My NSFW TG Channel - https://t.me/stefanfalkokmoth
My TG Channel - https://t.me/StefanFalkokAI
My TG Chat - https://t.me/+y4R5JybDZcFjMjFi
Thanks to creators for helping with some images - subscribe to see their cool NSFW works
Hi! I introduce my working workflow Wan 2.2 NSFW Workflow with 10 Steps I2V model + NSFW Loras (fp8 + gguf)
I have included 2 workflows for fp8/fp16 models and GGUF and MMAudio workflows
Recomend resolutions: from 576p to 720p
You need to have Wan 2.2 model 10 steps i2v high and low noise models (https://civarchive.com/models/2086218?modelVersionId=2360534), clip (https://huggingface.co/Comfy-Org/Wan_2.2_ComfyUI_Repackaged/blob/main/split_files/text_encoders/umt5_xxl_fp16.safetensors) and vae (wan 2.1 vae)
Also you need download SVI 2 PRO loras for long videos - https://huggingface.co/Kijai/WanVideo_comfy/tree/main/LoRAs/Stable-Video-Infinity/v2.0
(wan 2.2 experimental) WAN General NSFW model - https://civarchive.com/models/1307155?modelVersionId=2073605
WAN DR34ML4Y - All-In-One NSFW - https://civarchive.com/models/1811313?modelVersionId=2176505
Cunnilingus - Pussy licking Lora I2V - Wan2.2 - K3NK - https://civarchive.com/models/1954774/cunnilingus-pussy-licking-lora-i2v-wan22-k3nk
reverse suspended congress - https://civarchive.com/models/1894970?modelVersionId=2145089
Handjob (One/Two handed) - https://civarchive.com/models/2078995?modelVersionId=2352366
POV Paizuri / Titfuck - https://civarchive.com/models/2070335?modelVersionId=2342652
Oral Insertion [WAN 2.2] - https://civarchive.com/models/1874153?modelVersionId=2121297
Wan2.2 - Pussy (T2V/I2V) - https://civarchive.com/models/1895314/wan22-pussy-t2vi2v
Perfect Fingering - Wan 2.2 I2V - https://civarchive.com/models/1952032?modelVersionId=2209481
Cum/Facial - Wan2.2 - https://civarchive.com/models/1962545?modelVersionId=2460386
Deepthroat, Blowjob - Wan 2.X I2V & T2V - https://civarchive.com/models/1497390?modelVersionId=2152516
Ultimate DeepThroat I2V Wan2.2 Video LoRa - K3NK - https://civarchive.com/models/1874811?modelVersionId=2325788
WAN 2.2 I2V - Combo Handjob Blowjob - https://civarchive.com/models/1899045?modelVersionId=2149593
WAN 2.2 I2V - Handjob (One/Two handed) - https://civarchive.com/models/2078995?modelVersionId=2352366
WAN 2.2 Anime Cumshot Aesthetics – Precision Load I2V (Beta version) - https://civarchive.com/models/1869475?modelVersionId=2116008
WAN 2.2 I2V - POV Body Cumshot & Pullout - https://civarchive.com/models/2031069?modelVersionId=2298673
WAN 2.2 I2V - POV Double Blowjob - https://civarchive.com/models/1906148?modelVersionId=2164213
WAN 2.2 I2V - Teasing Sensual Blowjob - https://civarchive.com/models/2231076?modelVersionId=2511581
Wan I2V (2.2 & 2.1) - Assertive Cowgirl - https://civarchive.com/models/1566648?modelVersionId=2129122
Licking lips - I2V - Wan2.2 - https://civarchive.com/models/1978869?modelVersionId=2239947
Wan2.2 - French Kiss (I2V/T2V) - https://civarchive.com/models/2171333?modelVersionId=2445168
Wan Fingering Pussy - https://civarchive.com/models/2007360?modelVersionId=2272024
Anal insertion (I2V) - https://civarchive.com/models/2002532?modelVersionId=2266652
(Wan 2.2) Male finishing [Anime] - https://civarchive.com/models/1946997?modelVersionId=2579567
Deepthroat/Face Fuck- Wan2.2 I2V - https://civarchive.com/models/2023407?modelVersionId=2517513
WAN 2.2 I2V - Breast Play - https://civarchive.com/models/1967237/wan-22-i2v-breast-play
Wan 2.2 I2V Footjob - https://civarchive.com/models/1861926?modelVersionId=2107285
Wan 2.2 T2V Thigh Sex - https://civarchive.com/models/2020185?modelVersionId=2286441
WAN 2.2 I2V Mating Press - https://civarchive.com/models/2368333?modelVersionId=2663475
Tentacled - Wan2.2-14B - https://civarchive.com/models/1853189?modelVersionId=2543800
Wriggling motion LoRA (WAN 2.2 Release) [Tentacles!] - https://civarchive.com/models/1814871?modelVersionId=2161335
