⚠️This LORA is for WAN2.2 I2V video generation!!⚠️
The sample PNG image is meta information for video generation. Drag and drop it into ComfyUI to view the workflow and resources.
📢I have released new collaboration model.
Onaho Figure POV SEX (WAN2.2 I2V 14B) Vomiting Cum Bukkake
⚠️Do not refer to LoRA:[Vominting Cum Bukkake] simultaneously with this model.
[Vominting Cum Bukkake] is a LoRA model for video production that starts with the final frame image of a video created with this model.
Let's Sex ! with Your Real PCV Wife Figure!!
Please specify a full-body photo of the PCV Figure as the starting image.
🦋If you like images of fairies like Tinkerbell, feel free to use them.
💡If it negatively affects the facial features, reduce the strength of the LOW model.
⚠️This model is optimized for toy figures.
training dataset and config is here.
https://mega.nz/folder/Iw5RhBhY#a4L83fBzT7oYJtFKkS1smA
🎬WAN2.2 I2V Video Prompt
POV SEX,
A toy figurine is placed there.
A first-person hand grabs the toy figurine.
A first-person hand spreads the figurine's legs.
A first-person perspective hand deeply inserts its penis into the figure's vagina.
The figure's belly swells greatly.■sorry Japanese text■
本モデルはプラスチックフィギュアとのＰＯＶセックスに最適化してトレーニングしています。
プラスチックフィギュア以外でのご使用は・・・あなたのアイデア次第です。
動画投稿の際にはCivitaiの規約厳守をおねがいします。
[Onaho Figure Bukkake SEX / IL]
このILのモデル大好きだったんですよ。個人的に滅茶苦茶使いました。
WAN2.2のモデル作成技術も習熟してきました、作ってみましたWAN2.2モデル。
結果としては、ありとあらゆるフィギュア画像から理想のPOVセックスに展開できます。
フィギュアPOVセックス動画の起点動画になるので、最終コマを次のスタート画像に指定して、長編動画をいくらでも生成できるでしょう。
やったぜ❤
今後とも、よろしくお願いします。
モデル作成には、AIモデル作成支援ソフト『HENTAI APP』を利用しています。
詳細は以下記事を参照：https://civarchive.com/articles/26629/create-wan22-datasets-from-videos-optimized-software-has-been-released
📘Article from the WAN2.2 IV2 Training Guide
https://civarchive.com/articles/27705
私は本LORAの利用に、制限・権利・責任を一切主張しません。
Description
FAQ
Comments (15)
......!I'm going right now to go retrieve the PVZ Peashooter and Chomper models I left out in the backyard.
I want to train my own LoRA.I'd like to know how to create a video clip like this.
I know how to make a jump cut LoRA and can handle that kind of footage, but for coherent motion like in this LoRA, I don't know how to obtain the training data.
@qiaoy123
Thanks for your comment!
The training video for this model was created using WAN2.2.
The training video clip was created by specifying the start and end images, and referencing appropriate prompts and multiple LoRa signals.
If a creative clip doesn't exist in existing videos, you'll have to create it yourself.
However, with WAN2.2, creating clips is incredibly easy...though it might take a little time.
You can access the video clips and configuration files used for training by clicking on the dataset links in the article.
@suteakasuteakasuteka434 Thank you！
@suteakasuteakasuteka434 I've checked your training data, so I have a rough idea of what your dataset looks like. However, I'm still unsure how to obtain a video like that using the method described in your article.
The problem is that most of the footage we collect is extremely slow. Even speeding it up doesn't help. If we only pick a few frames, we end up with jump cut material.
Personally, I usually collect two images of the same character from image sites, set one as the starting frame, and generate a video from the other using I2V, then splice them together. I have tried the start-and-end frame method, but to be honest, the results were terrible—the transition was distorted. I still don't know how to acquire such smooth video footage like yours.
@qiaoy123
When specifying start and end images in WAN2.2, the background and character's clothing must match between the two images.
In conclusion, the start and end images are created using StableDiffusion. Images are combined to create a video that achieves smooth animation.
The video file and the workflow PNG for video generation are stored below.
You can check the video creation status by dragging and dropping the workflow PNG into ComfyUI.
I plan on doing terrible things with this
What a coincidence! I also do all sorts of things on my local PC that I can't even describe in writing. And those things absolutely cannot be shared on the internet. 🫢
great model thanks
i'm so bored of AI that now i want one of this silicone thing for real. Hard to find one wtf
Bro opened pandora's box
Oh, if I have instilled in you a fetish for RX-78 Gundam figures, I sincerely apologize.
Im always wanted to do that with GRENDIZER
Yikes!