Closed Desm0nt closed 1 year ago
Make sure you haven't modified StableDiffusionPipeline.py to remove the safety checker. Setting the safety checker as a dummy function confuses the script, because it can't figure out what kind of model it is.
Ideally for this usage it would be nice if the save_pretrained
function avoided failing in this case.
Thanks for the issue! @briansemrau answer seems right! We have verified the script with the orginal .ckpt
file and works as expected, please check if you have not modified anything.
I have the same problem, I can't convert a merged ckpt. Probably one of those was not an original model of stable diffusion? I haven't touched anything, I use the script in colab and always clone the diffuser repository at every runtime. Are there any other requirements to convert the ckpt? Maybe the size or a mandatory config.yaml?
Hey @Desm0nt or @jujpenabe,
Could you maybe upload the original .ckpt
file somewhere to the Hugging Face Hub: huggingface.co so that I can reproduce the error locally?
Closing to inactivity - we need access to a checkpoint to check this - will otherwise follow up on https://github.com/huggingface/diffusers/issues/671
Describe the bug
I have some ckpt files and try to convert it to diffusers in google colab. with this command
!python /content/diffusers/scripts/convert_original_stable_diffusion_to_diffusers.py --checkpoint_path /content/gdrive/MyDrive/Waifu_6e.ckpt --dump_path /content/stable-diffusion-alt
But got this error:
It works fine with the same model 2 days ago.
Reproduction
No response
Logs