Wan 2.2 Fucking Machine / Sex Machine / Dildo Machine I2V - https://civarchive.com/models/2448294?modelVersionId=2752896
WAN COWGIRL + REVERSE COWGIRL -- T2V & I2V LoRa - https://civarchive.com/models/1428098/wan-cowgirl-reverse-cowgirl-t2v-and-i2v-lora?modelVersionId=2156435
Wan 2.2/2.1 POV Missionary - https://civarchive.com/models/1331682?modelVersionId=2098405
Hip sway/Ass shaking dance Wan 2.2 I2V - https://civarchive.com/models/2074265/hip-swayass-shaking-dance-wan-22-i2v?modelVersionId=2347127
I2V - Sex from behind (front facing) - https://civarchive.com/models/2227622?modelVersionId=2545249
Doggy slider T2V I2V Wan2.2 Video LoRa - https://civarchive.com/models/1987345?modelVersionId=2249683
Penis Lora (+Blowjob, +Cumshot) [Taz] - WAN 2.2 14b - https://civarchive.com/models/1476909?modelVersionId=2267675
Wan Multiscene Photoshoot: Softcore Edition - https://civarchive.com/models/2316517/wan-multiscene-photoshoot-softcore-edition?modelVersionId=2606755
Wan 2.2 Spoon by MQ Lab - https://civarchive.com/models/1916930?modelVersionId=2169632
Wan 2.2 Massage tits by MQ Lab - https://civarchive.com/models/1952945/wan-22-massage-tits-by-mq-lab?modelVersionId=2210306
Wan 2.2 Doggy by MQ Lab - https://civarchive.com/models/1923354/wan-22-doggy-by-mq-lab?modelVersionId=2176892
Wan 2.2 X-ray NSFW [Anime] --I2V,T2V - https://civarchive.com/models/2079276/wan-22-x-ray-nsfw-anime-i2vt2v?modelVersionId=2631919
WAN 2.2 I2V - Face Down Ass Up - https://civarchive.com/models/1929094/wan-22-i2v-face-down-ass-up?modelVersionId=2183383
Wan 2.2 Fisting | Vagina & Anus - https://civarchive.com/models/2131482/wan-22-fisting-or-vagina-and-anus?modelVersionId=2411126
Wan 2.2 I2V Foot Worship / Toe Sucking - https://civarchive.com/models/2003375/wan-22-i2v-foot-worship-toe-sucking?modelVersionId=2267620
Footjob I2V Wan 2.2 - https://civarchive.com/models/2105640?modelVersionId=2382169
WAN 2.2 I2V Sucking Foot - https://civarchive.com/models/2328449/wan-22-i2v-sucking-foot
Breast Insertion [WAN 2.2] - https://civarchive.com/models/1972311/breast-insertion-wan-22
Groping/massage/Show off (fake)breasts - https://civarchive.com/models/1790891?modelVersionId=2605575
Jerking Off - https://civarchive.com/models/2224862/jerking-off?modelVersionId=2603468
Finger-Slap-Spread Pussy - https://civarchive.com/models/2279512/finger-slap-spread-pussy?modelVersionId=2565600
Slap (and Self Slap) - Wan 2.2 I2V / T2V - https://civarchive.com/models/2062140/slap-and-self-slap-wan-22-i2v-t2v?modelVersionId=2509620
WAN 2.2 I2V - Pegging - https://civarchive.com/models/2410338/wan-22-i2v-pegging?modelVersionId=2710056
Squatting Froggy - I2V - WAN 2.2 - https://civarchive.com/models/2332227?modelVersionId=2623479
Wan Dildo Classic - https://civarchive.com/models/1707303/wan-dildo-classic?modelVersionId=2405949
The Walk - Wan 2.2/2.1 - I2V & T2V - https://civarchive.com/models/1363473/the-walk-wan-2221-i2v-and-t2v?modelVersionId=2254373
PLEASE READ TERMS OF USE FOR EVERY LORA IN LORA'S PAGE LINK!!!
download my ComfyUI Build https://huggingface.co/datasets/StefanFalkok/ComfyUI_portable_torch_2.11.0_cu130_cp313_sageattention_triton - my new ComfyUI build with torch 2.11.0cu130 + sageattention + my workflows, also u need download and install CUDA 13.0 (https://developer.nvidia.com/cuda-13-0-0-download-archive) and VS Code (https://visualstudio.microsoft.com/downloads/)
Leave comments if you have trouble or you found the problem with workflows, also leave me comment or dm message if you see the violations
Description
Included workflow with SVI 2 Pro long video technique and many links with NSFW lora for Wan 2.2. Also put fixed file for MMAudio
FAQ
Comments (31)
Hello,
I’ve just downloaded your project and I’m starting to download the required files.
