search

apple / ml-stable-diffusion

Stable Diffusion with Core ML on Apple Silicon
MIT License
16.83k stars 941 forks source link

Instruction for model training #134

Open mozzipa opened 1 year ago

mozzipa commented 1 year ago

I am very new for this work. While generating is able from python to swift, I could not find applicable way to make customized model by myself. Dreambooth of Diffusers from huggingface is not working for my 16GB M1 mini. - Actually dreambooth for 8GB was able but it took more than 10 hours for their dog example. I expect that there will be better way to be applied for Apple silicon.

If there is several instruction or example for tunning or training to be applied ml-stable-diffusion, it will be very helpful for persons as like me.

Osub commented 1 year ago

M1 MAX is not support?

kartzke commented 1 year ago

+1 for adding the ability to train models natively