Free ComfyUI workflows, Runpod templates, and guides:
👉 hearmemanai.com
For more LoRAs and updates, Join my Discord:
https://discord.gg/ZVWVhT43GW
You can deploy MiniMax on Runpod with my template:
👉 https://get.runpod.io/minimax-template
👉 Deployment tutorial
For Innies version: https://civarchive.com/models/2853405/hminnie-innie-pussy-lora-for-minimax-h3?modelVersionId=3222484
V1:
4x dataset size, contains most pussy and anus types with support for shaved/hairy pussies, smaller file size.
V0.5:
Much better anuses, still not 100% there.
You don't have to stack the video LoRA with this one if you don't care about motion.
V0.1:
Initial version, mostly legible pussies, anuses are very hit or miss.
This LoRA is trained on fingering, spreading and just plain pussy/anus show off.
Two files, and both are needed. Load them together.
Vagina is stills-trained, run at strength 1.0. It restores legible structure in female genital anatomy, which the base model renders soft and vague.
hmpussy is video-trained, run at strength 0.35. It holds that structure together through motion. It is meant to sit under the stills file rather than carry a generation alone.
Trained on MiniMax H3 fl2va, bf16 base. LoKr adapters.
Strengths are not set in stone, you might get better results with different strengths.
I deliberately uploaded good and bad examples so you can see the current state.
Description
FAQ
Comments (60)
Thanks! Now I have to ask to leave work early...
Probably cause it will take 1 hour to generate my 10 seconds video, and every second matters.
@tickaestore799 Why don't you use google colab? Then you get a A100, it is not too expensive.
it's happening! skilled trainers are releasing h3 loras! yeeeh! haven't been this excited since illustrious released!
Thanks!
Just some feedback: In I2V i always get a hard cut to a zoomed in view, mostly without face even if i prompt otherwise. The anus is mostly barely visible or not existing at all. Both might be prompt issues, idk. Looking forwards for your updates :)
yeah it just did that to me too. :( Maybe with only 1 of the files for I2V would work better, I will have to test.
In reference to video I lowered the strengths to 0.7 and 0.2 and it works a little better, try playing around with that
@kaileykk I am mostly doing I2V but I will try. thanks :)
Great idea to split it. I was thinking of the same thing. Power details via image training, and then motion train in a second pass. Glad to see someone else trying it too.
anus? where?
Missionary only as god intended 😂
In version 0.5 :)
@pocketpie okay with missionary, but they can poop then?
@forfreelsd368 You'll have to go on discord for those sort of loras but they do exist I've seen it!
@pocketpie oh, I got one for wan2.2, from some dump on huggingface. Funny, that lora worked extremely good, almost every time picks right hole. I wonder why I keep such things (I test them on some pics with other loras, than just put them in thematic folder and forgetting them, but sometimes really funny to just dig it and remember how much stuff there, and when almost everything from wan2.2 i2v was squeezed... now comes h3, which almost everything do from the box, and I'm like: damn, I have no Idea what to try next, lets try to dig in wan2.2 stuff, and yes, poop was tried also, and... wrong hole. not like I worry too much, I was just excited how simple was to manage that, but without proper anus lora, separated from vagina - many stuff will just not work, as anal sex for example - when dick not already in). Oh yes, h3 can do b3sti4lity (ref2v, I'm not into i2v or t2v right now) better than normal sex, hm... I repeat, I'm not some pervert, but possibilities are really fantastic for someone with a good imagination (and that's not me, sadly).
Discord - not my stuff, I badly write on English, not mention to speak it, but I created couple of "discord loras only" myself for wan2.2, not very good because lack of experience in creating (sometimes - collecting) a good dataset and prompting it correct. Also funny thing: I created wan2.2 i2v lora on single scene where Kratos from GoW3 torns off Helios's head. And you know what? H3 knows how to do that from box, if describe it in colors. Amazing. Me, who created whole (well, only high noise for test) lora for that: wtf, h3?
Make one for Breast sizes , or Nipples
Working on this.
A please next time would've been nice.
Yeah, on behalf of all the people asking for more, I want to say THANK YOU.
Thanks buddy
New Feedback for v0.5 I2V / ref
- Anus spawned and visible :O
- i have no cuts anymore
Thank you!
Thanks buddy!
