A WebUI that integrates dataset acquisition, post process and PLoRA / LoRA training, aiming to simpler dataset operations and manual training.
Almost all functions are from DeepGHS' work, these pipeline things are awesome! However, this training interface has no queue capabilities now.
But we can do full-auto training now, select your favorite preset and fill in waifu name, then press the button!
We have both HCP & kohya-ss frameworks supported!
I didn't know that there was already webui support for plora training, so I tried writing this and later added some processing functions.
If you are interested in full-auto pipeline or image classification, it's recommended to follow the DeepGHS team,
there are something interesting there :)
- Python 3.10.6
- Dependencies will be installed on first start.
- If there are something broken, delete
venv. - Torch can be install automaticlly now, CUDA version 11.8 and 12.1 can be detected from device.
- If your CUDA is not supported, you can also install it manually.
Install Command
git clone https://github.com/LittleApple-fp16/LittleAppleWebUI.gitRun webui.bat on Windows
For Linux, run webui.sh instead.
- WARN: This is a lightweight toolbox, and most of the work will be done in RAM.
- I don't understand Python, and I can't guarantee too much functionality for this. However, if there are any issues with UI usage, feel free to let me know.
- By the way, if you want advanced plora training, please contact HCP-Diffusion developers for assistance.