easydiffusion / easydiffusion

Easiest 1-click way to create beautiful artwork on your PC using AI, with no tech knowledge. Provides a browser UI for generating images from text prompts and images. Just enter your text prompt, and see the generated image.
https://easydiffusion.github.io/
Other
9.42k stars 776 forks source link

Is there a way to train a lora for Easy Diffusion? #1669

Open fulcrumrebel opened 9 months ago

fulcrumrebel commented 9 months ago

I am looking for a way to train a Lora for faces for Easy Diffusion. I find a lot for Stable Diffusion and I know that they are both similar, but is there a tutorial specific for Easy Diffusion?

I am looking to make a lora of my face and my husband's face to make cool art pieces with of us as different characters for gifts. If you have any suggestions, I would appreciate it. I was trying to find a free google colab notebook to help me, but have only found ones I have to pay for.

AtomicLugia commented 9 months ago

https://civitai.com/models/22530/guide-make-your-own-loras-easy-and-free

I used this to make my own LoRA's. Works fine with Easy Diffusion

Skoteinos1 commented 4 months ago

https://www.youtube.com/watch?v=Un9SHPVAAbE

With one person it is simple. Just put bunch of images of desired person into your google drive and then run google colab lora trainer. It takes like 20 minutes Guess with 2 people it is same... but you really don't want to confuse AI. You may end up with mix of yours and partners face.

Consider just generating an image with 2 random people. Then split image in half in Paint, do face-swap for one person, than face-swap second person and finally join two halves of image in Paint together. You can get like 30 free daily face-swaps on seaart.ai.