I2V / ref with 2 loras + 4step lora. This is like a horror movie. :(
@stormstrike52382 in mine wf even without 4step loras those two sometimes broke face consistency and every time broke prompt adherence. Still think it's some wf stuff (used one from this site).
Anyways, I'm still don't get about this model many things, for example how to use bf16 or bf16_pruned version for generations, if white noise all the time I replace int8_convrot...
first t2v pussy lora for mmh3 which works. Congrats
lmfao like all of them have worked. H3 is more showing which users are literate and which arent
@makiaeveli so you really are crying in every single h3 lora comment section aren't you... do better
@MiraStudio lmfao I could leave a comment in each thread if you'd like. I've commented on most new model loras as they come out since before Hunyuan. "Do better" lmfao. And what, my liege, must I do instead? You could downvote, my liege. Or possibly, people could read the manuals and stop shitting on every other creator? Americans are something else lmfao. Seriously, who's the problem here? "Oh my eyes are so harmed by pixels you were allowed to place here, oh how dare thee!!!" I mean you aren't even attacking the content of my comments. This is the definition of ad hominem, so... do better? See, I was attacking the person's argument. That they were wrong to say all the other loras don't work. You are saying I should... shut up?
So strange... When works with Turbo LoRA, everything is fine. However, when I disable Turbo LoRA, the frame starts to distort.
look froward to testing, one small thing, maybe name the files so they obvioulsy related and need each other? with 1000s of lroas eventually it becomes tricky
That's a good idea.
Will do it in the next version
What strenghts when using both LoRas for realism?
thank you for your work!!!
To be honest, the results are fantastic—really realistic. But there's one issue: the LoRA seems to add a pussy to the male characters in the scene as well. Even when they have a penis LoRA applied, a pussy still shows up around the testicle area. I've tried lowering the weight to find a balance, but it doesn't seem to help. I'm using R2V.
I have a Penis LoRA coming up, will probably solve this
I seem to be having trouble with the lora fl2va prunted int8 convrot, I still seem to be getting stock version of the genitalia and not actual ones, I am running both on the suggested strengths, anything I am missing? thanks
innie ver please
In training 🚀
we need innie version
If i'm using this with HMBREAST what recommendation weight that u reccommend? Sorry that im asking this instead of testing by my self cause the generation time for me is really take so long for trial and error.
Now I just need a lora that stops the camera from zooming in all the time no matter what prompt I put in... not that the zoom in issue is specific to this lora
Lemme know when you find one of those looool
Do you manually prompt your videos or do you have an LLM arrange it? I feed the official prompt docs into the system prompt of a local vision model (gemma4 uncensored in this case) with a small explainer at the top explaining its job, and when I tell it to be a static shot with no camera motion it seems to get the message on zoom. But I couldn't tell you what it's doing to make that happen lol.
@sansmia I used manual prompts. I may have to try the LLM approach.
It works extremely well, but way too well, everyone's got pussies now, the men, the woman, the bed, the walls, a little too much. But it works great when it's solo play.
the bed and the walls 😅😆🤣
Needs normalization in the training data
LOL
Thank you! This lora solved the single biggest issue I have with MM. Protrusions coming out of pussies, I don't even need to add the trigger word, it just works!
Yeah its great when it works but 90% of the time it makes the male subject or random things have a vagina lol... I'll wait for the next version as its not reliable enough for me to waste multiple renders and hope to get a good one
each lora works and are life savers, but...
using two at once almost always puts the woman from i2v into a clear background.
trees/city/indoors, doesn't matter, turns bluescreen background.
Works really well, but all these ladies are shaved, and prompting won't produce pubes, at least not so far for me. Pube version please!
Most of the dataset are indeed shaved, will probably hairy one sometime soon
Base model is clearly only trained on shaved mutilated vags :p
Problem is, you'll never be able to train the anatomy of a vagina when its covered with pubes.
Now that the anatomy is pretty much perfect, one could do some additional fine-tuning steps to bring in some variation.
I think @HearmemanAI did the right thing by doing this in stages. First stage is recover the anatomy in Minimax h3, second stage is bring variance
Overall it's a great lora, and works well in most of the positions, expecially laying or from the back bending. But from the front when standing sometimes the clit and genitals starts a bit higher (like 1-2 cm above the expected position) at least compared to the cameltoe position in bikini.
A question, if I decide to use HMinnie, I need to stack the lora with the other two, or do I have to remove some? And what strenghts for the mix of the three to use, if valid?
best pussy lora, thx :)))