A double handjob blowbang.
Include the keyword bl0wb4ng in your prompt and wan will handle the rest.
v1.1
Tweaked a few knobs and pushed the training a bit further.
🟢 Less hallucinations
🟢 Better hand skills
🔴 Slightly reduced facial accuracy
Usage
Prompting: keep it simple; avoid NSFW vocabulary when possible.
Shift: keep it low. 7 is OK with most of schedulers and evenly split steps.
Cut: the cut happens at the 9th frame.
Lower Strengths (high-noise):
~0.5: makes a progressive transition instead of a cut.
~0.7: less crowded scene.
Misc
All the showcase videos have their workflow embedded.
Tested only:
with lightx2v LoRAs enabled (1030 or 1022 on high-noise and 1022 on low-noise).
in ComfyUI.
on the following base wan models: Q4, Q5, Q6 and nvfp4.
Training
v1.0: 67.5 epochs, 72 5-seconds clips (4860 steps).
v1.1: 80.0 epochs, 72 5-seconds clips (5760 steps).
Description
Same LoRA, more training
FAQ
Comments (36)
I was here before this gets deleted :D
Fine. If it doesn't get deleted I will block you then.
Read the terms of service for Civitai.red. Fucking moron.
Doubt it, his shoot cumshot lora is still here for months, it has the same blink transition.
@fronyax I hope so. They've removed so many blink-transition LoRAs already)
@zoefollonier so far i couldnt find a terms of service here that prohibit blink transistion.wan 2.2 is already capable for scene cuts.
@qdr1en Scene cut from a portrait to porn is a no go on CivitAI. They rarely last a day. Back it up on huggingface and it will be on the civarchieve.
@meropewantsstls Bonjour, on reste poli s'il vous plait.
@ROXBOT care to back it up with any mention of this in the terms and conditions of civitai.red?
You're just talking out of your ass.
@meropewantsstls Hello, you're not leaving me a choice, sorry. I suggest you to Read the terms of service for Civitai.red, because your insults will not go unanswered, And if you think your insults will disappear with this topic, you're wrong, it's too late.
@cumulus0033697 kindly asking someone to "fuck off" is not technically an insult. It's just general vulgarity.
I hope this helps.
@meropewantsstls Hello, let's imagine that someone treats you this way, what would your reaction be, knowing that you want to be on a quiet site, whether it's nice or not, can you make the effort to not be vulgar ?, I don't blame you, i blame to the words you wrote.
edit : I find it nicer to write pleasant words than words that could hurt.
By the way, thanks to the OP for the model
We can continue the conversation and sorry for my interruption.Hello All, Regarding the subject, it is possible that the policy will change, insofar as we must be responsible for the images and videos we create. If the TEAM decides to try to give us this responsibility, then it would be possible that certain I2Vs will no longer be prohibited, until some users transgress and the TEAM is forced to backtrack for the purpose of protection.
That's pure speculation.@cumulus0033697 I'd laugh about it, of course. You can't take people online seriously.
Especially the ones that are constantly posting things like "I was here before it got deleted lol" when it's not even against the terms of service of civitai.red.
@meropewantsstls hello, thanks for this conversation,
I have a comment about your sentence : We can't take people online seriously.
It's a serious problem, but there's nothing wrong with taking people online seriously. On the contrary, if we try to set a good example, some people might realize that behind the computer there's a human being, and now sometimes an AI ;-)
edit : That doesn't mean we can't make jokes.
have a nice day, and nice animations.
@cumulus0033697 The joke did not make me laugh at all, hence my unfriendly reply to the commenter.
A change in the policy would be truly awesome.
@qdr1en Hello,
I wasn't talking about joking on this subject, but in general, it's obvious that this wasn't a joke here, and I sincerely hope that the communication policy will evolve in the right direction.@meropewantsstls No, just talk from using this site for two years and watching the SFW to NSFW I2V LoRAs getting removed in mass. Go to CivArchieve>filter deleted/Civitat/Wan Vid 2.2 I2V-A14B. There are tons there that have been deleted. I have watched creators reupload and complain that it was removed because the LoRA was removed for going from SFW to NSFW. They would complain that it wasn't in the terms and conditions that it wasn't allowed. They likely linger longer because NSFW has been separated from the SFW LoRAs. Scroll down the list of LoRAs and ask why there aren't more SFW to NSFW LoRAs. There use to be tons of them. You can find them on CivArchieve.
