Train dreambooth on 11GB GPU with prebuilt docker image.

[ZeroCool22]

https://github.com/smy20011/efficient-dreambooth

https://www.reddit.com/r/StableDiffusion/comments/xt0x6b/train_dreambooth_on_11gb_gpu_with_prebuilt_docker/

This user claims that made works DB on a 1080 TI.

[avclubvids]

@AUTOMATIC1111 - I know there’s textual inversion in here now, but could you elaborate on what it would take to get Dreambooth working inside the same webgui? It would be supreme if there were a new tab for training so it was all in one place vs the current mess of online Dreambooth to update offline webgui

[ZeroCool22]

💡NEW UPDATE, now on 10gb VRAM!:

DreamBooth Stable Diffusion training in 10 GB VRAM, using xformers, 8bit adam, gradient checkpointing and caching latents.

[ClashSAN]

@ZeroCool22 have you gotten these running locally?