Wan 2.2 I2V Toejob/Footjob CUM LoRA - t0e
My second Wan 2.2 Image-to-Video (I2V) LoRA, showing promising results for NSFW workflows. I haven't had much change to test it properly so post away!
Overview
This LoRA is designed to be used in the second part of a full NSFW workflow for merging two videos - one video starts the NSFW act, and this LoRA helps with the finishing cum effect. The main objective of the Lora was for a penis cumming DURING a toejob/footjob.
used in conjunction with the footjob lora in the first scene to create the main first frame position.
Strength: 0.4 - 1.0 (Lower strength for more exaggerated effect and better prompt adherence)
Training Details
- Dataset: 52 clips (5-10 seconds each)
- Resolution: 320 x 192
- Content Mix: 90% realistic + 10% anime
- Anime addition helps with exaggerated action and fluid amount
Prompts
t03, The penis ejaculates lots of thick white cum as she strokes the penis with her toes and feet. Thick white cum is pouring out of his penis and onto her toes and feet. Lots of thick transparent white cum flows out as her toes and feet stroke the penis.
Notes
- All examples done with LightxV2 Speedups
Support
If you like what I'm producing and would like to support me and help me make more awesome Wan 2.2 I2V Lora's, a coffee can go a long way :). Also will do custom requests, just hit me up.
Thanks!
Description
FAQ
Comments (14)
this lora looks great, it just sucks that there isn't any good realistic checkpoint to gen good footjob pics other than anime...
Yeah that's why I use the Footjob I2V to create the first video aka position and use the last frame to begin the second video with this lora. Problem is the first video tends to leave the face out of focus.
There are, SDXL models with Illustrious LoRAs is the combo you're looking for...
@yedeb63824620 I tested many pony and illustrious lora on SDXL models none of them work or it just give really bad output that can't even inpaint it.
What SDXL models did you used with illustrious lora?
@bielzimdag346 Mainly mixes of BigAsp2 / BigLove or Lustify
Yo go to porn sites and use an extension on your browser and download videos of compilations. Use VLC to pause and snapshot frames. If they don't have a face use Flux Fill out-painting. If you have loras of specific people then you can use those but you will likely have to use a Lora to prevent flux messing with the genitals, lol. If you want to get fancy though download HYPIC from Bytedance its the best little photo manipulation app I promise and you can cut out the people and place them in new locations or something. It's mostly free and no ads. If you need loras go to Shakker.ai and you get 3 free ones. Want more? Just switch google accounts they are not that sophisticated telling who's who. If you have shorter clips and want every frame super clean with no blur AI can write you a python script to use CV2 and it will grab every frame in the video and lay them out as perfect images. Or I can just give it to you right now. You just have to change the name of the video to xme.mp4 and run this but (WARNING:be careful if its over 10 seconds like a whole movie get ready for a hundred thousand pictures as this is not sophisticated with a stop button its just going to do what its told so I recommend short clips only. Do the math 16-60 FPS x how many seconds = a lot of fucking pictures)
copy the code in Notepad++ and save it as a python file and name it something simple like xaf.py - place this inside your ComfyUI folder as well as the video renamed xme.mp4 - It should be the same place as your .venv You can open the terminal in Comfy and type what is inside the arrows only and as long as you named your mp4 xme.mp4 it will create the folder and rip it into pictures. If you don't feel comfortable with all this just ignore I ever said it.
--> Python xaf.py install <--
copy below:
import cv2
# Open the video file
video_path = 'xme.mp4'
cap = cv2.VideoCapture(video_path)
# Check if the video opened successfully
if not cap.isOpened():
print("Error: Could not open video.")
exit()
# Get the frame rate of the video
fps = cap.get(cv2.CAP_PROP_FPS)
print(f"Frame rate: {fps} FPS")
# Create a directory to save the frames
import os
output_dir = 'extracted_frames'
os.makedirs(output_dir, exist_ok=True)
frame_count = 0
# Read and save each frame
while True:
ret, frame = cap.read()
if not ret:
break
# Save the frame as an image file
frame_filename = os.path.join(output_dir, f'frame_{frame_count:04d}.png')
cv2.imwrite(frame_filename, frame)
frame_count += 1
# Release the video capture object
cap.release()
print(f"Extracted {frame_count} frames.")
About to give this lora ago, ill post my results once they have been completed. Thank you for the hard work youve put into making this.
Yes please! I want to add a more exaggerated anime effect but was actually struggling to find good samples for the dataset but looking forward to seeing your results. Lower strength might help more with flexibility and prompt adherence
This looks fantastic! Can we please get a T2V version, as well?
Can we have the workflow with the first part?
I've got quite a large workflow for many different starting. Ill have to clean it up but will work on it doing that when I have time
@aivantgarde Thanks, looking forward to it
we need a no cum version of this, since there are no such good FJ loras for the first part of the video...
This lora also generates realistic cumshots else than footjob scenes, i use it for handjob cumshots and works brilliant.
Details
Files
FTJ-LOW-EPOCH-100.safetensors
Mirrors
FTJ-LOW-EPOCH-100.safetensors
FTJ-LOW-EPOCH-100.safetensors
FTJ-LOW-EPOCH-100.safetensors
FTJ-LOW-EPOCH-100.safetensors
FTJ-LOW-EPOCH-100.safetensors
wan2.2-i2v-low-footjobcum.safetensors
FTJ-LOW-足交射精-100.safetensors
FTJ-LOW-EPOCH-100.safetensors
FTJ-LOW-EPOCH-100.safetensors
FTJ-LOW-EPOCH-100.safetensors
FTJ-LOW-EPOCH-100.safetensors