Home Artists Posts Import Register

Content

Current tutorial for the Stable Diffusion GUI:

Will keep updating the files to try to improve them in the future.

Prompt Examples:

https://docs.google.com/document/d/1f2vAJnwaJw4KisnHM_r8i0e7nWuGK1Wvc81o1U87NrY/edit?usp=sharing 

Create Inpainting/Outpainting:

https://docs.google.com/document/d/1_Fc36pHyPhiokTwyqvrD-qYg_4Q62NYduI98pMtv8pc/edit?usp=sharing 

Training Dreambooth:

https://docs.google.com/document/d/15gsTo2IaNxSfGox_RQatg1nuIyxs3Tm0fyUZMUHTXbw/edit?usp=sharing 

Downloading models from HuggingFace site:

https://docs.google.com/document/d/15_nOmXsQJxDeM3XDSIFLjXceB4bZMsmtWr53iZs98P4/edit?usp=sharing 

Comments

Anonymous

I used the dreambooth to train a couple custom models in areas where SD had no idea what to do with the text prompt. I used images of bellydancers in costume and now I can generate images of bellydancers. Previously, SD just made random garbage with that type of prompt. I also used about eight images of my wife from 25 years ago that I took at my photo studio, and now I can generate images of her, as well. I did one of her in costume as Wonder Woman and showed it around at thanksgiving. Nobody thought twice about it until I pointed out that the costume is from the new movies, and she was in here fifties when it came out, not her thirties. Lot's of fun, so thanks for that. I do have a question about training. Do I need to train all the images I want to include all at the same time using the text files, or can I use a custom model I made as the basis for adding more training with a custom prompt? For instance, I trained the dancers and my wife separately; could I use the dancers.pt file to train again using the images of my wife?

Anonymous

Might be a good idea to pin this to the top or add these links to pinned post

cool1

Yes you can train again with the source ai model being one you already trained on (and it will create a new output ai model like it normally does). I don't know if that's the best way though but it lets you.