This is a template to be used with my templated animation script and my vaginal depth model. You need both!
note: you need the model at https://civarchive.com/models/48341/vaginal-depth-concept for the animation to be smooth
Usage
install the script, over at https://civarchive.com/models/42798/templated-animation-script
to install the script just extract the py file to your webui/scripts directory and restart the webui
download this file, it's a zip, with a txt file in it
extract the txt file
open the webui
scroll to the bottom of the webui on the txt2img tab, select 'Prompts from templated file or textbox with GIF generation'
optional: use the generation data, from one of the gifs I set as a sample if you want to generate that exact gif,
set 'frame duration' to your preference, recommended 100, for 10fps
in 'replacement text' enter 'prompt' (NOTE: do not forget to do this or the webui will hang)
option: you can either copy past or upload the file
to copy paste, simply open the txt file from step 3, ctrl a, ctrl c, that file, and ctrl v it into the 'List of prompt inputs' box
OR: you can click on 'click to upload' and select the txt file
enter the details you want into the prompt box, ex,
long hair, red hair, black eyes, happy, happy sex, medium breasts, completely nude, mature female, (thigh grab:1.3), grab, freckles, (arms behind back:1.1), <lora:vaginalDepthConcept_v1:0.7>,click generate! wait for it to finish, might take 20 seconds, might take 20 minutes, depends on your hardware, but at the end, there should be a sexy scene animated with the character you described in your prompt box
Notes
For step 9 above, note that the lora vaginalDepthConcept_v1 was invoked in the prompt, you need that for the animations to be smooth. You can find it at https://civarchive.com/models/48341/vaginal-depth-concept
Also! Style loras help a lot for reducing flicker and making the frames more consistent.
Description
FAQ
Comments (26)
note! if you're one of the six people that downloaded this already, then you need to redownload the file. I uploaded the wrong one, sorry!
It's not that big a deal, there was just an arrant facing away in the prompts, so it was a bit harder to control.
Where it says to download there isn't a link. Is that just on my end, or did the download link get removed?
@Multimulti seems fine to me.
tip! you can do missionary, cowgirl, doggiestyle, and probably some others with this script, just make sure to put relevant info in your prompt.
The template and the vaginal depth model should handle the rest for you.
another tip! torso grab works better than thigh grab
This is pretty based. I haven't tried any of the animation ones yet because I lack understanding. I should try this soon and come up with solutions on how to make them work a little bit better and make them be more consistent. The hair is bugging me, but the rest is pretty awesome. How did you do this anyway?
Edit: You just made me realize we could probably get a consistent eye blink animation doing this sort of thing. I have heard people have managed to make walk cycles, but everything really needs a lot of work, you know? But we are on the verge of greatness and I think that's beautiful :)
the hair you could probably fix by finding a model or lora for the hair you want, which is basically how I did this,
the lora this uses, the vaginal depth one, I trained specifically to have distinct consistent stages of depth, then this makes use of that by having prompts that cycle through those stages.
@SkyLordAer blinks are pretty easy, check out my face_poser model, the old version has a wink animated, and my male_masturbation model has a girl closing her eyes while swallowing, and slight flinch when it first lands :), I have the prompts for those included on their pages.
For a blink, you don't even need a special model for it, all the models merged with NAI are good enough at open eyes, closed eyes, and half-closed eyes, that you just need to time the promtps right and they can do it.
If you are planning on doing it. I recomend using prompt editing. Much much easier to get consistent frames,
```
open eyes
[open eyes:half-closed eyes:0.2]
[open eyes:closed eyes:0.2]
[open eyes:half-closed eyes:0.2]
open eyes
open eyes
open eyes
open eyes
open eyes
```
you'd add whatever else you wanted, but for nine frames, if you do 100ms frames, that should be a decent pace for blinking.
@emberbyte Thanks. Although I don't understand how to do these scripts in the first place so it doesn't help me much yet
this setup as well as AnimatedDiff extension is amazing. With the AnimatedDiff you can set different motions etc. by making a break and then the frame# and what you want to start happening at that point. So with animatedDiff you put the prompt then hit enter and put 0: smile .... 8: open mouth ... 16: closed mouth, and it will animate them pretty well. For this setup, you just edit the text parameters that were provided for us to copy/paste as a basis.
Why when I click Generate, I can only generate one image
I think I know how to use it, but there is another place that makes me wonder, what is Replacement Text used for?
@zz3870471 replacement text is what gets replaced in the template. So, in this txt file, each prompt has in it the word "prompt" you need to put "prompt" no quotes in replacement text, so that the script knows to replace it with what's in your actual prompt box. That's what lets you set the character and whatnot.
@zz3870471 I just added a screenshot to this page of me using the script, hopefully that helps.
@emberbyte praise God
How would i go about getting more frames per second with the gifs? Would i need to add more frames to it and change the frame duration im guessing?
yep. the frame duration is in ms, so 100 is 10fps, lower the frame duration and you'll increase fps.
another option is to use frame interpolation to generate intermediate frames, flowframes is an easy to use tool for that
you could also use the prompt traveling extension to prompt travel between the frames in the list, I'm not sure how well that would work though.
not sure if you're asking how to get more frames (the base text file gives you 8) but the more lines of prompt you put in "List of Prompt Inputs" will equal the amount of frames you get. I just pasted the base text file twice to get it to gen 16 images. Only learned this because FlowFrames doesn't work with just 8 frames and I was very sad at first.
thank you emberbyte
this improve my life
could you fix it with controlnet?
大佬
Sorry, I used your parameters to test but raised an error:TypeError: Script.run() missing 6 required positional arguments: 'checkbox_iterate', 'checkbox_iterate_batch', 'frame_duration', 'show_frames', 'single_line_prompt', and 'prompt_txt'
Bro I kinda got it working and going to try to improve my results more. But why does it absolutely mess up the very first image. lol thats my only problem so far
Nevermind lmao I just had to reload the ui. It works pretty good thanks
I only get one photo from following the directions
Does it only work with automatic1111 or it can work with comfyui