You can click on the link below and try it out directly. If the effect is good, you can deploy it locally.
Online Workflow: Wan2.2 Remix i2v 3.0 Image-to-Video Automatic Reverse
Experience link: https://www.runninghub.ai/post/2033536710499373058/?inviteCode=rh-v1325
Online Workflow: Wan2.2 Remix I2V v2.1 SVI2Pro Fully Automatic Reverse Long Video
Experience link: https://www.runninghub.ai/post/2008833586203463682/?inviteCode=rh-v1325
Online Workflow: Wan2.2-Remix-T2V-v2.0 text-to-video workflow
Experience here: https://www.runninghub.ai/post/1991533843182264322/?inviteCode=rh-v1325
Online Workflow: Wan2.2-Remix-I2V-Comfy-Qwen3
Experience here: https://www.runninghub.ai/post/1986632318448267265/?inviteCode=rh-v1325
Now! Register to get 1000 points, daily login 100 points, play 4090! Experience the super power of 48G!
[B-site video]: https://www.bilibili.com/video/BV1hF2cBNECZ
[Hugging Face: Wan2.2-Remix]:https://huggingface.co/FX-FeiHou/wan2.2-remix
[Telegram Group]: https://t.me/wan22remix
The workflow is very simple to use, and I have uploaded a tutorial on how to use it on YouTube. You can try it out and follow the tutorial and Civitai's workflow to easily reproduce the results. Thank you. If you have any other questions, please leave a comment in the comment section
Model Overview — Wan2.2-Remix
Wan2.2-Remix is a research and creative model built to generate short, cinematic video clips directly from text prompts. It focuses on human dynamics, realistic motion, and scene consistency, producing lifelike movement and smooth transitions without requiring LoRA setup.
This model already includes NSFW capabilities, enabling the generation of high-quality NSFW-themed videos without the need for additional LoRA.
High-noise model T2V: Lightx2v Wan2.2 Lightning dyno version (quantized to fp8)
High-noise model I2V: Comfy-Org Wan 2.2 ComfyUI Repackaged (fp16, quantized to fp8)
Low-noise model: Comfy-Org Wan 2.2 ComfyUI Repackaged (quantized to fp8)
Enhancements: Blended with open-source motion LoRA data and refined pose training for enhanced realism, anatomical precision, and gesture diversity.
Integration: Fully compatible with
lightx2v, no extra setup required.Recommended CLIP: NSFW-Wan-UMT5-XXL
Recommended Settings:
Steps: 4–8
CFG: 1
Shift: 5–10
Sampler: Euler
Scheduler: Simple
This release is currently beta. While functional, it is not yet perfect—feedback and testing are highly appreciated.
The GGUF version of the Remix model is now available.
Thanks to @BigDannyPt for creating it!
Download link:Wan2.2-Remix I2V GGUF v2.0
Changelog
I2V v3.0
Added VBVR LoRA to enhance spatial understanding and improve scene consistency.
Introduced UnifiedReward optimization to improve overall video quality and generation stability.
Adjusted several NSFW LoRAs to improve quality and visual consistency in NSFW scenes.
This version is considered the final release of the Remix I2V series.Unless there are major technological breakthroughs in the future, no further updates are planned.
I2V 2.1
Enhance facial consistency in long videos
Adjust LoRA ratio to make video dynamics more realistic
The v2.1 version has reduced the weight of the NSFW in the low-noise model to enhance character consistency. If you want to get a better NSFW, you can mix and match the low noise model of version 2.0.
