ᴛʜɪꜱ ʟᴏʀᴀ ɪꜱ ᴘʀᴏᴠɪᴅᴇᴅ ᴀꜱ ɪꜱ ᴀɴᴅ ɪɴᴛᴇɴᴅᴇᴅ ꜰᴏʀ ʟᴀᴡꜰᴜʟ ᴀɴᴅ ʀᴇꜱᴘᴏɴꜱɪʙʟᴇ ᴜꜱᴇ.
ɪᴛ ᴀᴄᴛꜱ ᴀꜱ ᴀ ɢᴇɴᴇʀᴀʟ-ᴘᴜʀᴘᴏꜱᴇ ʀᴇꜰɪɴᴇᴍᴇɴᴛ ʟᴀʏᴇʀ, ɪɴꜰʟᴜᴇɴᴄɪɴɢ ʜᴏᴡ ᴛʜᴇ ʙᴀꜱᴇ ᴍᴏᴅᴇʟ ᴇxᴘʀᴇꜱꜱᴇꜱ ᴘʀᴏᴍᴘᴛꜱ ᴀɴᴅ ᴄᴏɴᴄᴇᴘᴛꜱ ʀᴀᴛʜᴇʀ ᴛʜᴀɴ ᴘᴇʀꜰᴏʀᴍɪɴɢ ᴅɪʀᴇᴄᴛ ᴛʀᴀɴꜱꜰᴏʀᴍᴀᴛɪᴏɴꜱ. ᴛʜᴇ ʀᴇꜱᴜʟᴛɪɴɢ ᴏᴜᴛᴘᴜᴛ ɪꜱ ᴀʟᴡᴀʏꜱ ᴅʀɪᴠᴇɴ ʙʏ ᴛʜᴇ ᴘʀᴏᴍᴘᴛ ᴀɴᴅ ᴛʜᴇ ᴄᴀᴘᴀʙɪʟɪᴛɪᴇꜱ ᴀʟʀᴇᴀᴅʏ ᴘʀᴇꜱᴇɴᴛ ɪɴ ᴛʜᴇ ᴜɴᴅᴇʀʟʏɪɴɢ ʙᴀꜱᴇ ᴍᴏᴅᴇʟ.
ᴛʜᴇ ʟᴏʀᴀ ᴅᴏᴇꜱ ɴᴏᴛ ʀᴇʟʏ ᴏɴ ᴘᴀɪʀᴇᴅ ɪɴᴘᴜᴛꜱ, ʙᴇꜰᴏʀᴇ/ᴀꜰᴛᴇʀ ᴇxᴀᴍᴘʟᴇꜱ, ᴏʀ ᴛʀᴀɴꜱꜰᴏʀᴍᴀᴛɪᴏɴ-ꜱᴘᴇᴄɪꜰɪᴄ ᴛʀᴀɪɴɪɴɢ. ɪᴛ ꜰᴏᴄᴜꜱᴇꜱ ᴏɴ ɪᴍᴘʀᴏᴠɪɴɢ ᴄᴏʜᴇʀᴇɴᴄᴇ, ᴅᴇᴛᴀɪʟ, ᴀɴᴅ ꜱᴛᴀʙɪʟɪᴛʏ ᴀᴄʀᴏꜱꜱ ɢᴇɴᴇʀᴀᴛɪᴏɴꜱ, ɪɴᴄʟᴜᴅɪɴɢ ᴍᴀᴛᴜʀᴇ-ᴛʜᴇᴍᴇᴅ ᴏᴜᴛᴘᴜᴛꜱ ᴀʟʀᴇᴀᴅʏ ꜱᴜᴘᴘᴏʀᴛᴇᴅ ʙʏ ᴛʜᴇ ʙᴀꜱᴇ ᴍᴏᴅᴇʟ.
ɴᴏ ꜱᴘᴇᴄɪᴀʟɪᴢᴇᴅ ᴅᴀᴛᴀꜱᴇᴛꜱ ᴛᴀʀɢᴇᴛɪɴɢ ɪᴅᴇɴᴛɪᴛʏ ᴇxᴘᴏꜱᴜʀᴇ ᴏʀ ᴄᴏɴᴛᴇɴᴛ ʀᴇᴄᴏɴꜱᴛʀᴜᴄᴛɪᴏɴ ᴡᴇʀᴇ ᴜꜱᴇᴅ ᴅᴜʀɪɴɢ ᴛʀᴀɪɴɪɴɢ. ᴀʟʟ ꜰᴀᴄᴇꜱ ᴀɴᴅ ᴘᴏᴛᴇɴᴛɪᴀʟʟʏ ɪᴅᴇɴᴛɪꜰɪᴀʙʟᴇ ꜰᴇᴀᴛᴜʀᴇꜱ ᴡᴇʀᴇ ᴍᴀꜱᴋᴇᴅ ᴏʀ ᴀɴᴏɴʏᴍɪᴢᴇᴅ ᴛʜʀᴏᴜɢʜᴏᴜᴛ ᴛʜᴇ ᴛʀᴀɪɴɪɴɢ ᴘʀᴏᴄᴇꜱꜱ.
