KeyError: 'state_dict' when trying to load checkpoints merged with Automatic1111 #5451
Unanswered
ClaraSanders
asked this question in
Q&A
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
EDIT: Thank God I'm not the only one. According to InvokeAI guys this happens because "Merged models from auto11 merge board are flat for some reason. The current behavior of invoke is not changed by this modification." I don't know what it means but I think you guys should check it out because it didn't happen before. Apparently, InvokeAI fixed it:
invoke-ai/InvokeAI#1766
Can a kind soul help me with this issue before I jump off the window, please? I always used Automatic1111 to mix and merge my dreambooth models until the update that added the safetensors option. Now I still can merge models and stuff but the merged checkpoints ONLY work on Automatic1111, when I load them on for example hlky or NMKD Stable Diffusion GUI I get the infamous Python error: KeyError: 'state_dict'
Any idea what the hell is going on? Why before the update all merges loaded perfectly anywhere and now they run only on Automatic1111? I tried to not rename the checkpoints, I tried sum weights and add difference. Nothing solves this issue.
Beta Was this translation helpful? Give feedback.
All reactions