@ROXBOT 7 days and counting, care to back it up with any mention of it in the civitai.red terms and conditions?
@meropewantsstls The closest would be :
AI workflows, models, or tools designed with the intention of removing clothing or otherwise "nudifying" real individuals.
This one doesn't have examples that nudify so that may be its saving grace.
I am speaking prior to the site being split in two. Even though blinks/smashcuts/jumpcuts or whatever you want to call them have been scrubbed since the split. We had to accept a new term of service which means changes have likely been made. I gave you detailed directions on how to see all the prior "Blinks" and similar LoRAs that were removed in the past. There is nothing to prove. Everyone that has been around Wan LoRA for a while knows it. We all watched iGoon's LoRAs get removed, he reposted, removed again, reposted again with SWF examples, removed again.
You need to calm down though. This isn't some great deception; no one is out to get you. Plenty of us have seen them get scrubbed. It happens.
@ROXBOT you could have just said "there isn't one, my bad"
@meropewantsstls never said there was one. Only stated what was observed over time. Go to CivArchieve and see all the ones that were removed. That was what I said.
@ROXBOT Oh, I'm sorry. I was under the wrong impression that you said:
"Scene cut from a portrait to porn is a no go on CivitAI. They rarely last a day" 10 days ago.
My bad.
@meropewantsstls I did and I pointed you to the laundry list of preexisting ones you can access offsite on the achieve that are listed as deleted from Civitai. As I also mentioned creators have complained about their LoRAs being deleted by Citivai for being scene transitions from SFW to NSFW. Go look at iGoonhard's repeated uploads and deletes. I am on creators' discords, Civ deleted their uploads. ToS or not, it is whatever the mods that sees feels like doing at the end of the day. Move on my man. You are arguing for no good reason. I am only stating what has been seen and creator testimony. ToS is just there for the site's protection, they can delete whatever they do or don't want. As I also mention, with the site split, maybe they are more lenient. Don't know, don't care. Go find something better to do that arguing points that I am not disagreeing with you on.
@ROXBOT But I literally just apologized. You clearly did not state that this type of LoRA is a "no go" on this platform and that it would not last a single day 12 days ago.
Again, my bad fam. Clearly, you're the authority on this subject.
This works quite well for solo bjs and hjs too.
How do you do it?
@qdr1en Just specify a single penis, male, or if the image already has a penis then it'll understand. Most loras can do a lot more than they're advertised for, other training is always present regardless if the trainer wants it or not.
thanks! It really seems to love tattoos though..
I didn't notice before you mention it. It may happen, but it should stay rare/marginal: on the presentations videos, there is 1 out of 10 with added tatoos.
Have you tried adding the word in the negative prompt?
@qdr1en I stand corrected. I had messed up my prompt I now see. Small window, bunch of enters and then somewhere explicitely asking for tattoos, which was a left over from a previous prompt...... lol. soryr :-)
Maybe because of the desiwa base model I tried it with, but I don't seem to get a cut, rather two (sometimes three) men walking in from the side and the action begins. Not necessarily a problem because I am getting 0 loss in character consistency. The movement and details are amazing!
I have had better results with base Wan than with any custom checkpoint.
@ss9999 It's normal. Custom checkpoints aren't trained, but have other LoRAs merged in instead. And those LoRAs conflict with this one.
@ss9999 Generally the same for me. When I first started learning about comfyui and wan I tried some of the "mega checkpoints" and everything was a mess, blurry, weird body shapes, and they would mangle loras. I think they can be great, but really only if you stick to what they are good at, and use loras that don't clash. The few that I still might try or use here and there are the checkpoints that are more focused on smooth movement, faster generation times etc. The desiwa checkpoints work pretty well with most loras, but I still mostly use base wan (KJ scaled).
I'm still blown away with what I can manage to generate with a rtx 3070 8GB vram, just takes a bit and uses up a shitload of system ram X)
You should generally avoid checkpoints that aren't entirely custom data.