stable-diffusion How to finetune inpainting on custom dataset？

How to finetune inpainting on custom dataset？ How to prepare image dataset and run train script？

Oct 24 '22 12:10 universewill

Hey @UniverseWill, did you have any luck with this? Would love to discuss.

Dec 06 '22 23:12 BenjaminIrwin

@universewill also interested in this. Anyone?

Dec 16 '22 02:12 RobertLucian

+1 mark

Jan 05 '23 03:01 hdjsjyl

+1, any pointers on how to get started?

Jan 19 '23 16:01 ovshake

+1 interested in this, looking forward to get any help. Please let me know if anyone has done it or wanted to done. We can sync up on this.

Jan 22 '23 06:01 RahulRangaraj

You may use this repo https://github.com/huggingface/diffusers/tree/main/examples/research_projects/dreambooth_inpaint

Feb 02 '23 22:02 AlexUv2

This may help you https://github.com/CreamyLong/stable-diffusion/blob/master/scripts/train/train_inpaint.sh

Oct 12 '23 08:10 CreamyLong

Any luck anyone? I am not sure which checkpoint to use? Can we use inpaint_big, but why does it have an absurdly large number of conditional channels?

Apr 22 '24 21:04 bhosalems