12/06 UPDATE - Futanaris and Males!

Added Loras to make males genitals and futanaris.

Futanaris and Males High showcase is only txt2vid generations made using "Wan 2.2 t2v High/LowNoise 14B".

Futanaris and Males Low showcase is only img2vid generations made using "Wan 2.2 i2v High/LowNoise 14B fp8 scaled".

Works for both Img2Vid and Txt2Vid!

• Recommended weight for Img2vid: 0.5~0.8

• Recommended weight for Txt2vid: 0.5~0.8

Usage with clothing: Vast majority of the dataset have the characters (futanaris, men and woman) completely nude - so high weights might result in nipples and penises going through clothes - use the recommended 0.5~0.8 weight range to prevent/mitigate that. Of course that won't matter if they are nude.

• Flaccid and Erect: It should differentiate flaccid and erect penises - use the prompts "flaccid", "flaccid penis", "erect", "erect penis" to get those results. The AI can be VERY stubborn on some poses sometimes, alternating between erect and flaccid state during the animation - I found dad adding something in the prompt that reinforces the state of the penis helps, such as: "the penis keeps flaccid and pointing downwards", "The penis stays erect","the penis keeps erect and pointing up" and so forth.

• Futanari/Male with a Female: its possible to generate normal females together with a futanari/male as well. To make it easier use "a futanari and a woman" or "a man and a woman" to do so - be sure to describe each character separately in the scene.

• No XXX content: there is no XXX content in the dataset - no even masturbation. Its using the basic actions actions and poses that Wan 2.2 recognizes.

11/04 UPDATE - SMOOTH XXX Animations Added!

Added Loras for NSFW animations!

Work for both Img2Vid and Txt2Vid

It covers several poses. None are tied up to an especific prompt so generic prompts - usually used on Images models - should work just fine, such as:

oral sex, vaginal sex, anal sex, handjob, sex from behind, deepthroat, titfuck, cowgirl position, reverse cowgirl position,

Showcase has all images used for the animations with their prompts. I tried to show several different poses and resolutions so you can have a good idea of what is possible.

• Img2Vid

The High and Low Showcases are all using weight 1.0 on both Loras and used SmoothMix I2V Checkpoint Model.

Very reliable and works really well on several poses. For more complex or niche poses combine it with other Loras.

• Txt2Vid

Can have trouble with complex and niche poses but can still help scenes and do alot. I recommend using with Loras that do the scene you want together with this!

Recommended weight 0.8 for Txt2Vid!

The videos on this post used both SmoothMix Txt2vid v.2 Model and the Standard Wan 2.2 16fp. The posts with Loras from other Creators used the Standard Wan 2.2 fp6 Model - The ones without used the SmoothMix Model.

Shout out for the autors of all Loras and their Loras that made training SmoothXXX Animations possible:

(wan 2.2 experimental) WAN General NSFW model by @CubeyAI

Penis Play - WAN 2.2 by @Sweet_Pixeline e

Anal insertion (I2V) by @blo01

Oral Insertion [WAN 2.2] by @LocalOptima

Wan 2.2/2.1 POV Missionary by @dtwr434

Wan I2V (2.2 & 2.1) - Assertive Cowgirl by @icelouse

Handjobs by @f1985

POV Titty Fuck -T2V & I2V by @playtime_ai

SmoothMix Animations

"Epic stuff or lewd stuff... make your choice!"

Make SmoothMix come to life with theses Loras!

There is a lot to explain so lets get started!

• ComfyUI Workflow Settings

All videos on all Showcases used this setup:

->Smooth Txt2Vid Workflow v2.0 on its default Low/Mid settings;

->Wan2.2-T2V-A14B-High/LowNoise-Q8_0.gguf;

->light2x_T2V_14B_cfg_step_distill_v2_lora_rank128_bf16 at weight 3.0 on High and 1.0 on Low.

• Trigger Words

All Loras, Animation and Style, use to the same trigger words:

'SmoothMixAnime' - Will make the video have the style of SmoothMix Illustrious/NoobAI/Illustrious2+NoobAI;

'SmoothMixRealism' - Will make the video have the style of SmoothMix Realism

• Differences

SmoothMix Animation High/Low - Trained with only videos. This Lora serves as a generic animation enhancer for SFW and NSFW animations. There isn't a lot of NSFW explicit content on it so for those you still have to use Loras that are specific for that type of content. All videos were made by using Img2Vid from images made with SmoothMix Illustrious/NoobAI/Illustrious2+NoobAI/Realism.

SmoothMix Style High/Low - Trained only with Images. This Lora servers as a texture enhancer for SFW and NSFW content. SFW and NSFW Images were used on it so it should work well in all cases. All images used were made by using SmoothMix Illustrious/NoobAI/Illustrious2+Noobai/Realism with Smooth Booster to ensure the video will have lots of background and character details.

• Usage examples

The Showcases demonstrate different suggested ways to use these Loras. Let check each of them:

1) Using Both with Priority to Animation:

Animation High and Style High Showcases display my PERSONAL favorite settings - SmoothMix Animation High/Low set to 0.8 with SmoothMix Style High/Low set to 0.2.

Gives a good amount of movement while increasing the details just enough to complement any scene.

2) Using Both with Priority on Style:

Sometimes SmoothMix Animation can give a little TOO much movement. In those cases, using a higher weight on the SmoothMix Style should be ideal - Style Low Showcase display use this settings - SmoothMix Animation High/Low set to 0.3 with SmoothMix Style High/Low set to 0.8.

Reduces the amount of jiggle and bounce to give a more natural animation while giving great amount of details to effects, backgrounds and characters.

3) Using only Animation:

Wind blows, grass grows, breasts bounce and butts jiggles - nothing stays still. If you have yourself a very busy animation in your mind, go all out on this. Animation Low Showcase displays what happens when you use only SmoothMix Animation High/Low set to 1.0.

4) No Trigger Word:

If the visual style of SmoothMix is not what you are looking for - just don't use the trigger words 'SmoothMixAnime' or 'SmoothMixRealism'.

The Loras by themselves will add detail and smooth animation to your videos. It all depends on the type of video you are going for; A video full of action an movements? use SmoothMix Animation; Its a more quiet and very sutle movements? SmoothMix Style.

This post displays video using Animation High/Low set to weight 0.5 with no trigger words.

This post displays video using Style High/Low set to weight 0.5 with no trigger words.

Thank you for all your support!

Please consider visiting my Deviantart page and subscribing! Any help counts.

Thats all for now!

Have fun! =)