for outputs. Ignores concept loras completely. Just full body gyrating.
I can confirm the same issue using the latest I2V version. It's pretty hilarious.
Hello and thank you for your feedback. It seems you’re not the only one based on the responses. I’ll take a look at it and try to fix it. To be honest, I mostly focused the model on sexual positions, but I think I can improve that.
I think I should rename the model to sexual pose while waiting for an update, it would be more honest. And thanks again. Don’t hesitate to tell me if you have any other issues.
I just ran a quick test, swapped out Lora with the following. Looking good, will test more.
https://civitai.com/models/1648982/wan-nsfw-posing-nude
@paultheblock Hello, and thank you for your feedback. This could be useful for others. I think I’ve already found the solution, but this time I prefer to run more tests. Thanks again, Paul.
I think I’ve fixed the issue, but to be sure I’d like to run more tests. Does anyone who has this problem have a prompt where it happens? I’d like to test it and post the result.
@taek75799 I have 3 prompt, all the same issue happend. thx for your effort!
1.
The camera is fixed on a stationary mount. The camera is fixed and steady throughout the video. A woman is sitting, lifting her straight legs up and apart to create a V-shape. Feet over head, Her hands are on the floor beside her hips to maintain balance as she looks straight ahead at the camera.
2.
handheld shot. she has an unrealistically upturned gaze, mouth wide open and long tongue sticking out, hands making a V sign.
3.
nsfw. explicit content. cinematic lighting.
(at 0 seconds: static shot, she is straddling a penis, moving her body up and down slowly, cinematic framing).
(at 2 seconds: camera starts a smooth 360-degree orbit around subject, motion fluid).
(at 4 seconds: crash zoom, The penis makes rapid in-and-out movements inside the vagina).
(at 5 seconds: camera dollies out steadily to a wide shot, background slightly blurred for cinematic effect).
@aawass0001167 Hello Aawass,
Thank you very much for your help. I hope I didn’t bother you too much with this.
I’m testing with the version available here and with the fix.
@aawass0001167 Here are the results.
I first tested your initial prompt on the posted version, and it’s clear there is an issue. There are movements that shouldn’t be there, and the prompt is only partially followed. She lifts her legs, but doesn’t form the V-shape.
Then I used the new model and made three tests. All three looked similar, so there is no cherry-picking. Only one test was done for the second prompt.
Finally, I’m showing you the adult scene: I used version 1.1 of my model, then V2 for the camera angles.
It’s important to note that adding NSFW elements will inevitably affect V2’s camera behavior. You can test this yourself: when using V2 with certain NSFW LoRAs, the camera angles become more static because these LoRAs are often trained on fixed shots.
is there a version with the loras included that is SFW? I love this model but it adds too much motion to things that shouldn't haha! Thanks!
Hello, and thank you for your feedback, it helps me a lot.
May I ask in which situations you feel there is too much movement? I’d like to understand what specifically bothers you so I can improve.
I did try to add a lot of dynamism to the scenes, so I’ll take your comments into account and adjust where needed.
To answer your question, check the description ;)
@taek75799 I think i had downloaded the wrong version. The Q4 seems to work just perfectly. Thank you again!
@HairyAI I'm glad everything is working well. Thank you for your feedback!
Any chance of a Q8 version of the new release? Thanks for your work!
It seems there was an issue unrelated to the intimate scenes. That’s why I renamed the checkpoint.
The women in the outputs appeared to be “moving too much,” especially around the hips, ass, tits, even when it wasn’t mentioned in the prompt.
I’ve corrected this behavior and I’m planning to release the new version soon in FP8.
If I receive good feedback, I’ll gladly create a Q8 version and possibly others as well.
the new sexualpose i2v seems to change the face much more than previous models. is that expected?
Hello,
it’s possible that some LoRAs might affect the faces. I tried to minimize this effect as much as possible, and I personally haven’t noticed significant issues, but I do take your feedback into account.
Have you tried increasing the resolution?
It turns out I was comparing v1 fp8 to v2 gguf. Switching to v2 fp8 I don't see the issue, and higher res does help. Thanks
@Kiefstorm Thank you for your feedback.
I’m planning to release an improved version of the NSFW model, and I hope you won’t encounter this issue anymore with it.
@taek75799 Awesome, it's a great model. Looking forward to it
@Kiefstorm Thank you! ;)
The camera tracking and prompt comprehension in V2 are amazing—really impressive work! Thank you so much!
However, in the NSFW version, the face tends to change more easily compared to V2, and the prompt adherence is weaker.
Since the NSFW quality in V2 isn’t very good, I tried an idea: using V2 for the High model and NSFW for the Low model.
