Hello everyone, I hope you're doing well!
This is an updated guide for stable diffusion fine-tuning methods, it covers 4 methods: Dreambooth, Textual Inversion, Hypernetworks, and LoRA. This tutorial is tailored for low VRAM users.
Used material links:
Automatic1111 Launcher repo: https://github.com/EmpireMediaScience/A1111-Web-UI-Installer
Kohya_ss GUI: https://github.com/bmaltais/kohya_ss
Kohya config file: https://drive.google.com/file/d/1oOw8iyeNxvyJMimTWkc4jxMDQvDt4P4z/view?usp=sharing
Git downloads: https://git-scm.com/downloads
Resize and Crop free tool: https://www.birme.net/
Civitai: https://civitai.com/
Paper Links:
Dreambooth: https://arxiv.org/abs/2208.12242
Textual Inversion: https://arxiv.org/abs/2208.01618
Hypernetworks: https://arxiv.org/abs/1609.09106
LoRA: https://arxiv.org/abs/2106.09685
Artist list for stable diffusion:
https://stablediffusion.fr/artists
Tools and Resources for AI Art:
https://pharmapsychotic.com/tools.html
--------------------------------------------------------------
Let's connect:
LinkedIn: https://bit.ly/3roXgQ2
GitHub: https://bit.ly/3CrfRRP
Kaggle: https://bit.ly/3C1mqZD
Twitter: https://bit.ly/3UR06e3
--------------------------------------------------------------
♪ Chillpeach - Daydream : https://youtu.be/tlGiSnhClxY
--------------------------------------------------------------
If you have any question, suggestion, or remark. Feel free to leave it in a comment below.
See you next time!
#MLWH
Chapters:
00:00 Intro
01:25 Finetuning Methods Explained
11:50 Installing Auto1111 & Kohya_ss UIs
19:16 Preparing Training Data
22:44 Textual Inversion
30:30 Hypernetworks
32:53 Dreambooth
42:23 LoRA
48:06 Results Comparison