T2V 2.0
Optimized character stability in generated videos, reducing jitter and unwanted motion artifacts
Enhanced video detail for more visually rich scenes
Added dynamic motion effects to improve realism and cinematic impact
Updated base model to original Wan2.2 version to minimize differences across random seeds
I2V 2.0
Significantly enhanced character motion and poses for more dynamic animations
Greatly improved camera movement and cinematic framing
Optimized NSFW content weighting to reduce jitter in generated videos when producing SFW scenes
Enhanced explosions and particle effects for more impactful visuals
SFW version will no longer receive updates
I2V 0.8
Updated lighting acceleration module to Wan2.1 version, fixing slow-motion character issues
Integrated over 20 open-source LoRAs and 10 custom-trained LoRAs to enhance motion dynamics
Removed several low-quality LoRAs to improve overall visual fidelity
T2V 1.0
Reduced dyno weighting to improve stability across different random seeds
Added more Asian facial datasets for greater diversity
Enhanced character motion dynamics for fluid animation
Improved motion amplitude, gesture diversity, and overall visual quality
T2V 0.9
High-noise T2V model: Lightx2v Wan2.2 Lightning dyno version (fp8)
High-noise I2V model: Comfy-Org Wan 2.2 ComfyUI Repackaged (fp16, fp8 quantized)
Low-noise model: Comfy-Org Wan 2.2 ComfyUI Repackaged
Refined character motion blending for smoother transitions
Enhanced anatomical articulation and natural body flow
This model incorporates the following Loras. Thank you to the Lora authors for their work. In order to facilitate management, I will change the file name uniformly after downloading Loras, so I cannot correspond to their real names one by one. Please understand. If there is any infringement, please contact me and the next version will remove it.
Wan2.1 - lightx2v_I2V_14B_480p_cfg_step_distill_rank256_bf16
Wan2.2 - I2V_A14b_high_noise_lora_rank64_lightx2v_4step_1022
Wan2.2 - I2V_A14b_low_noise_lora_rank64_lightx2v_4step_1022
Wan2.2 - Fun-A14B-InP-HIGH-MPS_resized_dynamic_avg_rank_21_bf16
Wan2.2 - Fun-A14B-nP-LOW-HPS2.1 resized dynamic avg rank 15 bf16
Wan2.1 - Skyreels-v2-T2V-720P LoRA rank64 fp16
Wan2.2 - PusaV1_lora_HIGH_resized_dynamic_avg_rank_98_bf16
Wan2.2 - PusaV1_lora_LOW_resized_dynamic_avg_rank_98_bf16
Wan2.1-T2V-BBFW - Better Buttholes For WANv1
Wan2.1-12V-General NSFW-Low
Wan2.2-12V-Hip Slammin Assertive Cowgirl
Wan2.2-12V-PENIS LORA e496
Wan2.2-12V-DR34ML4Y -All-In-One NSFW
Wan2.2-T2V-Body Shots v2 14B
Wan2.2-12V-Deepthroat, Blowjob
Wan2.2-T2V-Masturbating Rubbing 14B
Wan2.2-T2V-Doggy Style-14B -v3
Wan2.2-12V-COWGIRL+REVERSE COWGIRL-14b
Wan2.2-12V-Combo Handjob Blowiobsafetensors
Wan2.2-Bounce-Test
Wan2.2-T2V-GodPussy
Wan2.2-T2V-MysticXXX
My Train:
P001-Side-Lying-Sex-Wan2.2-i2v v10
P003-Sexy-Enhancer-Wan2.2-i2v
P004-Lying-Bouncing-Wan2.2-i2v v20
P005-Chinese-Dance-i2v-v20
P006-Squirting-Wan2.2-i2v
P008-Running-Wan2.2-i2v v10
P009-The-Walk-Wan2.2-i2v-Light
P010-whip-t2v-v10
F006-AsianGirls-Wan2.2-t2v v10Disclaimer
This model is released for research and creative purposes only. It must not be used for illegal, unethical, or harmful activities. The author assumes no responsibility for any misuse or consequences arising from its use.
Description
FAQ
Comments (79)
Do we need to use specific trigger words from the list of included LORAs? For example, BBFW - better but holes for WAN. If I was loading that LORA separately I would need to use it. Not sure how this works with the remix.
No specific trigger words are needed, just write prompt words in normal natural language
I get an error with this clip: Recommended CLIP: NSFW-Wan-UMT5-XXL
There are two versions, FP8 and BF16, give them a try
both clipfile versions are faulty somehow... bf16 throws clip error and fp8 throws clip missing.
