-
Notifications
You must be signed in to change notification settings - Fork 2.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Bump diffusers
version (and accelerate
, hugginface-hub
)
#7174
Conversation
Graph:
|
Tested CPU; with and without xformers; installation of all starter models; generated in many different ways on canvas, including controlnets (with all preprocessors) and ip-adapter. All LGTM |
Seen that before It's an issue inherited from Diffusers I think. I believe it's a combination of the scheduler config from the repo and the scheduler in use by InvokeAI. The load of the scheduler from default passes the config on , but sometimes they have bits in the config that the scheduler chosen in InvokeAI doesn't support. DreamShaper does it as its 'default' scheduler is DEIS which conflicts with DPM schedulers |
This param has been deprecated, and by including it (even when set to False) the safety checker automatically gets downloaded.
I tested on MacOS (SD1.5 text-to-image, SDXL text-to-image, SD1.5 IP-Adapter, SD1.5 Controlnet, SD1.5 LoRA). Everything runs ok. The image outputs are changed slightly for the same seed. The cause of this is unknown, and it's probably not worth the effort to track it down. |
Quality Gate passedIssues Measures |
Summary
Bump diffusers and its dependencies.
The main motivation for bumping now is for access to the latest SD3 code.
QA Instructions
I smoke tested the follwing on CUDA:
Remaining tests:
Merge Plan
No special instructions.
Checklist