How to Make AI Images of Yourself | Flux Local LoRA Training in 16GB VRAM
ฝัง
- เผยแพร่เมื่อ 19 ต.ค. 2024
- My Stable Diffusion Course:
hojenaia.com.br/
Become a Channel Member to gain many benefits:
/ @hojenaia
Links in this video:
github.com/bma...
github.com/koh...
developer.nvid...
Kohya installation:
git clone --recursive github.com/bma...
cd kohya_ss
git checkout sd3-flux.1
I needed to edit the requirements_linux.txt file in the root folder and put this in line 1:
torch==2.4.0 torchvision==0.19.0 --index-url download.pytor...
chmod +x ./setup.sh
./setup.sh
source venv/bin/activate
Copy the scripts and configs into the sd-scripts folder:
gist.github.co...
gist.github.co...
Modify the files according to the folders where you downloaded Flux, CLIP and T5. And according to your dataset.
cd sd-scripts
./train.sh
My Mac:
MacBook Pro M3 Pro: amzn.to/3Swm2eG
My PC:
Intel Core i7-12700KF: amzn.to/4b5R6Jv
Asus ROG STRIX Z790-F: amzn.to/3QwlLYr
NVIDIA RTX A4500: amzn.to/3ybOR9e
Kingston DDR5 FURY Beast Black: amzn.to/3UQR55F
SSD FURY RENEGADE: amzn.to/3WuDRh6
NZXT H5 Flow case: amzn.to/3QBLq1I
Elgato Wave 1: amzn.to/4bt8c48
Canon EOS Rebel T7: amzn.to/3Wrrlz5