I noticed that the description does not mention the z_image_turbo_bf16.safetensors file. However, in the code it appears to be used in UNETLoaderDistTorch2MultiGPU.
Is this file actually required? What exactly is it used for?
Hello. I just put an empty UnetLoaderDisTorch2MultiGPU. All that's left for you is choose a correct wan models for you - my 10 steps, Dasiwa and another Wan models
@Stefan_Falkok Undestand. I found the work incredible. I ran some tests and they showed good results. However, it seems that I’m encountering an anti-NSFW issue. When using the same LoRA and the same prompt that was already configured, adding only “a man appears” and performing an action (since the image initially didn’t include him), he doesn’t appear. When trying to remove clothes, for example, I feel some resistance. Of course, this might be because I’m using RunPod and not the local version provided in the description. In any case, my goal is to generate a woman wearing clothes who then removes them and performs some scenes, but as I mentioned, I had some difficulty achieving that. I tried using the gguf that is already saved and also tested it with dasiwan.
@Valorizando understand that problem. Try with another loras which I specified in the workflow and search in Civitai - it can help you
that part is broken, nothing more to install https://imgur.com/a/txWKwTn
Check missing custom nodes via manager and install missing nodes (it's MultiGPU and GGUF nodes if I'm right)
@Stefan_Falkok well i changed my browser and worked,thanks, the result at first time without many config with my loras are very good
Hey, for some reaason whenever i click on run to run the first batch it just skips to output and saays this in the logs:
"got prompt
Failed to validate prompt for output 102:
* KSamplerAdvanced 84:
- Return type mismatch between linked nodes: scheduler, received_type(['simple', 'sgm_uniform', 'karras', 'exponential', 'ddim_uniform', 'beta', 'normal', 'linear_quadratic', 'kl_optimal', 'bong_tangent']) mismatch input_type(['simple', 'sgm_uniform', 'karras', 'exponential', 'ddim_uniform', 'beta', 'normal', 'linear_quadratic', 'kl_optimal', 'bong_tangent', 'beta57'])
* KSamplerAdvanced 85:
- Return type mismatch between linked nodes: scheduler, received_type(['simple', 'sgm_uniform', 'karras', 'exponential', 'ddim_uniform', 'beta', 'normal', 'linear_quadratic', 'kl_optimal', 'bong_tangent']) mismatch input_type(['simple', 'sgm_uniform', 'karras', 'exponential', 'ddim_uniform', 'beta', 'normal', 'linear_quadratic', 'kl_optimal', 'bong_tangent', 'beta57'])
Output will be ignored
Failed to validate prompt for output 357:
* KSamplerAdvanced 108:
- Return type mismatch between linked nodes: scheduler, received_type(['simple', 'sgm_uniform', 'karras', 'exponential', 'ddim_uniform', 'beta', 'normal', 'linear_quadratic', 'kl_optimal', 'bong_tangent']) mismatch input_type(['simple', 'sgm_uniform', 'karras', 'exponential', 'ddim_uniform', 'beta', 'normal', 'linear_quadratic', 'kl_optimal', 'bong_tangent', 'beta57'])
* KSamplerAdvanced 118:
- Return type mismatch between linked nodes: scheduler, received_type(['simple', 'sgm_uniform', 'karras', 'exponential', 'ddim_uniform', 'beta', 'normal', 'linear_quadratic', 'kl_optimal', 'bong_tangent']) mismatch input_type(['simple', 'sgm_uniform', 'karras', 'exponential', 'ddim_uniform', 'beta', 'normal', 'linear_quadratic', 'kl_optimal', 'bong_tangent', 'beta57'])
Output will be ignored
Failed to validate prompt for output 124:
Output will be ignored
Prompt executed in 0.02 seconds"
im trying to solve it with grok now i have been trying for hours i really am struggling with it
update comfyui and custom nodes or install my new comfyui build
@Stefan_Falkok i have clicked update all today more times than i have used comfyui itself, i will download your build then, but im going to be honest im kind of a beginner so i dont know how to download your build and the cuda and all of these stuff since i use comfyui web and a virtual gpu
and honestly i solved the issue above by disconnecting the scheduler and making it a dropdown, but then everytime i press on run a new issue appears, i have been on this shit for about 4 damn hours
@marcelinorafik01419 just download my ComfyUI build - I put all necessary main loras you need and configured it to work without any problems
@Stefan_Falkok okay thanks a lot, but is there a specific place i have to download it in? do i also downloaded on jupyter? or on my own laptop
@marcelinorafik01419 you can place in any path you want
@Stefan_Falkok okay thanks a lot, i apologize for asking these kind of questions, but how would i import it onto the comfyui web, if it was a python file it would be different but i see all kinds of files so im honestly confused
@Stefan_Falkok i think i got it now, ill update you, thanks a lot btw
Hey stefan, i managed to download it, but honestly i dont know how to add it to vast to use it on vast.ai to use a higher gpu than what i have, because what i have is bad!!
