Great work! I've been playing with diffusion pipe as well and have been trying to figure out a good set of hyperparameters. Did you use the constant LR scheduler? What LR? I assume with the small dataset size just batch 1?

Point of order - did you really upload two different LoRAs with the exact same name?

This is the file I get for both versions:

genitals_helper_v1.0_e219.safetensors

Are they the same file?

Great work!

The performance in the anime style is not very good, and color difference may occur.

So realistic

Overall this LoRA does what it intends to do, and it does it well.

However, I can see almost no difference between "outie" and "innie", they all look innie in my gens no matter the prompt and LoRA/prompt strength.

It also seems to change the face too dramatically in I2V gens. Could possibly be because this is made for 480p and I'm using the 720p checkpoint for my videos... If that's the case, disregard this part.

Other people seem to be having more success than me with it though, which is great to see (and a testament to your skill).

Any tips for preventing the camera from zooming in, keeping a higher angle POV perspective, etc.?

IMPORTANT: INFORM ALL WAN USERS! I DISCOVERED HOW TO USE LORAS 'NSFW/SFW' IN WAN 2.2 T2V, GETTING THE FULL POTENTIAL OF "HIGH NOISE" AND "LOW NOISE." THE RESULTS WERE IMPRESSIVE!! I ran it on an RTX 3060 6GB, 32MB RAM, 480x480, length: 41, generation time: 5 minutes to generate each video. The quality is fantastic! For those of you with more powerful computers, the results will be even more incredible. I used the models: WAN 2.2 - Q4_1.gguf FOLLOW THE STEPS BELOW: 1: Don't use acceleration/enhancement LoRas (such as Lightx2v or other) on "high noise"; use only on "low noise" with strength 1. (I'm using this version of Lightx2v LoRa): lightx2v_T2V_14B_cfg_step_distill_v2_lora_rank64_bf16 2: Use 'NSFW/SFW' LoRas on both "high noise" and "low noise" models. On "high noise," use strength 2, and on "low noise," use half the strength, with strength 1. You can vary the strengths as you like, but always keep the "low noise" strength at half the strength of the "high noise." 3: Use these settings in KSamplers: "high noise": steps: 9 / cfg: 3.5 / euler/simple / start_at_step: 0 / end_at_step: 1 "low noise": steps: 9 / cfg: 1.0 / euler/simple / start_at_step: 1 / end_at_step: 10000 Note: To improve speed, I set the Windows system to "best performance," disable unnecessary background programs and applications, and lower the desktop resolution. This significantly speeds up video generation. I hope this helps.

Can I use this Lora on TensorArt with Wan2.2? Any idea how to do it?

Im getting outie even when prompting innie. Any trick to solve?

Bonjour, je suis nouveau dans la génération de photos de l'IA, et je cherche quelqu'un qui pourrait m'apprendre à générer des vidéos comme celle ci-dessus moyennant des frais ;) merci !

What an example to select lol

anyway to prevent its going half way up to belly button? :) I tried various strengths between 1 and 0.5 on both high and low on wan2.2.

UPDATE: works GREAT on wan 2.2, no unwanted zooming as long as i tell the model not to zoom, 5 stars!

Looks like the LoRA forces zooming in. I tried to put "zoom in" into negative prompt and "keep framing" and/or "do not zoom in" into positive, even adjusting strenght didn't help. Aesthetic is nice, then i think it's a necessary fix, i just want the model to do what i tell to.

Anyone been able to get this to work on mage.space

Seems to work well.

Hi everyone! I hope that is what I need to generate more accurate wg with wan 2.2 . My biggest problem is to get some sample workflow for that purpose (I2V and T2V) to obtain good results in erotic content creation.

This Lora worked for quick action sequences, where 2 other popular (and good) Loras failed.

that preview image needs to change, its so bad

HOLA GENTE SALUDOS PARA TODOS AGRADESERIA MUCHO ALGUNA ALLUDITA EN ESTE MUNDO SOY NUEVO POR AK VENDICIONES

Sadly on i2v on Wan2.2 this lora alters the characters face. I am unable to use for this very issue

Wan 2.2 (high noise / low noise) version coming soon?

I'm having issues with i2v where the vagina is upside down. Anyone have any solutions?

Theres a weird issue with this where it makes anime style characters develop luscious red lips. If you ever update this might want to try and filter out the red lips in the training or whatever! tag it better? idk how it works lol

am i doing something wrong? when I try to run all i get is noise and no picture.

This cant be used with other Loras such as SmoothmixXXX which diminishes much of it's utility.

Still one of the best loras for genitals, we're waiting for a wan 2.2 version. If you plan to create it, consider correcting one thing: this lora seems to force the woman touching herself, annoying when i want to make her just show.

Thass a lotta arby's

Works well, but on character lora be sure to start low, like 0.3. higher values can mess with my characters

how to get good and consistent result with wan 2.2? what strength for high and low?