Trained for 2000 steps with:
#|eyes closed%{random_prompt}, eyes closed:-0.1|eyes closed++:guidance=2
with https://github.com/ntc-ai/conceptmod
Eyes open animations are backwards, going from 0.7 to -0.7 with the trigger words "eyes closed"
Guide: https://civarchive.com/models/58873/conceptmod-tutorial-fire-train-any-lora-with-just-text-no-data-required
Description
FAQ
Comments (10)
what about not using your reference code and doing it locally.
@EvilFear https://civitai.com/models/50070/mania-concept-trained-without-data-tutorial Here's a tutorial for that. I put this together first but it's very technical and many people had trouble. The dockerfile I built is available here: https://github.com/ntc-ai/conceptmod/blob/main/docker/Dockerfile_train
@ntc alright, thank you :)
Is it possible to use SDXL or is it only for 1.5?
i dont understand it. pls help
Hi, I trained a lora using text to make a keyword more powerful. The technique is open source and you can train your own with the links provided.
Not sure if this helps, a more specific prompt might be helpful for me to explain better.
I meant how can I create the series of images.
@Stevem61 This might help https://civitai.com/models/57334/angry-trained-without-data-new-runpod-easy-animations-for-any-lora
Explains how it's done with the conceptmod github.
img2img not work
prompt: <lora:eyesclosed:1>, eyes closed,
it seems from -0.7 to 0.7，all pic are eyes closed with trigger words "eyes closed"
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.