This allowed me to maintain the High prompt quality while still supporting NSFW generation.
It seems that tuning the shift value is very important. If the SNR boundary isn’t around 0.9 (as recommended for WAN2.2 A14B), extreme outputs can occur on the Low side (for example, a sudden appearance of a male lower body).
For reference, the SNR boundary reaches around 0.9 with shift = 9.0 for 2+2 steps and shift = 6.0 for 2+3 steps.
Hope this helps!
Hello, and thank you very much for your feedback.
That’s a great idea you came up with!
To be honest, I was considering removing the sexual-pose version after some valid criticisms, but they helped me fix the issue and improve the model’s understanding.
I will publish the updated version soon, but before that, I plan to test your idea.
Thanks again!
By the way, I’m going to remove it. I’d rather avoid people wasting their time with a version that isn’t up to standard. Thank you.
For anime I've had a lot of luck using this for low noise and easywan in high noise. I'd really recommend staying away from nsfw merges. Just make yourself a few workflows for the different loras you need and switch between them. Infinitely better generations and less schizo than nsfw checkpoints.
Hello, and thank you for your comment.
I have to say I agree with you, but not entirely. I don’t use any NSFW model directly, I prefer to use my V2 model and pair it with certain NSFW LoRAs to have better control.
However, I know many people who don’t want to do it this way and prefer a ready-to-use model.
Not to mention the requests I’ve received privately:,just look at SmoothMix, which is the most downloaded model, followed by Dasiwa. People really seem to enjoy these two models.So, there is indeed a real demand.
But I completely agree with your advice.
Those who want to experiment will benefit from it, even if it takes a little more time.
Your model is extremely helpful! I suggest keeping the old versions and labeling them as v2, v3, etc., so that switching versions is less confusing. Thank you.
Hello, and thank you very much for your feedback.
I’m glad it satisfies you, and thank you also for your suggestion.
Actually, there are three models:
V1,
V1.1, which is just a minor update,
and V2, which introduces improved cameras.
By the way, if you haven’t tried it yet, check out the new node released in the description — it might suit you.
Thanks again!
Unless you’re talking about when you download the Wan model… in .safetensors?
If that’s the case, it is indeed a problem. I don’t know why the name gets completely changed when I post them, and I can’t manage to change it.
@taek75799 I used the previous version of SEXUALPOSE, and later its name seemed to change to v2. I noticed that in this version, many poses tend to be NSFW, and the variation has decreased a bit. So I planned to switch back to your earlier v2 version, but it seems that one’s no longer available. Now the current SEXUALPOSE is also gone, so I’m not sure if your current v2 is the same as the old v2, a mix of SEXUALPOSE’s v2, or an entirely new v2, haha—I’m confused.
@_Glint_ Haha, that’s not surprising — there are so many versions out there!
To be clear, I initially named my model “NSFW,” but after some feedback, it turned out that the bodies tended to move even outside of sexual poses. So I decided to rename it “Sexual Pose,” which felt more honest.
Honestly, I wasn’t really happy with the result. It worked great with my own prompts but didn’t perform consistently with others, so I decided to take it down.
The models available here are not NSFW. The V2 is excellent for handling different camera angles and provides more dynamic action.
And of course, I don’t like disappointing people — I promised an NSFW version, so I’ll be releasing it this weekend.
If the feedback looks good, I’ll make multiple GGUF versions afterward.
Thank you for your comment, it’s been really helpful!
What happened to the sexualpose version? 🤔
Hello, I’m sorry.
It had some issues outside of the sexual poses, as well as some prompt-understanding problems, so I preferred to remove it to avoid wasting anyone’s time.
I’m currently working on an update, and I hope this one won’t have any problems.
I’m doing my best, and I’ll post it this weekend.
every time I render something on my local machine the output is like blobby and grainy, and I can never get the desired results. anyone know why?
Hello, could you give me more details?
Which model are you using? How many steps on the 1st and 2nd k-sampler?
Ideally, you could post a video on Civitai and leave the metadata so I can retrieve it and see where the issue might be, if you’d like.
@taek75799 I have been using swarm UI. I have an init image, and I'm using Euler/simple, and there are 2 locations I can set CFG and steps.. Video generation, and in another place, but I'm confused about which one actually matters, I've been trying to set CFG to either 3 or 1, because I don't know which I should be using even though the generation parameters are posted, I'm still uncertain about which settings apply to general use. steps I have set between 4 and 8 because speed, and I also am still confused about which setting matters. Also I was told by AI that I need to use the LOW noise version on the refiner, and High noise as the primary generation. I don't know if that's right or not.
@GoronPlus Thank you for your response.