ʙʏ ᴅᴏᴡɴʟᴏᴀᴅɪɴɢ ᴏʀ ᴜꜱɪɴɢ ᴛʜɪꜱ ʟᴏʀᴀ, ʏᴏᴜ ᴀᴄᴄᴇᴘᴛ ꜰᴜʟʟ ʀᴇꜱᴘᴏɴꜱɪʙɪʟɪᴛʏ ꜰᴏʀ ɪᴛꜱ ᴜꜱᴇ ᴀɴᴅ ᴀɢʀᴇᴇ ᴛᴏ ᴄᴏᴍᴘʟʏ ᴡɪᴛʜ ᴀʟʟ ᴀᴘᴘʟɪᴄᴀʙʟᴇ ʟᴀᴡꜱ ᴀɴᴅ ᴘʟᴀᴛꜰᴏʀᴍ ʀᴜʟᴇꜱ.
Female ejaculation / SQUIRT LoRa for - I2V Wan 2.2 Video LoRa - K3NK
Trained on a dataset of 355 clips with various resolution, 1sec to 3sec
Various different source videos.
Besides the newest distilled loras, you can also try the old I2V Lightx2v, in some cases this lightx2v manages better fluids.
The workflows im using for wan2.2 is:
* you can also find the workflow inside the sample video's metadata
Example prompt:
visible liquid squirting upwards from her vagina, she is squirting a large amount of fluid in short bursts going upward. Her facial expression shows an open mouth, in state of arousal. She looks at the viewer. fluid is visible, indicating squirting.A woman with her legs spread wide with her (vagina squirting clear fluid upwards hitting the camera lens), Her facial expression shows an open mouth, in state of arousal. She looks at the viewer. (highly detailed, anatomically correct)Have fun! 😜
Description
FAQ
Comments (30)
I downloaded like a lightning since it is very most likely to be banned... Will you be able to make one for LTX 2.3 ?
tried the spitting lora and saliva was not there, i still have to find the correct parameters for each dataset.. kind of sucks..
@K3NK Looks different from Wan yea..
having trouble downloading the high noise.. is it broken? low lora was fine...
Same for me. This website is falling apart at the seams, if you haven't noticed.
Don't worry, their top people are on it. https://civitai.red/images/127846092
me too
yeah some server issues i guess, it download just fine for me
same issue with high
cant download high like the others, I wonder if "Unverified: Scan requested" has something todo with it. Perhaps its pending ?
can you put these into a zip as an 'other' file? not sure how long we'll last without this in our collections
Saw it was verified , still dont work. Weird.
I have the same issue, can't DL the high model. I saw something similar with a different model yesterday. Odd that some people can download it and others can't. Oh well...<shrug>
@K3NK You have a Lora that's gone, any object can be blown up.
ive reuploaded it 1h ago, still no luck?
@K3NK I tried clearing civit.red cache but still not able to download, clicking download just refreshes the page. Maybe add it as "training data" to the low file?
@iodrg244 added zip
@K3NK thanks! training data worked. thanks for doing that : )
@K3NK it doesn't seem to want to tag my image so not sure if this video is visible but I gave it a shot https://civitai.red/images/128121732
@iodrg244 your temporal frame count seems to be under the actual frame count, that's what tends to cause those lighting oddities :)
its on civit archive if anyone else is still struggling :) downloads fine there.
@K3NKi got it off civit archive mate, thanks :)
Was really hoping for a "realistic" one of these, kinda weird that no one's doing that
I haven't tested the new one yet, but at least with the previous version of this lora, I was able to achieve a less exaggerated effect by only using the low noise model and dropping its strength down to 0.5. Still not perfect but no fire hoses spurting up into the air
The realistic ones were removed, check'em on civarchive.
I tried many ways, but in the video, she squirts first from her mouth. Does anyone know how to resolve this?
Haven't played with this lora yet but if you are using any models or loras that are based off NSFW material they might be affecting your output. Suggestion is to go with the vanilla wan2.2 model but still use the any NSFW loras that you were using.
What are you using to caption your dataset? And what gpu do you use for lora training? Also do you train at 268 or 512?
epic awesome
Fantastic work!
Any plans on an LTX23 model?
Details
Files
wan22-squirt-i2v-162epoc-low-k3nk.safetensors
Mirrors
wan22-squirt-i2v-162epoc-low-k3nk.safetensors
wan22-squirt-i2v-162epoc-low-k3nk.safetensors
wan2.2-i2v-low-squirting.safetensors
wan22-squirt-i2v-162epoc-low-k3nk.safetensors
Female SQUIRT v2.0 I2V LOW - Wan2.2.safetensors
Female SQUIRT v2.0 I2V LOW - Wan2.2.safetensors
wan22-squirt-i2v-162epoc-low-k3nk.safetensors
wan22-squirt-i2v-162epoc-low-k3nk.safetensors
wan22-squirt-i2v-162epoc-low-k3nk.safetensors
wan22-squirt-i2v-162epoc-low-k3nk.safetensors
wan22-squirt-i2v-162epoc-low-k3nk.safetensors
wan22-squirt-i2v-162epoc-low-k3nk.safetensors