Was using your workload for image 2 video, but its not letting me input anything in positive prompt. I got all custom nodes and models. Not sure what I'm doing wrong
Just disconnect the line of the prompt input box that is connected to the Qwen3 node.
@FX_FeiHou Thanks! That worked
This is amazing! I'm noticing that when video comes close to ending the characters revert back to positions of original image. Is there a way to prevent that?
@Cheeze288 length<=81?
@qek I had length>81.
@Cheeze288 Too long, it may be the reason
@qek I had framerate for video above 16 at 26. Apparently higher frame rates make video last shorter, just took screenshot at end of vid and started new one
@Cheeze288 "higher frame rates make video last shorter" Interpolate frames instead
@qek Can you tell me how to do this? I have the interpolate custom node but I do not know how to apply it or find it in comfyui. I was using FX_ FeiHou workflow
@qek LOL I figured it out thanks
I love that you're doing T2V when it's being over looked by so many I2V models. Thank you! Keep up the amazing work <3
Edit. The female anatomy is kind of "odd". Noticing some weird things. With the men too 😢. Let me know if there's some way I can help to improve the next model. Maybe there's a way to hot fix 🤷
The quality is good but , maybe not trained on enough models. Each prompt produces the same looking models with slight hair changes, maybe bigger boobs, but same face each time, it can't do darker skin tones or smaller breasts, every woman has the same body proportions.
when you say "steps 4-8" do you mean total steps between 4 and 8, or 4 high steps and 8 low steps? thanks
No, steps in total
Does this also apply to txt2vid or only to img2vid?
Your model, with or without your workflow, completely ruins the character's face
The best Wan checkpoint in planet, at moment🔥🔥🔥 👍
* The I2V i mean...i don't work with T2V...
The T2V 2.0 version causes my character to look rather cartoonish and less lifelike. I didn't observe this issue with the previous version.
Is there a way to you make a workflow that doesn't use "sageattention"? No matter what I do I get a KSamplerAdvanced popup error saying "No module named 'sageattention'". Also can't enter input text in the I2V workflow, thanks!
Install it then
search how to install it here, on civitai, therewas a page tutorial about that
you can also turn it off and everything els still run. i think it cntrl + B
Love it, best results so far. The big D's can use some work though, very unrealistic compared to new lora's
I am running into some pretty basic issues with the the T2V models. My prompt mentions a girl with small breasts and black man with a big cock, and I am getting a girl with huge tits and a white man.
I am not using any LoRAs. I am using the workflows shared on google drive.
Version 2.0 of t2v cannot correctly generate men with penises on their own, and the men's skin is also poorly drawn. Version 1.0 did not have these problems.
I2v, the eyes is closed most of the time
t2v: Should I try increasing low or right steps first for prompts that go into a lot of detail?
Try various ratios and total step counts
And where to download Wan2.2 Remix NSFW GGUF ? Are there NSFW GGUF MODELS?
@qek this is understandable, but it does not indicate that these are models N,S,F,W, so the question is are these models N,S,F,W, or S,F,W, ? and by the way, I downloaded gguf models from this repository and did the generation without LORA and it turned out well, as I understand it, it is possible to generate without LORA?
@soyv4 What is the LORA?
@soyv4 they're nsfw - i used em - seem better with motion
My I2V are coming out extremely blurry, anyone have a modified workflow I can look at to see what changes I can possibly make?
do you have low and high reversed? i just did that on accident lol
@markdalias yes, start with High, refine with Low
The BF16 of that NSFW-Wan-UMT5-XXL
has been broken for months, are there alternatives? Would be nice to use a full bf16 for better accuracy but i can't find other sources online and the creator totally went ghost.