I’m not very familiar with SwarmUI, but that’s not important — only the correct settings matter.
So, if you’re using the base Wan model, you need to add the Lightning LoRAs (HIGH and LOW) so you can set just 2 steps for the HIGH and 2 steps for the LOW, or 3 and 3 — that works too.
In CFG, set it to 1; Simple Euler works well.
Of course, these settings apply when using the Lightning LoRAs.
Otherwise, I’m not sure if it’s been done, but you can use my V2 model in FP8: you won’t need to add the Lightning LoRAs, as they are already incorporated in the model.
You just need to apply the settings as I indicated, and feel free to give me feedback if you’d like.
@taek75799 Swarm UI is great, imagine a clean and minimalistic interface that has the ability to rearrange the interface based on a Comfy UI workflow (that literally loads on another tab that you can opt to make your interface adhere to) it manages your model dependencies too, for example if you load something that requires a specific model, it will download the other model so that you can use it without having to do a ton of setup. anyway, that aside since you didn't address the low to high thing, I'm assuming that AI was correct about how to use the Low and High Models.
@GoronPlus I tried SwarmUI a long time ago, but it seems like it has evolved a lot since then.
I’ve gotten used to ComfyUI, but I’ll give SwarmUI another try — it looks really interesting now.
Thanks to you, and don’t hesitate to let me know if you get good results; we can try to fix the issue together.
@taek75799 Thank you I'll do some more tests.
Hi, This is Great, fast and useful. Thanks for the checkpoint and the workflow.
Hello, and thank you for your feedback, have fun!
Amazing, thank you! Does this work with first frame last frame?
Yes, of course.
No FP16 ?
It seems I'm going to make the GGUF files for the BF16 version. Does that work for you?
@taek75799 Thanks for this model , very grateful been struggling to get camera prompt with lora and this is more than amazing.. anything of higher quality helps too for clients looking for high quality stuff.. although im good with the fp8 ..
@ApexArtist1 You're welcome. Have you tried the Q8 version? It's very similar to FP16 and much less resource-hungry
With a 5070 Ti and 16GB VRAM I'm aiming at the Q6 and Q4... not sure if i can get to the Q8.
I tried the V2 of taek75799 and they are working quite well for highly dynamic scenes.
Still struggling a bit with character consistency with Q4 when making multiple clip joined... so maybe I will try to find a way to make the Q8 work maybe is better.
@ai_phanotm_111517 Hello, are you using a workflow with the WAN Wrapper or the native version? I have a 4060 Ti with 16 GB of VRAM, same as you, and your card is more powerful.
I’m using the Q8 version: it’s slightly slower than the FP8 version (by a few seconds), and I can generate videos at 1024x1024. I haven’t tried increasing the resolution further yet.
@taek75799 I will share the workflow on my YouTube channel .. and I don’t use wrappers I hate it 😆
@ApexArtist1 Hello, no problem, do as you prefer 😉
@taek75799 the mystic lora is too strong the nipple is pink.. is there purely Cam control only version ? without the NSFW loras .. that be nice
@ApexArtist1 You can post your channel here if you want, so I can take a look. Maybe I already know it, and others will discover it.
@ApexArtist1 The camera version of my model. However, you can use it with other LoRAs, and they will work very well. The NSFW version is intended for users who want an all-in-one (AIO) model.
@ApexArtist1 You can still use it. The Mystic LoRA, its strength is not high, but it provides a real improvement.
@ApexArtist1 I found your channel, it seems very interesting. After just one video, I’m going to check out your content.
@taek75799 Faster card make it faster but for being able to run the main problem is the VRAM and we have the same ammount. So probably I can run the Q8 but than I will have less margin for extra lora or resolution or for frames.
With Q4 I can run with 121 Frames 10s clips and resolutions in the 832-1024 range with some space for 2 loras and also a good text encored.
I need to check the sweet spont also for the Q8 see how much i need to lower the video legth and resolution to make it run withour crashing.
@taek75799 yes "all-in-one (AIO)" models are super convenient, for consumer card if the lora is in the checkpoint you do not need to run it separately and it's all VRAM spared. Really appreciated your effort in this direction.
This is the best custom Wan2.2 model I've seen. I tried several versions and all worked perfectly. I am waiting for a new NSFW version. Thank you so much for your work.
Hello, and thanks for your comment. I’m glad you like the model. I’ll upload the NSFW version, and if you want to go even further with movements and camera work, feel free to check out the workflow I posted. If you don’t want to bother with my workflow, still take a look at the page of my WF on Civitai: there’s a link to a very useful node for that. You just have to replace the node with another one it’s very simple.