Errors: Converting inpainting diffusers to OG stable diffusion

[BenjaminIrwin]

Am I right in saying the diffusers to OG stable diffusion script (link below), shouldn't work for a non-standard architectures?

I've finetuned an inpainting model from Runway's SD 1.5 Inpainting checkpoint. It has 9 input channels instead of 5. I created a .ckpt file from this finetuned model with the script below, but it's crashing Automatic111's repo when I run it.

https://github.com/huggingface/diffusers/blob/main/scripts/convert_diffusers_to_original_stable_diffusion.py

[patrickvonplaten]

Hey @BenjaminIrwin,

We don't actively maintain the diffusers to OG stable diffusion script. Maybe someone from the community can help here though :-)

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.

[zaryabmakram]

Hi @BenjaminIrwin,

I would really love to have a detailed guide on how you were able to finetune the Runway's SD 1.5 Inpainting checkpoint on a custom dataset for inpainting. It would be great if you could describe to me your process on how I can do the same on my own image-mask dataset.