@qek sweet thanks
@qek sweet thanks
@sneedingonmyligma420 you're welcome
@sneedingonmyligma420 you're welcome
This model is perfect for nsfw needs, its the only one i've used that nailed dildo riding prompts and undressing prompts with no fuss. I could never use that t2i dildo riding lora everyone seemed to use for that kind of prompt, girls would always half or partially ride it. With this model, they fully go down on that thang.
hi. I've encountered a completely opposite situation. The dildos are not realistic, they're not where they should be, and they create feet. Can I ask for an example of the prompt you're using?
P.S
sorry for the crooked Google translation.
can anyone tell me what is the best current model to use to generate nsfw images? like also image to image ( like turning my model from clothes on to clothes off)?
Flux.1 Kontext with a lora
qwen image edit 2509 with lora
@jokim0820 I just looked at all the LORA specialists on NSFW and I'm not impressed. He draws breasts poorly, and as for the genitals, they're a plastic surgeon's dream. So far, I haven't found anything better than a pony. But there are other problems there.
@qek Qwen Rapid, I use 23 ver. https://huggingface.co/Phr00t/Qwen-Image-Edit-Rapid-AIO
unusable with lora
wow amzing
Is anyone else having issues for example a doggy position but the penis inserts to the side or doesn't align? Lots of insertions that go straight into the cheeks.
I keep getting the error ":> PathchSageAttentionKJ No module named 'sageattention'" even though I installed the KJ nodes with no issue, help?
Remove the node
yes, ctrl + b maybe you are on amd graphics card
If you are using comfyui(setup = not portable) version, I don't know how to solve it, okay?
And 64GB of regular RAM I assume? lol
Noticed that in i2V "first-last frame" workflow, last image is saved in a very bad quality. Much worse then original expected. Looks like model does not understand properly, how to work with start-end workflow. A bug?
Using the Comfy workflow, the final frame effect is acceptable without adding Lora,
Try my workflow: Wan2.2- Remix: First List Frame Loop Wallpaper Workflow
https://www.runninghub.ai/post/1998577590703128577/?inviteCode=rh-v1325
This is a pretty good model and does do a lot of things right. My only issues are that it does seem to degrade quality faster than the vanilla Wan models (due to all of the Loras?) and it also makes all of the woman incredibly horny - they're always grinding their hips/etc even when you don't want them to!
nsfw_wan_umt5-xxl_bf16.safetensors
是放在text_encoders文件夹吗？不知道为什么运行的时候显示 “加载CLIP invalid tokenizer”，求解
我也不能用，奇怪了
Thanks, fixed version,It works.
@qekHello, the link for version FP16 is broken. Do you still have the model? I really need it.
thank you for sharing your work, what is your go to pipeline right now for generating long, consistent videos with Wan 2.2 using the Remix models?
Woaaahhh it's awesome!
I am using your workflow, with your models (T2V), but my output images looks like they are made of plastic (cartoonish). No photo-realism, depth of details etc. What am I doing wrong ?
Details
Files
wan22RemixT2VI2V_t2vLowV20.safetensors
Mirrors
Wan2.2_Remix_NSFW_t2v_14b_low_lighting_v2.0.safetensors
Wan2.2_Remix_NSFW_t2v_14b_low_lighting_v2.0.safetensors
Wan2.2_Remix_NSFW_t2v_14b_low_lighting_v2.0.safetensors
Wan2.2_Remix_NSFW_t2v_14b_low_lighting_v2.0.safetensors
Wan2.2_Remix_NSFW_t2v_14b_low_lighting_v2.0.safetensors
Wan2.2_Remix_NSFW_t2v_14b_low_lighting_v2.0.safetensors
Wan2.2_Remix_NSFW_t2v_14b_low_lighting_v2.0.safetensors
Wan2.2_Remix_NSFW_t2v_14b_low_lighting_v2.0.safetensors
Wan2.2_Remix_NSFW_t2v_14b_low_lighting_v2.0.safetensors
Wan2.2_Remix_NSFW_t2v_14b_low_lighting_v2.0.safetensors
Wan2.2_Remix_NSFW_t2v_14b_low_lighting_v2.0.safetensors
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.