@marcelinorafik01419 I can't help with it because I'm using my PC for local generation
@Stefan_Falkok damn okay thanks a lot for your help and your build G
@Stefan_Falkok sorry one more question, is it okay if i just download the custom nodes in the default comfyui instead of your custom one? would that fix all the issues with the ksampler advanced (scheduler) and the vae errors and etc?
@marcelinorafik01419 that's okay if you download missing custom nodes
Hey Stefan, my bad, been asking a lot of questions, but i ran into this error:
RuntimeError: mat1 and mat2 shapes cannot be multiplied (154x768 and 4096x5120)
im trying to fix it with gemini or grok but they are telling me to remove the cliptext encoder and replace it with wan text encoder, but i honestly dont want to do anything that could ruin the workflow.
This workflow appears when it starts running the first ksampler advanced
@Stefan_Falkok that worked thanks a lot mannn!! im really hoping i dont run into any other issuesss! 😫😫
@Stefan_Falkok hey, for some reason this was the result (NSFW CONTENT IN THE LINK), https://streamable.com/hbiubj.
IS IT A LORA ISSUE? these are my loras: <lora:NSFW-22-H-e8:0.90:0.90> <lora:NSFW-22-H-e8:0.90:0.90> <lora:Elara (Lora):0.10:0.10> <lora:Sensual_fingering_v1_low_noise:1.00> <lora:wan22-i2v-rub-pussy-os-high:1.00:1.00>
the S3nsu4l f1ng3r1ng is turned off^^
just wanted to let you know, i asked grok ai, he said i should change the steps to 12-16 instead of 8 and change the scheduler and the sampler
@marcelinorafik01419 llms give you a wrong answers. Set 8 steps and download my 10 steps wan models
@Stefan_Falkok the fp8_12Vhighnoise and the fp8_12vlownoise?
@Stefan_Falkok it works now, but i think this is actually the last and final issue, i appologize, bare with me please, https://drive.google.com/file/d/1bKKUBm_4lspLsDhxt7v9Z3oZgUOsKr2s/view?usp=sharing
this is the video that i generated after doing what you told me to, but for some reason the animation is very weird and bad, is it a prompt issue? or lora issue?
@marcelinorafik01419 it can be an issue with prompt and lora at the same time. I recomend writing prompts better
@Stefan_Falkok Hey Stefan, i thought this was solved but apparently it isnt solved, im pretty sure if we solve this i wont have anymore issues. I keep on getting this error:
INFO | - | got prompt
18:00:47 | ERROR | - | Failed to validate prompt for output 357:
18:00:47 | ERROR | - | * KSamplerAdvanced 84:
18:00:47 | ERROR | - | - Return type mismatch between linked nodes: scheduler, received_type(['simple', 'sgm_uniform', 'karras', 'exponential', 'ddim_uniform', 'beta', 'normal', 'linear_quadratic', 'kl_optimal', 'bong_tangent']) mismatch input_type(['simple', 'sgm_uniform', 'karras', 'exponential', 'ddim_uniform', 'beta', 'normal', 'linear_quadratic', 'kl_optimal', 'bong_tangent', 'beta57'])
18:00:47 | ERROR | - | * KSamplerAdvanced 85:
18:00:47 | ERROR | - | - Return type mismatch between linked nodes: scheduler, received_type(['simple', 'sgm_uniform', 'karras', 'exponential', 'ddim_uniform', 'beta', 'normal', 'linear_quadratic', 'kl_optimal', 'bong_tangent']) mismatch input_type(['simple', 'sgm_uniform', 'karras', 'exponential', 'ddim_uniform', 'beta', 'normal', 'linear_quadratic', 'kl_optimal', 'bong_tangent', 'beta57'])
18:00:47 | ERROR | - | Output will be ignored
18:00:47 | ERROR | - | Failed to validate prompt for output 102:
18:00:47 | ERROR | - | Output will be ignored
18:00:49 | INFO | - | Prompt executed in 1.89 seconds
I duplicated your workflow and change ksampler advanced to the comfyui's original one and made the schedular a dropdown on the ksampler advanced and i dont get this error, but on the original workflow i get this error, I dont know why i got this error again even though yesterday i could run the workflow normally