Originally posted to HuggingFace by Norod78 and shared on Reddit.
Description
Trained on top of SD 2.
Note: To use this you'll need to also download the accompanying YAML config file and place it alongside the model.
FAQ
Comments (12)
Don't know about the rest of the world, but I'm getting:
RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cpu and cuda:0! (when checking argument for argument index in method wrapper__index_select)
Only seems to happen on this ckpt
MythicaDec 12, 2022
Same issue here, can't get it work.
Sorry guys. I failed to attach the YAML file that you need to include for this to work correctly.
I'll include that now.
JustMaierDec 12, 2022
We're working on a longer-term fix for this now.
meritrash6350Dec 13, 2022
I'd also suggest that you actually match up the names of your .yaml and. ckpt for the sake of avoiding confusion for those who don't know they need to match.
I trained an SDv1.5 version in addition to the SDv2.0 one so it could be used without a custom .yaml file
please make/add one that's not just for Xmas, i love claymation!!!
You'd probably like this model then:
https://civitai.com/models/82/clazy
The HuggingFace page indicates that it was trained from SD 2 - does this mean it won't work in an implementation that doesn't yet support 2.0? (DiffusionBee, Invoke-AI, etc)
richpizor323Dec 13, 2022
Answering my own question: this model absolutely requires SD 2.0 compatibility. Would be appreciated if that could somehow be labeled.
Thank you for pointing this out, I trained an SDv1.5 in addition so to not limit anyone
richpizor323Dec 15, 2022
Thank you!