Dreambooth Stable Diffusion training in just 12.5 GB VRAM, using the 8bit adam optimizer from bitsandbytes along with xformers while being 2 times faster.

[ZeroCool22]

Tested on Nvidia A10G, took 15-20 mins to train. We can finally run on colab notebooks.

Code: https://github.com/ShivamShrirao/diffusers/blob/main/examples/dreambooth/

More details huggingface/diffusers#554 (comment)

https://www.reddit.com/r/StableDiffusion/comments/xphaiw/dreambooth_stable_diffusion_training_in_just_125/

[dvztimes]

This is really cool. You got it running locally?

I just clone that repo into its own folder, run the commands listed there, and it will just work? I have a 16gb mobile video card.

[ZeroCool22]

This is really cool. You got it running locally?

I just clone that repo into its own folder, run the commands listed there, and it will just work? I have a 16gb mobile video card.

No, i don't even tried because i have a 1080 TI (11gb).